Prashant M.

Pawar ·
R. Balasubramaniam ·
Babruvahan P. Ronge ·
Santosh B. Salunkhe · Anup S. Vibhute ·
Bhuwaneshwari Melinamath Editors

Techno-Societal
2020
Proceedings of the 3rd International
Conference on Advanced Technologies
for Societal Applications—Volume 1
Techno-Societal 2020
Prashant M. Pawar · R. Balasubramaniam ·
Babruvahan P. Ronge · Santosh B. Salunkhe ·
Anup S. Vibhute · Bhuwaneshwari Melinamath
Editors

Techno-Societal 2020
Proceedings of the 3rd International
Conference on Advanced Technologies for
Societal Applications—Volume 1
Editors
Prashant M. Pawar R. Balasubramaniam
SVERI’s College of Engineering Precision Engineering Center
Pandharpur, Maharashtra, India Bhabha Atomic Research Centre
Mumbai, Maharashtra, India
Babruvahan P. Ronge
SVERI’s College of Engineering Santosh B. Salunkhe
Pandharpur, Maharashtra, India SVERI’s College of Engineering
Pandharpur, Maharashtra, India
Anup S. Vibhute
SVERI’s College of Engineering Bhuwaneshwari Melinamath
Pandharpur, Maharashtra, India SVERI’s College of Engineering
Pandharpur, Maharashtra, India

ISBN 978-3-030-69920-8 ISBN 978-3-030-69921-5 (eBook)

https://doi.org/10.1007/978-3-030-69921-5

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature
Switzerland AG 2021
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse
of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Contents

Sensor Image and Data Driven Societal Technologies

Office Monitoring and Surveillance System . . . . . . . . . . . . . . . . . . . . . . . . . 3
Vishal Patil and Yogesh Jadhav
Categorizing Documents by Support Vector Machine Trained
Using Self-Organizing Maps Clustering Approach . . . . . . . . . . . . . . . . . . . 13
Vishal Patil, Yogesh Jadhav, and Ajay Sirsat
Bandwidth Improvement of Multilayer Microstrip Patch
Antenna by Using Capacitive Feed Technique for Broadband
Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Anil K. Rathod, Md. M. Bhakar, M. S. Mathpati, S. R. Chougule,
and R. G. Sonkamble
Use of Median Timbre Features for Speaker Identification
of Whispering Sound . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Vijay M. Sardar, Manisha L. Jadhav, and Saurabh H. Deshmukh
Intelligent System for Engine Temperature Monitoring
and Airbag Deployment in Cars Using . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Akshay A. Jadhav, Swagat M. Karve, Sujit A. Inamdar,
and Nandkumar A. Admile
Analysis and Prediction of Temporomandibular Joint Disorder
Using Machine Learning Classification Algorithms . . . . . . . . . . . . . . . . . . 51
Roopa B. Kakkeri and D. S. Bormane
Machine Learning Approach in Cooperative Spectrum Sensing
for Cognitive Radio Network: Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Vaishali S. Kulkarni, Tanuja S. Dhope(Shendkar), Swagat Karve,
Pranav Chippalkatti, and Akshay Jadhav

v
vi Contents

Study on Detection of Leukemia in Human Blood Pattern Based

on Microscopic Photographs Using Image Processing Techniques . . . . . 73
Swagat M. Karve, Pravin Kishrsagar, Akshay A. Jadhav,
and M. Aravind Kumar
Brain Tumor Detection Using Deep Neural Network . . . . . . . . . . . . . . . . . 85
Rajshree B. More and Swati. A. Bhisikar
Design and Simulation of Different Structures of Micro Strip
Patch Antenna for Wireless Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Anil. J. Kokare, Mahesh. S. Mathpati, and Bhagyashri. S. Patil
State Context and Hierarchical Trust Management in WSN
for Intrusion Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Ranjeet B. Kagade and J. Santhosh
Portable Camera Based Assistive Text and Product Label
Reading from Hand Held Object by Using Android App
for Blind Person . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Somnath Thigale and Ranjeet B. Kagade
Automatic System for Identifying Cholesterol . . . . . . . . . . . . . . . . . . . . . . . 125
Mohua Biswas, Pragtee Tathe, Geeta Unhale, and Papiya Biswas Datta
Design of an IoT Based System for Monitoring and Controlling
the Sub-Station Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Pranali Bodke and A. A. Kalage
Implementation of Iridology for Pre Diagnosis of Brain Tumor . . . . . . . 141
Pragtee Tathe, Mohua Biswas, Anup Vibhute, Geeta Unhale,
Mrunmayi Raut, and Papiya Biswas Datta
Wireless Communication Using Light Fidelity Network . . . . . . . . . . . . . . 153
Nimisha Deval, Prajakta Satarkar, Akshata Jadhav, and Rupali M. Shinde
Smart Trolley with Automatic Master Follower and Billing
System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
R. Arpita, K. S. Shashidhara, and Veerendra Dakulagi
Analysis and Design of E Shaped Dual Band Microstrip Textile
Antenna for Wireless Communication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
Husain Bhaldar, Sanjay Kumar Gowre, Mahesh S. Mathpati,
Ashish A. Jadhav, and Mainaz S. Ustad
Enhancement Technique for Early Stage Lung Cancer Detection
Using Foldscope Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
Vanita D. Jadhav and Lalit V. Patil
Contents vii

Foldscope to Detect the Growth of Microorganisms on Various

Materials and Vessels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
Vanita D. Jadhav, Richa Tamhane, Kiran Kedar, Shruti Kawade,
and Aboli Gaikwad
Implementation of A* Algorithm for Real-Time Evacuation
During Fire Situation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
Shilpa K. Rudrawar, Pallavi Ghorpade, and Dipti Y. Sakhare
Involuntary Traffic Control System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
Shriniwas V. Darshane, Ranjeet B. Kagade, and Somnath B.Thigale
Early Detection of Diabetic Retinopathy Using Machine Learning . . . . . 215
Vishal V. Bandgar, Shardul Bewoor, Gopika A. Fattepurkar,
and Prasad B. Chaudhary
The Effective Use of Deep Learning Network with Software
Framework for Medical Healthcare . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
Padmanjali A Hagargi
Detection of Brain Tumor Using Image Processing and Neural
Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
Vanshika Dhillo, Dipti Sakhare, and Shilpa Rudrawar
Automatic Guided System for Image Resolution Enhancement . . . . . . . . 249
Neeta P. Kulkarni, J. S. Kulkarni, and S. M. Karve
Residual Network for Face Progression and Regression . . . . . . . . . . . . . . 257
Dipali Vasant Atkale, Meenakshi Mukund Pawar,
Shabdali Charudtta Deshpande, and Dhanashree Madhukar Yadav
Design and Simulation of 2-Element, Circular Shaped MIMO
Antenna for C-Band Frequencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
Ashish Jadhav, Nagashettappa Biradar, Husain Bhaldar,
Mahesh Mathpati, Manoj Deshmukh, and Renuka Wadekar
Attribute Inspection of Product Using Image Processing . . . . . . . . . . . . . 281
Anup S. Vibhute, Reshma R. Deshmukh, P. S. Valte, B. D. Gaikwad,
and Shrikant Pawar
A Proposed Method for Audio Steganography Using Digital
Information Security . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
Pratik Kurzekar and Shrinivas Darshane
Spine Diseases Detection Using SVM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
Jyoti M. Waykule and V. R. Udupi
BCI Integrated Wheelchair Controlled via Eye Blinks and Brain
Waves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
Sredha Prem, Jeswin Wilson, Shelby Mathew Varghese, and M. Pradeep
viii Contents

Classification of Soil Nutrients Using IoT . . . . . . . . . . . . . . . . . . . . . . . . . . . 333

Sandesh Koli, Dhaval Khobare, Amol Salunke, and Ranjeet B. Kagade
Non-invasive Methodological Techniques to Determine Health
of a Bone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
Meghana R. Khare and Raviraj H. Havaldar
Face Detection and Recognition Using Raspberry Pi . . . . . . . . . . . . . . . . . 351
P. R. Dolas, Pratiksha Ghogare, Apurva Kshirsagar, Vidya Khadke,
and Sanjana Bokefode
Human Tracking Mechanism for Institutions Using RFID
and Facial Recognition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
Rameez Shaik and L. V. Patil
Monitoring Power Consumption and Automation Using IOT . . . . . . . . . 365
Riya Jain, Revati Awale, Neha Kothari, Swarali Shah, and Amit Kore
Light Fidelity (Li-Fi): Future 5G Wireless Connectivity
to Empower Rural India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377
Prajakta A. Satarkar and Girish V. Chowdhary
Optimized Dynamic Feature Matching for Face Recognition . . . . . . . . . . 387
Ganesh Gopalrao Patil and Rohitash Kumar Banyal
Novel Secure Routing Protocol for Detecting and Presenting
Sybil Attack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393
S. M. Sawant, S. M. Shinde, and J. S. Shinde
IoT Model for Heart Disease Detection Using Machine Learning
(ML) Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
Madhuri Kerappa Gawali and C. Rambabu
Security Threats and Their Mitigations in IoT Devices . . . . . . . . . . . . . . . 411
Saurabh Gupta and N. Lingareddy
Age and Gender Classification Based on Deep Learning . . . . . . . . . . . . . . 425
Tejas Agarwal, Mira Andhale, Anand Khule, and Rushikesh Borse
Handling of Auxiliaries in Kannada Morphology . . . . . . . . . . . . . . . . . . . . 439
Bhuvaneshwari C. Melinamath
Dysarthria Detection Using Convolutional Neural Network . . . . . . . . . . . 449
Pratibha Dumane, Bilal Hungund, and Satishkumar Chavan
AR for Maintenance Training During COVID-19 Pandemic . . . . . . . . . . 459
Jyoti Pawar and Trupti Bansode
Face Recognition Based Attendance System Using Cv2 . . . . . . . . . . . . . . . 469
Vedant Khairnar and C. M. Khairnar
Contents ix

IoT Enabled Secured Card Less Ration Distribution System . . . . . . . . . . 477

Shilpa K. Rudrawar, Kuldeepak Phad, and Prajwal Durugkar
Voice Assisted Bots for Automobile Applications . . . . . . . . . . . . . . . . . . . . . 489
Shilpa K. Rudrawar, Nikhil Choudhar, and Ankit Meshram
Content-Based Image Retrieval Using Color Histogram and Bit
Pattern Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499
Nandkumar S. Admile, Akshay A. Jadhav, Swagat M. Karve,
and Anil A. Kasture
KNN and Linear SVM Based Object Classification Using Global
Feature of Image . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 509
Madhura M. Bhosale, Tanuja S. Dhope, and Akshay P. Velapure
MURA: Bone Fracture Segmentation Using a U-net Deep
Learning in X-ray Images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 519
Komal Ghoti, Ujjwal Baid, and Sanjay Talbar
Effective Usage of Oversampling with SMOTE for Performance
Improvement in Classification Over Twitter Data . . . . . . . . . . . . . . . . . . . . 533
Deepak Patil, Poonam Katyare, Parag Bhalchandra, and Aniket Muley
Multi-Classification of Breast Histopathological Image
Using Xception: Deep Learning with Depthwise Separable
Convolutions Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 539
Suvarna D. Pujari, Meenakshi M. Pawar, and Madhuri Wadekar
Dense Haze Removal Using Convolution Neural Network . . . . . . . . . . . . 547
Mayuri Dongare and Jyoti Kendule

ICT Based Societal Technologies

Diabetic Retinopathy Detection with Optimal Feature Selection:
An Algorithmic Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 559
S. Shafiulla Basha and Syed Jahangir Badashah
Students Perception and Satisfaction Towards ICT Enabled
Virtual Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 569
Moshina Rahamat and B. Lavanya
An Appointment Scheduler: A Solution to Control Covid-19
Spread . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 579
Apeksha M. Gopale
Bandobast Allocation and Attendance System . . . . . . . . . . . . . . . . . . . . . . . 587
Prashant S. Bhandare, Somnath A. Zambare, Amey Bhatlavande,
and Shamsundar Bhimade
Fire Fighting Un-manned Air Vehicle for Remote Areas . . . . . . . . . . . . . . 593
N. Shashank Bhat, K. S. Shashidhara, and Veerendra Dakulagi
x Contents

Human Age Classification and Estimation Based on Positional

Ternary Pattern Features Using Ann . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 603
Shamli V. Jagzap, Lalita A. Palange, Seema A. Atole, and Geeta G. Unhale
Object Recognition Using Fuzzy Classifier . . . . . . . . . . . . . . . . . . . . . . . . . . 613
Seema A. Atole, Shamli V. Jagzap, Lalita A. Palange, and Akshay A. Jadhav
An Effective Approach for Accuracy of Requirement
Traceability in DevOps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 623
Vinayak M. Sale, Somnath Thigale, B. C. Melinamath, and Siraj Shaikh
Clustering of Fruits Image Based on Color and Shape Using
K-Means Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 639
Vidya Maskar, Kanchan Chouhan, Prashant Bhandare, and Minal Pawar
Modern Education Using Augmented Reality . . . . . . . . . . . . . . . . . . . . . . . 651
Vishal V. Bandgar, Ajinkya A. Bahirat, Gopika A. Fattepurkar,
and Swapnil N. Patil
OSS Features Scope and Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 661
M. K. Jadhav and V. V. Khandagale
Text Summarization and Dimensionality Reduction Using
Ranking and Learning Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 667
Dipti Bartakke, Santosh Kumar, Aparna Junnarkar, and Somnath Thigale
Properties of Extended Binary Hamming [8, 4, 4] Code Using
MATLAB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 683
N. S. Darkunde, S. P. Basude, and M. S. Wavare
Identification of Fake News on Social Media: A New Challenge . . . . . . . 689
Dhanashree V. Patil, Supriya A. Shegdar, and Sanjivini S. Kadam
A Smart and Secure Helmet for Safe Riding . . . . . . . . . . . . . . . . . . . . . . . . 703
Ramesh Kagalkar and Basavaraj Hunshal
On Some Properties of Extended Binary Golay [24, 12, 8] Code
Using MATLAB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 713
N. S. Darkunde, S. P. Basude, and M. S. Wavare
Encoding Using the Binary Schubert Code [43, 7] Using
MATLAB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 725
M. S. Wavare, N. S. Darkunde, and S. P. Basude
Data Mining Techniques for Privacy Preservation in Social
Network Sites Using SVM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 733
Vishvas Kalunge and S. Deepika
Near Field Communication (NFC) Technology and Its
Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 745
R. D. Kulkarni
Contents xi

Performance Analysis of Convolution Neural Network

with Different Architecture for Lung Segmentation . . . . . . . . . . . . . . . . . . 753
Swati P. Pawar and S. N. Talbar
A Secure Data Sharing Platform Using Blockchain
and Fine-Grained Access . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 761
Shamsundar Bhimade, Prashant Bhandare, Amey Bhatlavande,
and Bharati Deokar
Efficient and Interactive Fuzzy Type Ahead Search in XML Data . . . . . 769
Laxman Dethe, Geeta Khare, Avdhut Bhise, and Somnath Zambare
IOT Based Interactive Motion Detection Security System Using
Raspberry Pi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 779
Geeta Khare, Subhash Pingale, Avdhut Bhise, and Sharad Kawale

Commercially Successful Rural and Agricultural Technologies

Cognitive Intelligence of Internet of Things in Precision
Agriculture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 789
Rahul Keru Patil and Suhas Shivlal Patil
Flooring: A Risk Factor for Fall-Related Injuries in Elderly
People Housing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 797
Unesha Fareq Rupanagudi
Analysis of Construction Readiness Parameters for Highway
Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 805
Harshvardhan R. Godbole and R. C. Charpe
Biogas Generation by Utilizing Agricultural Waste . . . . . . . . . . . . . . . . . . 819
Prathamesh Chaudhari and Shivangi Thakker
Effect of Change in the Resilient Modulus of Bituminous Mix
on the Design of Flexible Perpetual Pavement . . . . . . . . . . . . . . . . . . . . . . . 829
Saurabh Kulkarni and Mahadeo Ranadive
Analysis of Three-Stage Open Pan Heat Exchanger Working
on Dual Fuel for Jaggery Making . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 839
Abhijeet N. Kore and Sanjay S. Lakade
The Performance and Emission Analysis of Diesel Engine
with Sunflower Biodiesel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 847
Aniruddha Shivram Joshi and S. Ramesh

Deployable Environment or Healthcare Technologies

Experimental Analysis of Effect of Bio-lubricant Between
Tribological Systems of Piston Ring Under the Jatropha Oil . . . . . . . . . . 863
Mhetre Rahul Sanjay and L. B. Abhang
xii Contents

Garbage Monitoring and Collection System Using RFID

Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 873
Amol A. Kadam, AksahyAjadhav, Dhanraj P. Narsale,
Anil M. Kasture, S. M. Karve, and Manoj A. Deshmukh
A Survey on Mental Health Monitoring System Via Social
Media Data Using Deep Learning Framework . . . . . . . . . . . . . . . . . . . . . . . 879
Satyaki Banerjee and Nuzhat F. Shaikh
An Approach to Heart Disease Prediction with Machine
Learning Using Chatbot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 889
Chinmay Nanaware, Arnav Deshmukh, Nikhil Chougala, and Jaydeep Patil
Automated Early Detection of Diabetic Retinopathy . . . . . . . . . . . . . . . . . 897
Supriya Shegdar, Ameya Bhatlavande, Dhanashree Patil,
and Sanjivani Kadam
Product Lifecycle of Automobiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 905
V. K. Bupesh Raja, Ajay Shivsharan Reddy,
Suraj Ramesh Dhavanapalli, D. R. Sai Krishna Sanjay,
BH. Jashwanth Varma, and Puskaraj D Sonawwanay
Analysis of Heavy Metal Pollutants in the Sediments
from Coastal Sites of Al-Hodiedah Governorates, Yemen . . . . . . . . . . . . . 911
Majeed Hazzaa Nomaan, Dipak B. Panaskar, and Ranjitsinh S. Pawar
Development of Low Cost PCR Product Detection System
for Screening and Diagnosis of Infectious Diseases . . . . . . . . . . . . . . . . . . . 923
Patil Yogesh Navalsing, Suri Vinod Kumar, Suri Aseem Vinod,
Kar Harapriya, and Thakur Mansee Kapil
Determination of Viscous, Coulomb and Particle Damping
Response in SDOF by Forced Oscillation . . . . . . . . . . . . . . . . . . . . . . . . . . . 935
S. T. Bodare, S. D. Katekar, and Chetan Chaudhari
Assessment of Godavari River Water Quality of Nanded City,
Maharashtra, India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 947
P. R. Shaikh, Girish Deore, A. D. Pathare, D. V. Pathare, and R. S. Pawar
Groundwater Quality Assessment in an Around Thermal Power
Plants in Central India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 963
V. U. Deshmukh, D. B. Panaskar, P. R. Pujari, and R. S. Pawar

Micro, Nano Manufacturing, Fabrication and Related

Applications
XML Based Feature Extraction and Process Sequence Selection
in a CAPP System for Micro Parts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 973
G. Gogulraj and S. P. Leokumar
Contents xiii

To Study and Optimize the Effects of Process Parameters

on Taper Angle of Stainless Steel by Using Abrasive Water Jet
Machining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 983
Meghna K. Gawade and Vijaykumar S. Jatti
Effect of Process Parameters on Response Measures of Cartridge
Brass Material in Photo Chemical Machining . . . . . . . . . . . . . . . . . . . . . . . 995
Bandu. A. Kamble, Abhay Utpat, Nitin Misal, and B. P. Ronge
Thermal Performance of Two Phase Closed Thermosyphon
with Acetone as Working Fluid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1005
Shrikant V. Pawar and Abhimanyu K. Chandgude
Effect of Slip on Vortex Formation Near Two-Part Cylinder
with Same Sign Zeta Potential in a Plane Microchannel . . . . . . . . . . . . . . 1013
Souvik Pabi, Sumit Kumar Mehta, and Sukumar Pati
Optimization of WEDM Parameters During Machining of Ni-75
Using AHP-MOORA Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1023
S. A. Sonawane and S. S. Wangikar
Structural Analysis of Novel Mini Horizontal Axis Wind
Turbine (NMHAWT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1031
Pramod Magade, S. P. Chavan, Sushilkumar Magade, and Vikram Gaikwad
Fabrication of Tree Type Micro-Mixer with Circular Baffles
Using CO2 Laser Machining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1039
Sachin R. Gavali, Sandeep S. Wangikar, Avinash K. Parkhe,
and Prashant M. Pawar
Sensor Image and Data Driven Societal
Technologies
Office Monitoring and Surveillance
System

Vishal Patil and Yogesh Jadhav

Abstract Facial recognition is a biometric software category that mathematically

maps the facial features of a person and stores the data as a face-print. Using machine
learning algorithms, the software compares a live capture or digital image to the stored
face print to verify an individual’s identity and help automate authentication. Facial
recognition will increase protection, recognize unauthorized entry and keep a track
of visitors. ID passes are yesterday’s technology Our project’s main task is to identify
if the person is an employee or a visitor by using a face recognition system where in
security guards job is to watch over the process and stepping in only when the system
says that the person is not an employee or when they see something suspicious.

Keywords Face recognition · FaceNet · OpenCV · Office monitoring

1 Introduction

A facial recognition system is a technology capable of recognizing and validating

an individual from a digital picture or a video source. Facial recognition usually
functions in conjunction with captured facial features from a given picture using
faces in a database. It is also recognized as a biometric technology based on arti-
ficial intelligence, which recognizes the person by detecting patterns based on the
facial textures and shape of a person’s face. Even though a face recognition system
was initially implemented with the help of the computer, a wider implementation is
seen on mobile devices as well and also in different technologies such as robotics
wherein it is used as access control in the security system and can be different from
other biometrics such as fingerprinting or eye iris recognition systems. It has also

V. Patil (B)
Computer Engineering Department, K. J. Somaiya Institute of Engineering and Information
Technology, Mumbai, India
Y. Jadhav
Computer Science and Engineering Department, Amity School of Engineering and Technology,
Amity University, Mumbai, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 3

P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_1
4 V. Patil and Y. Jadhav

become popular recently as a marketing and commercial device. Many implementa-

tions include automated interaction between humans and computers, video analysis,
automatic indexing of images and video archives. In today’s application, facial recog-
nition can be approached in different ways. The primary component analysis (PCA),
Linear Discriminant Analysis (LDA), and deep learning are examples of such vari-
ables. PCA is a form of dimensional reduction sometimes used to reduce dimensions
on a fairly wide dataset. You do this by turning variables into smaller sections in large
datasets, while still using large datasets details. The PCA approach is very commonly
used as the principle is very straight forward and has a computer-efficient algorithm
as illustrated in following Fig. 1.
FaceNet is one of the implementations of facial recognition using deep learning.
This is a one-shot approach for learning with Euclidean space to measure each
face’s resemblance size. FaceNet is a fairly recent approach used with the Deep
Convolutional Network system in Google study in 2015. There were two separate
systems in the analysis, the network system for Zeiler and Fergus and the newest
network system for Inception. CNN in train practices Stochastic Gradient Descent
(SGD) implementing back propagation and AdaGrad standards. FaceNet provides
up to 99.63% accuracy from datasets that have labeled Faces in the Wild (LFW)and
95.12% accuracy on YouTube Faces Database. The benefit of using the FaceNet
approach is that this model requires only limited alignment to cut the face region
relatively tightly. FaceNet’s drawback is that the training cost is very costly because
it requires a CPU. Office Monitoring and Surveillance System is a face recognition-
based system for providing an efficient security solution for commercial buildings
and offices. The project does this by performing facial recognition of people entering
the premises and checking if the person is an employee of the company or a Visitor.
If the system detects the personas an employee, he will be granted access to the
company. If the system does not detect the person as an employee, they will be
provided with a welcome screen. The welcome screen will contain options whether
the person has reserved a meeting, and with which employee. The guest user will
choose the employee they want to meet and their video will be live-streamed to
the employee. If the employee recognizes the guest, they will be granted access;
otherwise, the guest will be directed to a nearby consultant. Along with this system
also does object detection and fire recognition. The system uses Google FaceNet,
OpenCV, SVC model to perform face recognition. The system contains a web-based
application that will be used by all the employees to get notified about the presence of
a visitor willing to visit them in the office premises. The same web-based application
will be used for storing the visitor details and monitoring them. The application is
developed using the Laravel framework. The proposed system will help to reduce the
physical procedures, time & efforts required for the same and will save the paperwork
as well; to provide security for offices; to ensure the safety of employees and also
maximize the efficiency of guards.
Office Monitoring and Surveillance System 5

Fig. 1 System flow

6 V. Patil and Y. Jadhav

2 Literature Survey

Wei Wu, Chuanchang Liu, Zhiyuan Su have presented the result of the analysis
and the methodology of applying FaceNet. Recent studies show that deep learning
approaches to this challenge can produce impressive results. To overcome these chal-
lenges, this program uses a profound learning approach. This comes with a module
to detect the faces and identification of the faces. The MTCNN facial detection
module is very simple, accurate and robust enough for variations in lighting and
contexts. This module is based on FaceNet, which teaches you to map the facial
images directly at the position of the Euclidean space, where the two-pointed space
inside the Euclidean space corresponds directly to the similarity of two facial images.
When this Euclidean space is generated, facial images are transferred as facial vectors
to the FaceNet embedding system. Feature vectors are then listed in a SVM to help us
easily recognize the face images. The experimental findings indicate that the device
is highly accurate, low computational complexity, and has a strong implementation
benefit for real time identification [1].
Many businesses have widely adopted video monitoring devices to increase
protection in their workforce. There are few visible cameras because of their suscep-
tibility to eye strain A kind of control method is used in this paper to solve the
problem. The network is made up of wireless and video monitoring devices. When
security threats are detected, an alert may be activated by the wireless control device
and the video stream from the camera to the computer is enabled. Therefore, the
system avoids dangers resulting from machine breakdown or user error and thereby
increases industrial efficiency [2, 3].
Facial recognition is an essential process to identify and separate the facial region
from the majority of the photos of every face detection device. There are numerous
techniques proposed, ranging from the simple detection of edges to the advancement
of techniques like pattern recognition. The review contrasts two facial recognition
approaches the characteristics of them and the local binary patterns focused on impact
rate recognition and speed detection. The algorithms have been tested on Microsoft
Visual C++ 2010 Express with OpenCV software. Test findings show the reliability
and trust of the functionality of the local binary template for implementing a facial
recognition device in real time [4].
Paper 4 introduces the presentation on Jetson TX2 with FaceNet and MTCNN
algorithms of an advanced and multi-camera face recognition surveillance device.
Presents the analysis of the modern, multi-camera facial recognition tracking system
FaceNet and MTCNN algorithms on Jetson TX2. The suggested portable system
tracks the target or suspect along with the ID/location of the camera and documents
its presence through various camera installations in the database. Triplet loss is one of
the most significant things we find in this article. The gap between the anchor and the
result, which implies a separate identity, is reduced using Triplet loss procedure, if the
result is accurate and indicates the same identity; the gap between the anchor and the
sample is then maximized. Triplet loss is one of the easiest ways to practice effective
128-dimensional embedding for any image. The representation of the anchor here is
Office Monitoring and Surveillance System 7

the frame from which the triplet loss was measured. Three pictures are required to
quantify the triplet loss: an anchor, positive and negative [4, 5].
Study of paper 5 describes a way of monitoring individuals and obtaining canon-
ical frontal pictures that fit the sensor network model in 2D world coordinates. In front
photos, monitoring and identity management have a specific desirability because they
are essentially invariant with daily appearance changes. This method has been applied
and tested on the FaceNet wired camera prototype network. Their key contribution
is to demonstrate how to sense the trajectories of moving objects to acquire canon-
ical vision of high quality while maintaining node capacity. The authors present an
overview of the method and display the task algorithm on data obtained from the
FaceNet camera network [6–11].
The purpose of project implemented in paper 6 is to create a complete hardware
and software framework for the surveillance of entry into a secure facility with
modern facial recognition technology (FRT) and lightweight, cheap components.
For visualization in the operation, FRT senses and recognizes individuals as they
enter and exit the area protected. The project uses FaceNet FRT to boost facial
recognition accuracy by means of a deep-learning approach. The FaceNet FRT is
implemented on a Raspberry Pi using the Neural Compute device Intel Movidius
with input from two portable cameras [7].
Substantial advancement in the field of identity testing and facial detection has
been made lately, taking advantage of advances in the convolution neural networks
(CNN). The findings on human level have been exceeded by FaceNet’s performance
particularly in terms of accuracy in “Labeled Faces in the wild (LFW)” datasets
and “Youtube Faces in the wild (YTF).” The triplet loss in the FaceNet has been
proven successful for facial verification. However, if a large dataset is used to train
the model, the amount of triples available is explosive. In this article, we propose a
simple class-by-class triple loss based on intra-class distance metric learning which
can substantially reduce the number of potential triplets to learn, but simplifying the
classic three-point loss function hasn’t deteriorated the approach. Experimental tests
of the most commonly used LFW and YTF benchmarks show that the model with
a class-wise simple triplet loss can achieve the latest results. And the simulation of
the dispersal of the learned features based on the MNIST data set has shown the
effectiveness of the proposed method of better class separation and discriminating
against the other state-of-the-art loss functionality [12–14].

3 Proposed System

Our system provides automated surveillance and monitoring for offices. This system
will work on employee’s faces for distinguishing a person from a visitor and an
employee of the organization using Face Recognition. The project subject matter
is a solution to orthodox methods of logging in employees in physical handwritten
logbooks, keeping security guards on duty 24 × 7, relying solely on sensors to fire
breakouts, etc. So, to make up for these fairly unreliable methods, we have come
8 V. Patil and Y. Jadhav

up with a plan that this project will increase the quality and efficiency of security in
office premises. The main objective of our project will be to recognize an employee
to let them in the premises and greet visitors respectfully keeping security as a
higher priority. The model is a deep convolutional neural network, which is learned
by means of a three-fold loss function, which induces vectors for one identity to
become more identical (smaller distance). A significant breakthrough in this research
was the emphasis on teaching a computer to establish embedding directly (instead
of removing them from an intermediate layer of a computer).
The system will be fed live image frames through a CCTV [15], which will be
used to detect faces coming up in the frame. For the detection of faces, we are using
OpenCV. Multiple faces can be detected at once. Once the faces are detected, the
faces will be cropped to size and by analyzing the eye orientation, the faces will be
aligned to a proper axis. Once Face Alignment is achieved, these images will be fed
to Google FaceNet. FaceNet will then extract the features from the image received.
Then, these face embedding systems are used as the basis for training classifier
systems on standard face recognition benchmark datasets, achieving state-of-the-
art performance. These preprocessed images will be used by the Support Vector
Classifier to train the model. This training will be done for each new employee that
joins the organization. Once the training phase is completed, the preprocessed data
received realtime, will be fed to the Support Vector Classifier Inference model which
will recognize the face as an employee or not. Changes (recent scars, beard/non-beard,
different lighting conditions) to faces are anticipated to expand its classification
performance as illustrated in Fig. 2.

4 Experimental Results

We require 40–50 facial data of each employee. Above images are used for training
the system. Accuracy achieved ranges from 95 to 100% as shown in Fig. 3.

5 Conclusion and Future Scope

We have seen various kinds of techniques in this paper that allow us to detect the
face and recognize it in various situations such as dark and light and even in various
clothing where the face is partially visible. All the techniques such as viola jones
used for object detection in OpenCV and haar cascade classifier which is also used
by the help of OpenCV. However, we observed that FaceNet provide significantly
better results than the previously mentioned techniques. Additional features in this
project can be further added after successful implementation, for example, visitors
who have been granted access in the office can be monitored for behavioral analysis;
sentiment analysis to check whether the visitor becomes hostile and creates nuisance,
etc. Furthermore, additional features in which employees can be monitored to keep a
Office Monitoring and Surveillance System 9

Fig. 2 Working of face recognition

Fig. 3 Face recognition

10 V. Patil and Y. Jadhav

check on the performance in work can be implemented. In restricted areas within the
company, only employees who are allowed access will be granted entry otherwise
no.

References

1. Wu W, Liu C, Su Z (2017) Novel real-time face recognition from video streams. In: Conference
on computer systems, electronics and control (ICCSEC), Dalian, pp 1149–1152. https://doi.
org/10.1109/ICCSEC.2017.8446960
2. Xiugang G, Yang L, Ruijin Z, Liang Y (2011) Study on monitoring systems with video capture
cards and wireless sensors. In: IEEE conference on electronics, communications and control
(ICECC), 03 November 2011
3. Saputra DIS, Amin KM (2016) Face detection and tracking using live video acquisition in
camera closed circuit television and webcam. In: 1st international conference on information
technology, information systems and electrical engineering (ICITISEE), Yogyakarta, pp 154–
157. https://doi.org/10.1109/ICITISEE.2016.7803065
4. Kadir K, Kamaruddin MK, Nasir H, Safie SI, Bakti ZAK (2014) A comparative study between
LBP and Haar-like features for face detection using OpenCV. In: 4th international conference
on engineering technology and technopreneuship (ICE2T), Kuala Lumpur, pp 335–339. https://
doi.org/10.1109/ICE2T.2014.7006273
5. Jose E, Greeshma M, Haridas MTP, Supriya MH (2019) Face recognition based surveillance
system using FaceNet and MTCNN on Jetson TX2. In: 5th international conference on advanced
computing & communication systems (ICACCS), Coimbatore, India, pp 608–613. https://doi.
org/10.1109/ICACCS.2019.8728466
6. Heath K, Guibas L (2007) Facenet: tracking people and acquiring canonical face images in a
wireless camera sensor network. In: First ACM/IEEE international conference on distributed
smart cameras, Vienna, pp 117–124. https://doi.org/10.1109/ICDSC.2007.4357514
7. Boka A, Morris B (2019) Person recognition for access logging. In: IEEE 9th annual computing
and communication workshop and conference (CCWC), Las Vegas, NV, USA, pp 0933–0936.
https://doi.org/10.1109/CCWC.2019.8666483
8. Chien Y-T, Huang Y-S, Jeng S-W, Tasi Y-H, Zhao H-X (2003) A real-time security surveillance
system for personal authentication. In: IEEE 37th annua international Carnahan conference on
security technology. Proceedings, Taipei, Taiwan, pp 190–195. https://doi.org/10.1109/CCST.
2003.1297558
9. Qezavati H, Majidi B, Manzuri MT (2019) Partially covered face detection in presence of
headscarf for surveillance applications. In: 4th international conference on pattern recognition
and image analysis (IPRIA), Tehran, Iran, pp 195–199. https://doi.org/10.1109/PRIA.2019.
8786004
10. Cao Y, Pranata S, Nishimura H (2011) Local binary pattern features for pedestrian detection
at night/dark environment. In: IEEE international conference on image processing, Brussels,
pp 2053–2056. https://doi.org/10.1109/ICIP.2011.6115883
11. William I, Ignatius Moses Setiadi DR, Rachmawanto EH, Santoso HA, Sari CA (2019) Face
recognition using FaceNet (Survey, Performance Test, and Comparison). In: Fourth interna-
tional conference on informatics and computing (ICIC), Semarang, Indonesia, pp 1–6. https://
doi.org/10.1109/ICIC47613.2019.8985786
12. Ming Z, Chazalon J, Luqman MM, Visani M, Burie J (2017) Simple triplet loss based
on intra/inter-class metric learning for face verification. In: IEEE international conference
on computer vision workshops (ICCVW), Venice, pp 1656–1664. https://doi.org/10.1109/
ICCVW.2017.194
13. Geng X, Zhou Z-H, Smith-Miles K (2008) Individual stable space: an approach to face
recognition under uncontrolled conditions. IEEE Trans Neural Networks 19(8):1354–1368
Office Monitoring and Surveillance System 11

14. Ibrahim R, Zin ZM (2011) Study of automated face recognition system for office door access
control application. In: IEEE 3rd international conference on communication software and
networks, Xi’an, pp 132–136. https://doi.org/10.1109/ICCSN.2011.6014865
15. Celine J, Sheeja Agustin A (2019) Face recognition in CCTV systems. In: International confer-
ence on smart systems and inventive technology (ICSSIT), Tirunelveli, India, pp 111–116.
https://doi.org/10.1109/ICSSIT46314.2019.8987961
Categorizing Documents by Support
Vector Machine Trained Using
Self-Organizing Maps Clustering
Approach

Vishal Patil, Yogesh Jadhav, and Ajay Sirsat

Abstract This paper mainly emphasis on the use of machine learning algorithms
such as self-organizing maps (SOM) and support vector machines (SVM) for clas-
sifying text documents. We have to classify documents effectively and accurately to
different classes based on their content. We tested classification of self-organizing
map on Reuters R-8 data set and compared the results to three other popular machine
learning algorithms: k-means clustering, k nearest neighbor searching, and Naive
Bayes classifier. Self-organizing map yielded the highest accuracies as an unsu-
pervised method. Furthermore, the accuracy of self-organizing maps was improved
when used together with support vector machines.

Keywords Document Categorization · Self-organizing maps (SOM) · Support

vector machines (SVM)

1 Introduction

It is very important to store information in a way that is easy for retrieval. This
becomes even more important in the case of electronic information. Every day we
search for documents on our computer is it at home and at work. With each passing
day the amount of information increases. The problem which we need to overcome is
manually storing all related information together and searching for that information.
This may take a lot of effort and time. The solution to this problem is to classify
and label documents. Once the documents are classified and ordered, we can go

V. Patil (B)
Computer Engineering Department, K. J. Somaiya Institute of Engineering and Information
Technology, Mumbai, India
Y. Jadhav
Computer Science and Engineering Department, Amity School of Engineering and Technology,
Amity University, Mumbai, India
A. Sirsat
Computer Engineering Department, St. John College of Engineering and Management, Mumbai,
India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 13

P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_2
14 V. Patil et al.

through the documents and search for information according to our needs. Consider
a scenario where documents are already ordered and. In such a scenario, we can
browse through documents effectively. This is where machine learning comes into
focus.
Machine learning algorithms can classify documents automatically making it
easy to store them. These algorithms can be supervised or unsupervised. Supervised
machine learning algorithms involve human interference. Since it may not be the case
that class labels are already known for each document, this may give rise to human
error. Also, many a times the document has no title or has an ambiguous title, which
makes it difficult to understand the category of the document. Thus we have focused
on unsupervised machine learning algorithms for the task of document classification.
We have used Kohonen’s maps or self-organizing maps to classify documents. Self-
organizing maps (SOMs) [1] is an effective data Visualization method which can be
used to reduce the dimensions of data by using self-organizing neural networks [2].
It is a very popular method for clustering [3] and classification of text documents [4,
5]. We compared self-organizing map with one unsupervised [6] and two supervised
machine learning algorithms.
Self-organizing maps gave better performance as compared to k nearest neighbor
and k-means clustering with accuracies as high as 90–92%. However, the draw-
back of SOM is Curse of Dimensionality (COD) because of which its efficiency is
affected. Because of this, SOM is not able to differentiate between classes that give
the same Euclidean distance even though they have different features. Hence, we
have implemented a technique which uses SOM and SVM. SOM takes the unlabeled
text documents and forms clusters for these documents which form the training data
set for the Support Vect4or Machine (SVM) [7].
This method has a two-fold advantage; it reduces the effect of COD while also
eliminating human interference needed to label the training documents. Support
Vector Machine have been one of the most widely used machine learning algorithm
in most of the text classification application given as input for many a times for large
datasets manually classifying and labeling training documents in the preprocessing
stage leads to an increase in time and cost complexity. Thus, the unsupervised clus-
tering methods are used which eliminate the need of human involvement for grouping
together set of documents. The output of these methods is then Training phase of
SVM. This hybrid model gives excellent accuracy while also eliminating the effect
of COD. Moreover, we can see that the accuracy of text classification increases after
using this hybrid approach.

2 Literature Survey

Support vector machine is one of the most widely used supervised machine learning
algorithms. It is one of the most commonly used methods and has found use in a
variety of applications due to its excellent accuracy [8] (Fig. 1).
Categorizing Documents by Support Vector Machine … 15

Fig. 1 The optimal

separating hyper plane

However, for some of the classification domains, it might not always be possible to
label the training data set of Support vector machine. For certain domains, labeling the
training data might be difficult when human expertise is unavailable or when it may
be time-consuming and costly to involve humans. Hence, in cases where it may not
be possible to label training data, unsupervised clustering methods are introduced
as they can perform their task without training the dataset. For domains wherein
training data is unavailable, we can use unsupervised machine learning methods,
such as SOM and k-Means to group documents based on its content.
A SOM is a single layer neural network. The name artificial neural network is
because of the fact that they were originally inspired by the way how biological
neurons were thought to work. SOM is made up of multiple units lying along a grid
of n-dimensions. In the input layer for every unit has the number of weights and
input patterns are same. Training a Self Organizing Map when provided with an
input pattern requires calculating the distance between that pattern and every other
unit in the neural network [9]. A unit that has the least distance to Best Matching
Unit (BMU) is then selected and that pattern is then mapped onto that unit. There
will be a one to one correspondence between the mapping of patterns that lie nearby
in the input space and the units lying nearby in the output space. This will be visible
in the visual map generated by SOM [9, 10]. Thus, SOM is topology preserving as
neighborhoods of a node are kept intact through the mapping process (Fig. 2).

3 Proposed System

The data set we have selected is a variant of the Reuters- 21,578 collection. This
dataset from the year 1987 includes 21,578 English Reuters news documents. More
than one class labels to some the documents Therefore; we have discarded documents
16 V. Patil et al.

Fig. 2 SOM architecture

with multiple class labels and selected only those documents which have only one
class label.
We then selected the 10 largest classes to obtain the collection of 8008 documents,
collection of 8008 documents, consisting of 2254 test documents and 5754 training do
over the past two decades, many research works focusing on document classification
by machine learning algorithms have been carried out. A large number of these papers
focus on improving the efficiency of SVM. Many researchers have proposed a hybrid
method which combines unsupervised learning algorithms such as SOM with SVM
[7]. Wu’s research shows how we can improve the accuracy of SVM [11] analyzing
zebra fish gene expression [12] by analyzing data in the training data set by using
SOM. SOM filters the training data for Support vector machines thereby reducing
the train dataset by keeping only those neurons in every category of SOM that are
their BMUs while it eliminates the remaining we then performed conventional text
preprocessing on the dataset as shown in Fig. 3.
1. The dataset we have used is the Reuters R8 dataset having 5485 training and
2254 test samples respectively.
2. For preprocessing, firstly words we retransformed to their stems and then, stop
words were removed.
3. Next, we calculated the frequencies of resultant words and computed document
vectors for all documents.
4. Finally, tf-idf weights were calculated for all resulting word stems.
5. For using a self-organizing map in the document classification task we have to
label the map nodes with class labels of the training data set.
6. The labelled nodes then represent the document classes. The map can then
classify new documents by mapping them.
SOM is primarily used for labeling and organizing the data samples which are not
labeled. As mentioned earlier in this paper, the SOM [2] is affected by COD because
it measures similarity using Euclidean distance [7]. Reducing the dimensions of
original data in their principle components might be of some help. However, this
reduction is unsuitable for some datasets, because every feature consists of some
information. Using this may result in loss of information if any one of the dimensions
Categorizing Documents by Support Vector Machine … 17

Fig. 3 Flowchart showing

Preprocessing steps
performed on the data set
18 V. Patil et al.

is removed [7]. The obtained labels are coupled with the related original data samples
for creating a machine labeled training set for the SVM which was machine organized.
The SOM begins the step of clustering the documents by deciding which important
features are needed for efficient clustering of the data into groups [7]. A process of
statistical cluster analysis is usually used to better the feature extraction which is
iterative in nature. The dimensions of data are reduced to two dimensions which
simplify viewing and interpretation is obtained using the feature maps [7].
Self-organizing maps generate a map which represents the dataset [7]. When
different neurons are fired for each data point a greater number of neurons will give
larger number of classes. Thus, to prevent this, the neurons are first grouped in small
clusters using Euclidean distance. Many neurons together make a cluster [7]. Every
cluster is given a class number. Using the Euclidean distance all the data points are
grouped together around these clusters. A number of a particular cluster is assigned
to data points with the closest proximity to that cluster [7].
The combination of SOM and SVM which results in a method which uses unsuper-
vised learning and has better generalization ability [7]. This is assumed from the prop-
erties of the SVM. SVM is known to be excellent classification methods with respect
to performance [7]. The generalization attribute of Support Vector Machine because
we implement structural risk minimization which selects an optimal separating hyper
plane which guarantee classification with very high accuracy (Fig. 4).

4 Results and Conclusion

First we compared different algorithms used for clustering based on their accuracy.
The results are as shown in the Table 1 and the graph in Fig. 6. We can see from the
Table 1 that the accuracy of SOM is 92.45% which is very good when compared to
methods such as k means and k nearest neighbor whose accuracies are 90.23% and
89.9% respectively. The accuracy of Naive Bayes is 95.24%. As seen in Fig. 6 the
accuracy of the hybrid model of SOM and SVM is 96.44% which is more than the
accuracy obtained from Naive Bayes (Fig. 5).

5 Conclusion

We tested Self Organizing Maps for classification of text documents and compared
it with some of the commonly used text classification methods. Naive Bayes had
the best accuracy overall, but in unsupervised algorithms, SOM turned out to have
the best accuracy for text classification. Sometimes, the knowledge and expertise to
label and organize documents into separate classes might be unavailable and thus
unsupervised methods viz. SOM proves to be useful. Self-organizing maps gave over
90% classification accuracy in the Reuters R-8 Data Set making it is effective for the
document classification. Also, the accuracy of self-organizing maps was comparable
Categorizing Documents by Support Vector Machine … 19

Fig. 4 Block diagram implemented hybrid classification algorithm from [7]

Table 1 Comparison of
Algorithm Accuracy (%)
clustering methods
k means 90.23
kNN 89.9
Naive Bayes 95.24
SOM 92.45

to some of the supervised machine learning methods that we tested. The training
phase of SOM generates an intuitive visual map, which is useful in a variety of
applications.
The learning procedure of this method though costly, is advantageous. Theo-
retically, the learning phase of SOM is more time consuming as compared to the
learning of k nearest neighbor. However, practically it is the exact opposite. The map
has significantly less nodes compared to the training dataset. This result to 10 time’s
quicker classification when compared to that obtained from k nearest neighbor using
20 V. Patil et al.

Fig. 5 Clusters formed after using SOM

Fig. 6 Comparison of various machine learning algorithms

the same data set. We don’t have to build a new map whenever we get a new input
document in the dataset, we can map it and perform the learning phase afterwards
when there is no new input data available. The results presented above show that
when SOM is used along with SVM the accuracy obtained is near to that which we
get when the data used for training is classified by supervised machine algorithms.
Categorizing Documents by Support Vector Machine … 21

References

1. Saarikoski J, Laurikkala J, Jrvelin K, Juhola M (2011) Self-organising maps in document clas-

sification: a comparison with six machine learning methods. Adaptive and Natural Computing
Algorithms, pp 260–269
2. Haykin S (1999) In neural networks: a comprehensive foundation, Second Edition. Prentice
Hall, Upper Saddle River
3. Mary Amala Bai V, Manimegalai D (2010) An analysis of document clustering algo-
rithms. In: International conference on communication control and computing technologies,
Ramanathapuram, pp 402–406
4. Ko Y, Seo J (2000) Automatic text categorization by unsupervised learning. In: Proceedings
of the 18th international conference on computational linguistics, (COLING2000), pp 453459
5. Indu M, Kavitha KV (2009) Review on text summarization evaluation methods. In: International
conference on research advances in integrated navigation systems (RAINS), Bangalore, 2016,
pp 1–4
6. Rui W, Liu J, Jia Y (2016) Unsupervised feature selection for text classification via word
embedding. In: ICBDA
7. Shafiabady N, Lee LH, Rajkumar R, Kallimani VP, Akram NA, Isa D (2016) Using unsu-
pervised clustering approach to train the support vector machine for text classification.
Neurocomputing 211:4–10
8. Brucher H, Knowlmayer G, Mittermayer MA (2002) Document classification methods for
organizing explicit knowledge. In: Proceedings of the 3rd European conference on organiza-
tional knowledge, learning and capabilities, (ECOKLC02), Institute of Information Systems,
University of Bern, Engehaldenstrasse Bern, Switzerland, Athens, Greece, pp 124–126
9. Lobo VJAS (2009) Application of self-organizing maps to the maritime environment. In:
Popovich VV, Claramunt C, Schrenk M, Korolenko KV (eds) Information fusion and
geographic information systems. Lecture Notes in Geoinformation and Cartography. Springer,
Berlin, Heidelberg
10. Sigogne A, Constant M (2009) Real-time unsupervised classification of web documents. In:
IMCSIT, pp 281–286. IEEE
11. Li TS, Huang CL (2009) Defect spatial pattern recognition using a hybrid SOMSVM approach
in semiconductor manufacturing. Expert Syst Appl 36(1):374–385
12. Wu W, Liu X, Xu M, Peng J, Setiono R (2004) A hybrid SOM-SVM method for analyzing
zebra fish gene expression. In: Proceedings of the 17th international conference on pattern
recognition (ICPR04) vol 2, pp 323–326
Bandwidth Improvement of Multilayer
Microstrip Patch Antenna by Using
Capacitive Feed Technique
for Broadband Applications

Anil K. Rathod, Md. M. Bhakar, M. S. Mathpati, S. R. Chougule,

and R. G. Sonkamble

Abstract In this research paper, Bandwidth improvement is investigated in E-shaped

Micro strip patch antennas by using multilayer and capacitive feed techniques. Capac-
itive feeding technique has been used to cancel the inductive impedance of probe by
capacitive patch. The Rectangular capacitive patch has been used separately near the
radiating E-shaped design model. Performance analysis of E-shaped conventional
and suspended capacitive feed Micro Strip Patch Antenna is done to enhance its
parameters like Bandwidth, Gain, and Directivity. The proposed E-shaped multi-
layer capacitive feed Micro strip patch Antenna is shown improved bandwidth of
238 MHz and Gain of 6.071 dB as compared with conventional candidates. The
proposed antenna design models have a working center frequency in the range of
2.36 GHz to 2.59 GHz, which can be used for wireless applications. The Return
Loss and VSWR have been investigated in an acceptable range. E-field and Current
flow of the antenna are within the desired radiating patch. It has observed that the
capacitive feed technique E-shaped design model improves the Bandwidth, Gain,
and other antenna parameters over the conventional probe feed antenna.

Keywords E-shaped · Bandwidth · Gain · Directivity · Suspended techniques ·

Capacitive feed technique

A. K. Rathod (B) · Md. M. Bhakar

Department of Electronics & Communication Engineering, G.N.D. Engineering College, Bidar,
India
M. S. Mathpati
Department of Electronics & Telecommunication, SVERI’s COE, Pandharpur, India
S. R. Chougule
Department of Electronics & Telecommunication, KIT’s COE, Kolhapur, India
R. G. Sonkamble
Department of Computer Science and Engineering, SGU, Kolhapur, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 23

P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_3
24 A. K. Rathod et al.

1 Introduction

Nowadays, antennas are an essential part of any wireless communication system.

With technological advancements, antennas are playing a very vital role in commu-
nication [1, 2]. Micro strip antennas (MSAs) have been reported that they have
lightweight, small-volume several advantages and that they can be made conformal
to any type of host surface. Besides, PCB Technology can be used the physical
manufacturing of Micro strip patch Antenna so that a large amount of production
can be done at a small cost [1]. Defense and other commercial applications of Micro
Strip Patch Antenna are alternating choices to replace the conventional Antenna.
However, the MSA has been reported that they are suffering from Narrow Band-
width. Increasing the Bandwidth of the MSA has become a primary requirement of
research in the field. This is reflected in a large number of papers on the subject
published in journals and conference proceedings [1]. However, different types of
Broadband MSA design models have been investigated in the last few decades. The
coaxial or probe feed techniques can be used in many design methodologies. To
impedance matching, we have to do the soldering of the Patch and center conductor
of the coaxial connector. We can place at any desired location of feed point to match
input impedance; this is the advantage of the probe feeding method [1, 3]. A new
-shaped micro strip antenna is reported, which increased the bandwidth of the
antenna by converting rectangular to E-shape and -shaped micro strip antenna [4].
The stacking of two substrates and slotted patches were used to get the wide band-
width is investigated [5]. Defected Ground Structure (DGS) is a novel technique
used to modify the ground plane to improve the gain and bandwidth is reported
[6]. Multilayer stretchable conductors (SCs) are demonstrated for the first time as
an alternative cost-effective method for implementing conformal and reconfigurable
antennas over flexible substrates [7]. Some of Researchers have been designed patch
antenna on single, double, and multilayer dielectric substrate introduced by microma-
chining process [8]. A miniature patch antenna is proposed for satellite applications
in (C-band, X-band, and Ku-band). The designed multiband antenna is sensible for
satellite applications [9]. A fractal antenna is designed at 2.4 GHz frequency band
which is suitable for Wireless Local Area (WLAN) applications and FR-4 dielec-
tric substrate material is used [10]. To improve the bandwidth and gain of Micro
Strip patch antenna deferent methodologies investigated and anyone technique can
be adopted.

2 Proposed Antenna Design Model

Figure 1 shows the proposed E-shaped mulilayer MSA with capacitive feed. The
design Model consists of a copper ground plane, radiating Patch, air gap, and FR-4
as a dielectric substrate. The Capacitive probe feed technique has been used for the
impedance matching with the radiating patch.
Bandwidth Improvement of Multilayer Microstrip … 25

Fig. 1 E-shaped suspended MSA with capacitive feed patch

The air gap acts as a dielectric substrate having a dielectric constant is equal to
one. Here the proposed model has become Multilayer consisting of a copper ground
plane, air gap with a height of 6 mm, a dielectric substrate with a height of 1.6 mm
easily available in the market, and an E-shaped radiating patch [11]. The permittivity
can be selected near to 1 to obtain a wide bandwidth. If the dielectric substrate has
larger thickness then it gives rise to an increase in probe reactance for the excitation
of surface waves and coaxial feed, which reduces the efficiency of the antenna so that
dielectric height should be selected properly. This design model is used to investigate
the antenna parameters like Operating Frequency, Bandwidth, Return Loss, VSWR,
Directivity, Total Gain, Radiation pattern, and E-field.

3 Simulation Results of the Proposed Design Model

3.1 Bandwidth and Return Loss of E-shaped Suspended MSA

Figure 2 represents the Bandwidth and Return loss of E-shaped Suspended MSA
which gives the bandwidth of 238 MHz and Return Loss −12.98. The return loss of
Antenna is in acceptable range.
26 A. K. Rathod et al.

Fig. 2 Bandwidth and return loss of E-shaped suspended MSA

Fig. 3 VSWR of E-shaped suspended MSA

3.2 VSWR of E-shaped Suspended MSA

Figure 3 represents the VSWR of E-shaped Suspended MSA. It shows the VSWR
1.5778 must be less than 2 as per the standards. The given antenna can be radiate
and fabrication process can be done.

3.3 Directivity of E-shaped Suspended MSA

Figure 4 represents the Directivity of E-shaped Suspended MSA is 6.071 dB.

Bandwidth Improvement of Multilayer Microstrip … 27

Fig. 4 Directivity of E-shaped suspended MSA

3.4 Gain of E-shaped Suspended MSA

Figure 5 represents the Gain of E-shaped Suspended MSA is 6.0196 dB, which is
better than the conventional Antena.

3.5 Radiation Pattern of E-shaped Suspended MSA

See Fig. 6.

3.6 E-field of E-shaped Suspended MSA

Figure 7 represnts E-field of E-shaped Suspended MSA.

4 Result Analysis Table of E-shaped Suspended MSA

Table 1 consists of result analysis table for Hexagonal MSA & Suspended Hexagonal
MSA with performance Parameters like Operating Frequency in GHz, Bandwidth in
28 A. K. Rathod et al.

Fig. 5 Gain of E-shaped suspended MSA

Fig. 6 Radiation pattern of E-shaped suspended MSA

Bandwidth Improvement of Multilayer Microstrip … 29

Fig. 7 E-field of E-shaped suspended MSA

Table 1 Result analysis table

Parameters E-shaped suspended MSA
for E-shaped suspended MSA
Operating frequency (GHz) 2.361 − 2.599
Bandwidth (MHz) 238
Return loss(dB) −12.989
VSWR 1.577
Directivity (dB) 6.071
Total gain (dB) 6.019

MHz, Return Loss in dB, VSWR, Directivity in dB and Total Gain in dB as shown
in Table 1.
The Bandwidth can be defined as a reappearance extends over which VSWR is
always under two (which relates to an arrival loss of 9.5 dB or 11% reflected force).
The VSWR or impedance BW of the MSA is characterized as the recurrence extend
over which it is coordinated with that of the feed line inside determined cutoff points.
The Bandwidth of proposed E-shaped multilayer MSA is 238 MHz, which is
within the specified limits as shown in Fig. 2. Return Loss has been observed in an
acceptable range. VSWR can be defined in terms of the input reflection coefficient
(G) as:

1 + ||
V SW R = (1)
1 − ||

VSWR of E-shaped multilayer MSA is less than 2, which is shown in Fig. 3. As

shown in Fig. 4 for Directivity and Fig. 5 for Gain of E-shaped suspended MSA.
Figure 6 shows the radiation pattern for the proposed antenna. E-field of E-shaped
multilayer MSA is shown in Fig. 7. All the proposed antenna parameters have been
observed in the acceptable range.
30 A. K. Rathod et al.

5 Conclusion

In this paper, Bandwidth improvement is investigated in E-shaped Micro strip patch

antennas by using multilayer and capacitive feed techniques. The design model of
E-shaped multilayer MSA is operating at 2.45 GHz center frequency. The proposed
E-shaped multilayer Micro strip patch Antenna is shown improved bandwidth of
238 MHz and gains 6.019 dB, as compared with normal Antenna. The return loss
of the proposed designed model is −12.989 dB which is less than −10 dB. The
Voltage Standing Wave Ratio (VSWR) has been observed that 1.577, which is in the
acceptable range. FR-4 dielectric substrate material is used to model design having
relative permittivity (A) of 4.4. It has less cost compare to other materials. The
proposed E-shaped multilayer Micro strip patch Antenna can be considered a better
candidate for Broadband Applications.

Reference

1. Kumar G, Ray KP (2003) Broadband micro strip antennas. Artech House Boston, London
2. Al-Halawani MG, Al-Najjar M, Abdelazeez MK (2016) Microstrip antenna design for UWB
applications. IEEE. 978-1-5090-2886-3/16/$31.00
3. Faouri YS, Awad NM, Abdelazeez MK (2019) Hexagonal patch antenna with triple band
rejections. In: Proceedings on 2019 IEEE Jordan International Joint Conference on Electrical
Engineering Information Technology JEEIT 2019, pp 446–448
4. Deshmukh AA, Ray KP (2013) Analysis of broadband Psi ()-shaped micro strip antennas.
IEEE Antennas Propag Mag 55(2)
5. Tiwari K, Pujara D (2015) Multilayer slotted microstrip antenna for Wi-Fi application. IEEE.
978-1-4799-7815-1/15/$31.00
6. Kachare AB, Mathpati MS, Rathod AK (2015) Effect of size and location of the DGS on char-
acteristics of rectangular micro strip patch antenna. Int J Comput Appl (0975–8887), Confer-
ence on Emerging Applications of Electronics System, Signal Processing and Computing
Technologies (NCESC 2015)
7. Liyakath RA, Takshi A, Mumcu G (2013) Multilayer stretchable conductors on polymer
substrates for conformal and reconfigurable antennas. IEEE Antennas Wireless Propag Lett 12
8. Saha R, Maity S, Trigunayat N (2015) Enhancement of gain, bandwidth and directivity of a
patch antenna by increasing dielectric layers of the substrate through micromachining technique
for RFID application. In: 2015 international conference on advances in computer engineering
and applications (ICACEA), IMS Engineering College, Ghaziabad, India, IEEE. 978-1-4673-
6911-4/15/$31.00
9. Nageswara Rao L, Santhosh Kumar V (2020) A novel multiband micro strip patch antenna for
satellite applications. Int J Adv Sci Technol (IJAST) 29(3s):177–183. ISSN: 2005-4238
10. Ali KM, Nataraj C et al (2020) Bandwidth analysis of microstrip antenna with fractal design
for 2.4 GHZ wireless applications. Int J Adv Sci Technol (IJAST) 29(1):1344–1358. ISSN:
2005–4238
11. Hong S, Kim Y, Won Jung C (2017) Transparent microstrip patch antennas with multilayer
and metal-mesh films. IEEE Antennas Wireless Propag Lett 16(c):772–775
Use of Median Timbre Features
for Speaker Identification of Whispering
Sound

Vijay M. Sardar, Manisha L. Jadhav, and Saurabh H. Deshmukh

Abstract Identifying speaker from the whispered voice is difficult task contrasted
to neutral as voiced phonations are absent in the whisper. The accomplishment of
the speaker identification system for the most part relies on the selection of proper
audio features reasonable for the type of database and type of application. This paper
examines the various audio features available and emphasizes on the use of selected
timbrel features which are sorted by Hybrid Selection Algorithm. The limited number
of timbrel features namely MFCC, Roll-off, Brightness, Roughness, and irregularity
which are found outperforming when tested on CHAIN database. Likewise, the
possibility of using the MEDIAN based features is investigated by analysis. The
use of Median timbrel features reported an enhancement in speaker identification
accuracy by 2.4% compared to timbrel features only in whisper train-whisper test
scenario.

Keywords Whisper · Speaker identification · Timbrel feature · K-NN classifier ·

Median

1 Introduction

Speaker identification requires the main processes as (i) Pre-processing, (ii) Feature
extraction (iii) Training and (iv) Testing. Pre-processing may involve noise reduc-
tion, pre-emphasis, silence removal and equalization so forth. Audio features are
the speaker-specific information represented in a compact form. They will make it
possible to compare the speakers for identification purpose with minimum processing
speed and data. In the training phase, all extracted audio features are utilized for
modelling the speakers which will be used as the templates in the testing phase.

V. M. Sardar (B)
Jayawantrao Sawant College of Engineering, Pune, MS, India
M. L. Jadhav
MET’s Institute of Engineering, Nasik, MS, India
S. H. Deshmukh
Maharashtra Institute Of Technology, Aurangabad, MS, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 31

P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_4
32 V. M. Sardar et al.

While testing phase, features of query speaker is compared for the closest speakers
among the database. There are two types of variants in the speaker identification
system, in particular, closed-set and open set. In the open set scenario, the unknown
speaker not available in the database may also be used, but that task will be speaker
verification [1]. The speaker verification is a binary decision to whether a claimed
speaker is the same or not by comparing with the single audio template. While in
an identification task, every speaker is compared with every other speaker in the
database [2].
Identifying the person from the whispered voice is a challenging task due to
its hidden and multi-dimensional attributes. Unmistakable separation ability of the
rich phonation is misplaced from a whisper [3]. The ADs applicable for the neutral
database is not found effective for the whispered database. Hence, the adapted
approach as shown in Fig. 1 where a selection of the proper ADs is carried out
in the beginning only. This process of selecting the appropriate ADs reasonable for
the type of database and application will enhance the identification rate.
A majority of low-level audio features are discussed in the literature. It includes
primary features such as the zero-crossing rate, speech bandwidth, the spectral
centroid, and energy [4, 5]. A second high level and complex feature used for speech
and speaker recognition are like Mel-frequency cepstral coefficients (MFCC), roll-
off and brightness etc. Such features use the parametric analysis of the spectral
envelope [6]. It is expected that the features used for the classification need to be
more uncorrelated. Typical classification results using better non-correlated MFCC
features found correct up to 95% for music and speech classification which is a rela-
tively less complex task [7]. Besides, these results have reported for clean speech
scenario. For other databases including noise, music background, and telephone
speech etc. the performance degrades considerably [8, 9]. The block diagram of
speaker identification system is as shown in Fig. 1.
This paper is organized as follows. We portrayed the classification of audio
features, timbrel audio descriptors, MIR toolbox, and the impact of timbrel feature

Fig. 1 Block diagram of the speaker identification system

Use of Median Timbre Features for Speaker Identification … 33

selection on the identification accuracy in Sect. 2. The K-nearest neighbor classifier

with the most suitable variants for the whispered speaker database is described in
Sect. 3. The performance of the system with timbrel features and K-NN classifier
are presented in the result section (Sect. 4). The possibility of improvement based
on the MEDIAN of feature values is discussed in Sect. 5.

2 Audio Descriptors

2.1 Classification of Audio Descriptors

We can classify the audio features comprehensively in two classes as follows. The
global descriptor is the class of feature where computation is done on the complete
signal as a whole. E.g. attack-time of a sound is known from the complete duration
of an audio signal. Instantaneous descriptors is another class which computes on a
short period (say 40 ms) of audio signal called a frame. The spectral centroid of a
signal can vary with the time, hence it will be termed as an instantaneous descriptor.
As an instantaneous descriptor produces multiple values for the number of frames,
use of statistical operations (like the mean or median, the standard deviation, and or
inter-quartile range etc.) is essential to derive a single value representation. In the
CUIDADO project [10], a listing of 166 audio features is provided.
Depending upon the type of process used for extraction of the feature, we can
further differentiate:
• Features that are directly extracted from the time domain waveform of audio signal
like the zero-crossing rate.
• A transform like FFT, wavelet etc. is applied on a signal to extract the features.
E.g. MFCC.
• Feature extracted based on a signal model like the source or filter model.
• Features which converges to the human auditory response (response on bark or
ERB scale).

2.2 Timbrel Audio Descriptors

Timbre is the perceptual and multi-dimensional feature of sound. As the exact defi-
nition of timbre is very difficult, it can be analyzed by following the attributes
[11].
1. Harmonic analysis: The number, relative strengths, structure of harmonics.
2. Partial analysis: Phase, inharmonic partials, the content of partials.
3. Time-related parameter like rise-time.
4. Steady-state and attack slices.
34 V. M. Sardar et al.

2.3 Timbrel Audio Descriptors MIR Toolbox for MATLAB

Musical Information Retrieval (MIR) toolbox is for the most part intended to enable
the study of the relation between musical attributes and music-tempted sensation.
MIR toolbox uses a modular outline. It is well known that the common algorithms are
used in audio processing like segmentation, filtering, framing etc. with an addition of
one or more distinguished algorithms at some stage of processing. These algorithms
are available in a modular form and the individual blocks can be integrated to capture
some features [12].
The philosophy to integrate the appropriate modules is proposed in Fig. 2. For
example, to measure irregularity and brightness, we need the implement the algo-
rithms like reading audio samples, segmentation, filtering, and framing as the
common processes between them. In the final stage, due to inherent differences, irreg-
ularity needs peaking algorithm and brightness needs spectrum analysis (Fig. 2).
Even, the integration of different stages depends upon parameter variations. E.g.
mirregularity (…, ‘Jensen’), where the adjoining partials are taken into considera-
tion and mirregularity (…, ‘Krimphoff’) which considers the mean of the preceding,
same and next amplitude [13]. The flow diagram of algorithm modules and their
integration to extract the selected timbrel features are shown in Fig. 2.
Roll-off frequency: Roll-off is assessed from the foremost energy (85% or 95% as
a standard) contained below the predefined frequency.
Roughness: It estimates the average disagreement between all peaks of the signal.
It is also an indicator of the presence of harmonics generally higher than the 6th
harmonic.
Brightness: It is the measure of the percentage of energy spread above some cut-off
frequency.

Fig. 2 The logical flow of timbrel feature implementation in MIR

Use of Median Timbre Features for Speaker Identification … 35

Irregularity: It may be calculated as the sum of the square of the difference in

amplitude between adjoining partials or the sum of the amplitude minus the mean of
the past, the same component and subsequent amplitude.
Miraudio: This command loads the appropriate format of an audio file. E.g.
miraudio (‘speaker.wav’).
Mirsegment: This process splits a continuous audio signal into homogeneous
segments.
Mirfilterbank: A set of filters are required which are useful to select neighboring
narrow sub-bands that cover the entire frequency range. The effect like aliasing in the
reconstruction process is avoided e.g. mirfilterbank (…, ‘Gammatone’) processes a
Gammatone filterbank decomposition. The frame decomposition can be performed
using the mirframe command. The frames can be specified as follows: mirframe
(x,…, ‘Length’, w, wu).
mirspectrum: Discrete Fourier Transform decomposes the energy of a signal (be
it an audio waveform, or an envelope, etc.) along with the frequencies.
Mathematically, for an audio signal x;


N −1
−2ikn
Xk = xne N k = 0, . . . , N − 1 (1)
n=0

This decomposition is performed using a Fast Fourier Transform by the mirspec-

trum function.
Mirpeaks: Many features like irregularity require the Peaks analysis. Peaks are
calculated from any data x produced in MIR toolbox using the command mirpeaks
(x).

2.4 Timbrel Features for Better Speaker Identification

For better speaker identification, the accompanying conditions should be satisfied:

(i) The least variation in the inter-speaker feature (ii) Maximum discrimination in an
inter-speaker feature. We have selected a limited well-performing feature (MFCC,
Roll-off, Roughness, Brightness, and Irregularity) from MIR toolbox by using Hybrid
Selection Algorithm. It uses the iterative process of testing for the performance of
audio features. The accuracy of the system is tested with the independent features
and then by successively appending the new feature in conjunction with the previous
combination. To conclude, it is proven that a combined vector having timbrel features
namely MFCC, Roll-off, Brightness, Roughness and, Irregularity is found to be the
best [14]. An association among the intra-speaker samples and dissociation among
the inter-speaker samples are confirmed by correlation analysis. It predicts that the
selected timbrel features are well-performing for speaker identification in a whispered
voice. It is validated by the identification experiments in subsequent sections.
36 V. M. Sardar et al.

3 System Description

The system block diagram is presented in Fig. 3 which is described in succeeding

sub-sections.

3.1 Speaker Database

This undertaking makes use of the CHAIN database developed at the School of
Computer Science and Informatics University College Dublin [15]. It consists of
a total of 36 speakers with 33 samples each, with a good mix of male and female
voice samples (20 males and 16 females). The speech samples of 2–3 s duration are
recorded at a sampling frequency of 44.1 kHz. The phrases/sentences are selected
from CSLU and TIMIT database which ensures the phonetic balance in the corpus.

3.2 Hybrid Selection Algorithm

Hybrid selection is the iterative process which starts with the targeted timbre class
of ADs and further investigates the AD which maximizes the identification result
[16]. This algorithm also used for liver tissue pathologicssal image classification in
[17]. After every iteration, the feature which maximizes the classifier accuracy is
appended in combination with the previous feature/s. The process continues until no
further increase in accuracy is observed. The algorithm proved that MFCC, Roll-off,
Roughness, Brightness, and Irregularity are found to be best suitable.

Feature
Hybrid
Database

Extraction Classifier
Selection

Speaker
ID

Fig. 3 System block diagram

Use of Median Timbre Features for Speaker Identification … 37

3.3 K-Nearest Neighbor (K-NN)

K-NN is a simple and non-parametric algorithm which separates the data points
into several classes and a new sample point is used to anticipate the class. The non-
parametric technique of classification does not make any assumptions on the distri-
bution of data. This methodology is helpful as the greater part of the real world data
which does not obey the generally assumed pattern (e.g. linear regression models).
While managing with this classifier, the following parameters are used: the number
of nearest neighbors (k), a distance function (d), decision rule and n labelled samples
of audio files Xn . A class label to the unknown sample is identified based on the
minimum distance from the training classes. In another word, K-NN calculates a
posteriori class probabilities P (wi|x) as

ki
P( wi |x) = .P(wi ) (2)
k

where ki is the number of vectors which belongs to class within the subset of k vectors
[18]. For our system, all the variants of KNN classifier are verified to maximize the
identification accuracy. The variations tested for the number of nearest neighbor
are 1-NN, 2-NN, and 3-NN. Two distance functions Euclidean and City-block are
investigated. The rules namely nearest and consensus are also tested. After a variety
of experiments, it is concluded that a combination of 3-NN neighbors, City-block
distance and Nearest rule give the maximum identification accuracy in [19].

4 Median Features for Improvement in System

Performance

The intra-speaker variability is observed from the observations of the extracted

feature values. The extent of variations is analyzed for the selected features in [15]
using Standard Deviation (σ). However, incorporating standard deviations in the
feature modifications is difficult. As the value of standard deviation is to be either
added or subtracted from the feature value (i.e. feature value ±σ). It may need a
complex decision algorithm to decide the addition or subtraction operation for every
feature value of every speaker sample. Besides, modifying the feature values with σ
may exceed the normalization range (i.e. all distinct feature values are normalized in
0 to 1). The values of selected features (Roll-off, Brightness, Roughness and, irregu-
larity) given in Table 1. Till now, we have used the absolute values of all these features
in a vector form for the conducted experiments above. Here, we have analyzed all
these feature values for the MEDIAN value (Table 2). The median of a finite list of
numbers can be found by arranging all the numbers (elements) in ascending order.
If there are an odd number of elements, the middle one is picked. If there is an even
number of observations, then there is no single middle value; the median is then
38 V. M. Sardar et al.

Table 1 Feature values with their median

Features speaker sample Roll-off Roughness Brightness Irregularity
1_1 0.121857 0.322853 0.004915 0.083185
1_2 0.458324 0.486716 0.069048 0.127183
1_3 0.523269 0.664555 0.072613 0.459385
1_4 0.46465 0.369261 0.023994 0.296143
1_5 0.72602 0.76634 0.048737 0.402269
1_6 0.339668 0.372256 0.027058 0.465064
1_7 0.424056 0.428354 0.023742 0.351125
1_8 0.452389 0.518894 0.042518 0.465195
1_9 0.206797 0.367267 0.041907 0.394578
1_10 0.476325 0.357569 0.108497 0.199835
MEDIAN 0.455356 0.400305 0.042212 0.372852

Table 2 Intra-speaker distance absolute values and the median values

City block distance in inter-speaker sample
Speaker sample 1_1 1_2 1_3 1_4 1_5
1_1 0 0.944 (0.544) 1.321 (0.467) 0.860 (0.220) 1.762 (0.871)
1_2 – – 0.779 0.434 0.859
1_3 – – – 0.670 0.603
1_4 – – – – 0.902
1_5 – – – – 0
Values in bracket shows the distance from MEDIAN
– (empty spaces) are due to repeated distance values between combinations

usually defined to be the MEAN of the two middle values. So MEDIAN may be
considered as the fully sheared mid-range.
MEDIAN can be calculated by the following formula which will be generalized
for odd and even number of values in the set:
   
a# 2x + a# 2x /0.5
MEDIAN = (3)
2
Comparing the feature values of query (test) speaker with the MEDIAN values
of the particular assigned class (at the time of training), distance from MEDIAN
is recommended. As the intra-speaker features are distributed in some range of
its variability, the median as a reference of class value will certainly improve the
identification accuracy. The following analysis will confirm our conclusion.
In the above illustration, five samples (1_1, 1_2, 1_3, 1_4, and 1_5) of the same
speaker (speaker_1) are listed for the intra-speaker City-block distance. Then in
another approach, the distances of the speakers from the MEDIAN of a class are
Use of Median Timbre Features for Speaker Identification … 39

Table 3 Comparative identification accuracy by using MFCC, timbre only and timbre (Median)
Speech mode % Accuracy
Training–Testing MFCC Timbre Timbre (Median)
Neutral–Neutral 91.5 95.0 95.6
Whisper–Whisper 78.8 86.5 88.1
Neutral–Whisper 66.0 73.0 76.4
CHAIN database of 35 speakers, MFCC only/timbre feature/timbre (median), and K-NN
Classifier

calculated (shown in bracket). It is seen that the distances of the inter-speaker samples
are less when calculated concerning MEDIAN as compared the absolute values. It
means that the possibility of identification of all samples of speaker_1 increases as the
intra-speaker distances are reduced when calculated from the reference of MEDIAN
values.

5 Results

This section presents the results with the following parameters. Total 35 speakers
with whispered samples from a CHAIN database are used for the experiments. The
samples of each speaker are selected with a choice of 70% samples for training
and 30% for testing. KNN classifier settings are as—Rule: nearest, Neighbor: 3,
and distance-Metric: City-Block distance. The selected Timbre features by Hybrid
Selection Algorithm are Roll-off, Brightness, Roughness and, irregularity.
Mel Frequency Cepstral Coefficient (MFCC) is widely used features for speaker
identification task. Hence, the performance using MFCC features is compared with
timber features and the Median values of it (Table 3).
The above results in Fig. 4 are compared with a baseline speaker identification
system for whispered speech using the CHAIN database [20]. Further, an enhance-
ment of 2.75% in the identification accuracy system in whisper train-whisper test
scenario is reported in [21] are reproduced in Table 4.
The proposed timbre features using Median value shows that the identification
accuracy is increased by 2.4% in whisper train-whisper test scenario which is the
focus of this undertaking.

6 Conclusion

A large number and variety of audio descriptors are available which are selected
according to the application. As the attributes of whisper are drastically change
compared to the neutral voice, the perceptually motivated, multidimensional timbre
40 V. M. Sardar et al.

100

80

60

40

20

0
Neutral-Neutral Whisper-Whisper Neutral-Whisper
% Accuracy MFCC % Accuracy Timbre % Accuracy Timbre(Median)

Fig. 4 Comparative results using MFCC only/timbre feature/timbre (median)

Table 4 Baseline results of

Speech mode % Accuracy
speaker identification
accuracy by timbre features Training Testing
Neutral Neutral 95.0
Whisper Whisper 86.5
Neutral Whisper 73.0

features are expected to be most suitable. However, it is customarily recommended

to use limited and only well-performing features for high speed and performance. By
Hybrid Selection Algorithm, a set of limited features namely MFCC, Roll-off, Bright-
ness, Roughness, and irregularity is used for the speaker identification with a whis-
pered voice. When compared with the traditional MFCC features, the identification
accuracy using selected timbrel features is found to be enhanced by 7.7%. Taking into
consideration, the intra-speaker variability, the performance can be further increased
if the MEDIAN value of training class is used to identify the new speaker query. It is
due to decreased intra-speaker distance according to the analysis in this undertaking.
Hence, the use of median values of timbrel features further enhances the results by
2.4%.

References

1. Hourri S, Kharroubi J (2020) A deep learning approach for speaker recognition. Int J Speech
Technol Springer 23:123–131
Use of Median Timbre Features for Speaker Identification … 41

2. Jahangir R et al (2020) Text-independent speaker identification through feature fusion and deep
neural network. IEEE Access 8:32187–32202
3. Singh A, Joshi AM (2020) Speaker identification through natural and whisper speech signal.
In: Janyani V, Singh G, Tiwari M, d’Alessandro A (eds) Optical and wireless technologies.
Lecture Notes in Electrical Engineering, vol 546. Springer, Singapore
4. Fan X, Godin KW, Hansen JHL (2011) Acoustic analysis of whispered speech for phoneme
and speaker dependency. In: Proceedings of the annual conference of the international speech
communication association, INTERSPEECH, pp 181–184
5. Bhattacharjee M, Prasanna SRM, Guha P (2018) Time-frequency audio features for speech-
music classification. Project: Broadcast Video Analytics
6. Davis SB, Mermelstein P (1980) Comparison of parametric representations for monosyllabic
word recognition in continuously spoken sentences. IEEE Trans Acoust Speech Signal Process
ASSP s28:357–366
7. Hermansky H, Malaya N (1998) Spectral basis functions from discriminant analysis. In:
International conference on spoken language processing
8. Toonen Dekkers RTJ, Aarts RM (1995) On a very low-cost speech-music discriminator.
Technical Report 124/95, Nat. Lab. Technical Note
9. Dobrowohl FA, Milne AJ, Dean RT (2019) Timbre preferences in the context of mixing music.
Appl Sci (2076–3417) 9(8):1695–1695
10. Peeters G (2004) A large set of audio features for sound description (similarity and
classification) in the CUIDADO project
11. Park TH (2004) Towards automatic musical instrument timbre recognition, PhD thesis, the
department of music, Princeton University
12. Albert-Ludwigs-Universität Freiburg (2007) A Matlab Toolbox for Music Information. In:
Proceedings of the 31st annual conference of the Gesellschaft für Klassifikation e.V., March
7–9, pp 261–268
13. Lartillot O (2011) MIR toolbox 1.3.3 (Matlab Central Version)—User’s Manual. Finnish centre
of excellence in interdisciplinary music research, University of Jyväskylä, Finland
14. Sardar VM, Shirbahadurkar SD (2019) Timbre features for speaker identification of whispering
speech: selection of optimal audio descriptors. Int J Comput Appl (Taylor Francis, U.K.). ISSN
1206–212X
15. Cummins F, Grimaldi M, Leonard T, Simko J (2006) The chains speech corpus: characterizing
individual speakers. School of Computer Science and Informatics University College, Dublin
16. Deshmukh S, Bhirud SG (2012) A hybrid selection method of audio descriptors for singer
identification in North Indian Classical Music. In: Fifth international conference on emerging
trends in engineering and technology, pp 224–227
17. Liu H, Jiang H, Zheng R (2016) Computational and mathematical methods in MedicinePB—
Hindawi Publishing Corporation
18. Shah JK, Smolenski BY, Yantorno RE, Iyer AN (2015) Sequential k-nearest neighbor pattern
recognition for usable speech classification. Signal Processing Conference, IEEE Xplorer
19. Sardar VM, Shirbahadurkar SD (2018) Speaker identification of whispering sound: effective-
ness of timbral audio descriptors. In: International conference on power, communications and
sustainable energy system, Chennai
20. Wang J-C, Chin Y-H, Hsieh W-C, Lin C-H, Chen Y-R, Siahaan E (2015) Speaker identification
with whispered speech for the access control system. IEEE Trans Autom Sci Eng 12:1191–1199
21. Sardar V, Shirbahadurkar S (2018) Speaker identification of whispering speech: an investigation
on selected timbrel features and KNN distance measures. Int J Speech Technol 21. https://doi.
org/10.1007/s10772-018-9527-4
Intelligent System for Engine
Temperature Monitoring and Airbag
Deployment in Cars Using

Akshay A. Jadhav, Swagat M. Karve, Sujit A. Inamdar,

and Nandkumar A. Admile

Abstract The CAN bus has emerged as vital means of communication within the
automotive sector. Intrinsically varied applications are being implemented using the
Controller Area Network Protocol. In this paper, two major automotive applications
namely engine temperature monitoring and control and airbag deployment mecha-
nism are implemented with the assistance of ARM7 based microcontroller nodes.
Thus the essential CAN protocol has been implemented for real time embedded
automotive applications.

Keywords ARM7 · Basic CAN Protocol · CAN Bus · Nodes · Real-Time data

1 Introduction

Controller Area Network Bus is an everlasting version of the communication bus

systems utilized in automotive industry. CAN BUS may be a serial digital commu-
nication protocol invented by German BOSCH Corporation within the early 80s to
understand the info exchange between numerous controllers and measuring instru-
ments in modem automobile. It’s a multi-master bus; the communication medium
are often a double stranded wire, coax or optical fiber. Communication speed is
up to 1 MBPS. Bus communication interface integrates the CAN protocol physical
level and therefore the link layer function, and it can complete the framing of the
communication of knowledge processing, including the position filled, the block data
code, the circulation parity check, the priority distinguished and other works. CAN
communication protocol feature is to encode the info block. Length of the info is up
to eight bytes, which may meet the electrical bus control commands, working status
and test data requirement. Meanwhile, the 8 bytes won’t take the bus for an extended
time, so it ensures real-time communication [1]. In this paper, two major automotive
applications namely engine temperature monitoring and airbag deployment mech-
anism are implemented. The project consists of two nodes namely sensing node
and therefore the controlling node. The sensing node and therefore the controlling

A. A. Jadhav (B) · S. M. Karve · S. A. Inamdar · N. A. Admile

SVERIs College of Engineering Pandharpur, Solapur, Maharashtra, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 43

P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_5
44 A. A. Jadhav et al.

node consists of the ARM7 based microcontroller (with inbuilt CAN Controller) and
may Transceiver each. The important time data from sensors of the sensing node
is processed by the microcontroller LPC2129. The CAN Controller takes care of
the CAN broadcast message data. This data is then sent onto the CAN Bus. This
data is then received by the controlling node from the CAN bus. The controlling
node then further takes appropriate action. For controlling the temperature, a lover
is often used; while the airbag deployment mechanism is often achieved by turning
on a relay connected to the controlling node. This ensures the engine temperature
being monitored and controlled further and therefore the driver’s safety; that has
drawn special attention thanks to increasing number of accidents day by day caused
by high risky speeds. Thus the essential CAN protocol has been implemented for
real time embedded automotive applications.

2 System Design

The Fig. 1 shows system block diagram.

The sensing node and therefore the controlling node consist of LPC2129 micro-
controller each. The sensing node senses the sensor data. The sensed data is then sent
onto the bus. This data is then received by the controlling node. The controlling node
then takes the controlling actions accordingly. Here the fan is often wont to control
the temperature of the engine while the relay depicts the deployment of airbag. The
amount of nodes is often added to the bus to extend the amount of applications served.

Fig. 1 Block diagram

Intelligent System for Engine Temperature … 45

2.1 System Hardware Components

A CAN bus automotive electronic system consists of every controller through the
CAN bus interconnected together to exchange information (such as speed, engine
speed, engine ambient temperature etc.). This mutual connection helps the controllers
to use the info at an equivalent time. On the opposite hand it helps to extend the
throughput of the system, to develop updated features within the system and reduce
the system cost. Electronic automotive instrument cluster node can fully reflect the
car factor, product design and technical standards. Intelligent sensor nodes accurately
receive various sorts of signals and at an equivalent time can eliminate interference
signals [2].
1. CAN Bus Electrical Characteristics: CAN transmission medium formed by
the two, One is called high-level transmission line CANH and the other one is
called low-level transmission line CANL.
2. Hierarchical structure of CAN protocol based BUS: According to CAN BUS
network structure, Classified into the following five levels, as shown in Fig. 2.
3. CAN Message Transmission and Frame Type: CAN message use frame as
transmission unit. In the CAN 2.0B specification is provided in standard format
data frame, Given in two different 2.0B frame format, the difference is the length
of different identifiers: Frame with n-bit identifier called the standard frame, the
frame contains 29 identifiers is called extended frame. Message transmission
has the following four different types of frames: data frame, remote frame, error
frame, overload frame [3].
4. Can Bus Node Design: Figure 3 shows the CAN Bus node bus which consists
of two nodes namely sensing node and the controlling node. The sensing node
consists of following sensors: temperature sensor and shock sensor. The real
time data from the sensors is sent to the ARM7 based LPC2129 microcon-
troller and then to the CAN Bus via CAN Transceiver MCP2551. This data is
then received by the controlling node and appropriate action is taken. Thus the
ambient temperature of engine can be monitored. Also to control it further, a

Fig. 2 Hierarchical structure of CAN bus

46 A. A. Jadhav et al.

Fig. 3 Circuit design of CAN node

cooling fan can be started. The other application is of Airbag deployment mech-
anism. As soon as the data is received from the shock sensor, the controlling
node takes appropriate action. The action can be making a relay on as way to
deploy the airbag further.
The sensor used for sensing temperature is LM35 while the sensor used for shock
sensing is 801S. The shock sensor is micro vibration detection sensor and is useful
for the demo purpose. The microcontroller used for the node is LPC2129 (ARM7
based and Philips made microcontroller). It has several advantages including inbuilt
CAN Controller, supports high speed CAN operation, low power consumption etc.
The CAN Transceiver plays a major role here. It is used for communication between
the node and the CAN bus and then the other nodes. The data is sent onto the bus
via the high speed CAN Transceiver. The CAN Transceiver MCP2551 is compatible
with CAN 2.0B. It can provide differential transmit capability to the bus and differ-
ential receive capability to the CAN controller. The chip was originally designed for
automotive high-speed communications is fully compatible with ISO/DIS standard,
and the moment of interference in the anti-car environment, protecting the ability
of the bus. It is compatible with ISO-11898 standard at speeds up to 1 Mb/s, can
resist the strong interference the moment of interference such as the automotive
environment. General internal bus of circuit protection and limiting circuit, and has
a low current standby mode and slope control to reduce radio frequency interfer-
ence [4]. The sensing node may control the temperature by turning a cooling fan
on. Also the airbag deployment is depicted by turning on a relay. Thus two appli-
cations namely engine temperature monitoring and airbag deployment mechanism,
have been implemented.
Intelligent System for Engine Temperature … 47

3 System Software Components

The system software is written in embedded C and is compiled using ARM Keil Micro
vision v4 software. Flash Magic is used to upload the program to the Microcontroller
chip. The software flow consists of CAN transmission and Reception of messages.
The flow is depicted below with the aid of flowcharts as shown in Figs. 4, 5.

3.1 Transmission flow

See Fig. 4.

Fig. 4 Transmission flow

48 A. A. Jadhav et al.

Fig. 5 Reception flow

3.2 Reception Flow

The CAN Controller is initialized first. The data from one node is sent onto the bus
and then received by the other node. The flowcharts of transmission and reception
of data are shown in figs. 4 and 5 [5]. The received messages are then monitored on
the terminal of a personal computer (Fig. 6).

4 Results

As shown in Fig. 6 the real time data can monitored through UART on a personal
computer. The messages with corresponding ids can be seen on the terminal. A
specific message can be dedicated to certain sensor data. The real time data thus
can be monitored and corresponding controlled actions can be taken further. The
engine temperature can be controlled by turning on a cooling fan; while the airbag
deployment mechanism can be achieved by turning on a relay.
Intelligent System for Engine Temperature … 49

Fig. 6 Data monitored at the terminal of PC

5 Conclusion and Future Scope

The paper proposes implementation of two salient automotive applications namely

engine temperature monitoring and airbag deployment mechanism. The engine
temperature thus is often controlled further which is one among the factor that helps to
extend its efficiency. Also the airbag deployment mechanism ensures driver’s safety.
The scope of the paper is often extended further by increasing the amount of nodes
like respective applications. Thus more and more automotive applications are often
deployed in cars to form them as efficient as possible.
50 A. A. Jadhav et al.

References

1. Piao C-H, Chen L, Cao J (2010) A design for controller area network bus real-time monitoring
system. In: 2011 International conference on computer science and network technology, pp
1621–1624
2. Bo L, Tao J (2012) The design of monitoring system based on CAN bus. In: International
conference on measurement, information and control (MIC), pp 137–140
3. Ran L, Junfeng W, Haiying W, Gechen L (2010) Design method of CAN BUS network
communication structure for electric vehicle. In: IFOST Proceedings. 978-1-4244-1849-7/08/
4. Li R, Liu C, Luo F (2008) A design for automotive CAN bus monitoring system. In: IEEE
vehicle power and propulsion conference (VPPC), Sept 3–5, Harbin, China
5. Quanqi W, Jian W, Yanyan W Design of vehicle bus data acquisition and fault diagnosis system.
978-1-61284-459-6/11/
Analysis and Prediction
of Temporomandibular Joint Disorder
Using Machine Learning Classification
Algorithms

Roopa B. Kakkeri and D. S. Bormane

Abstract Temporomandibular joint disorder (TMD) includes specifically a series

of musculoskeletal disorders that may affect the masticating system. Roughly 30–
40 percent of adults today have oral problems, and the most common cause of oral
problems is TMJ. This disorder is very prevalent in the general population, but it
affects more women and young people. The focus of this research review was on the
methods for detecting TMJ disorder using machine learning algorithms. Propelled
with the rise in use of machine learning techniques in the research dimensions of
medical diagnosis, in this paper there is an attempt to explore different classification
for predicting the TMJ disorder. The proposed techniques are evaluated on real time
TMJ datasets. Dataset related to TMJ screening in subjects had 84 instances and 11
attributes. After applying different machine learning techniques, results suggest that
Naïve Bayes and Adaboost models work better with higher accuracy of 93% and
92%.

Keywords Temporomandibular joint disorder · Surface electromyography ·

Machine learning · Classification algorithm · SVM · Naïve Bayes · Adaboost

1 Introduction

Today, the issue of TMJ disorder has been growing systematically across all ages
of the human population. Early diagnosis of this musculoskeletal condition will
significantly aid in maintaining the mental and physical health of the subject [1–
3]. With the increase in the use of machine learning-driven models in the forecasts
of various human diseases, their early detection now seems to provide promising
results based on different health and physiological constraints. We were inspired
by this reason to raise awareness in the diagnosis and study of TMJ condition to

R. B. Kakkeri (B)
AISSMS Institute of Information Technology, Sinhgad Academy of Engineering, Kondwa, Pune,
India
D. S. Bormane
AISSMS College of Engineering, SPPU, Kennedy Road, Pune, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 51

P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_6
52 R. B. Kakkeri and D. S. Bormane

enhance improved technique of treatment. The round end of the lower jaw, known
as condyles, slips around the socket of the temporal bone joint when we open our
mouth [3]. TMJ disease diagnosis becomes a problem when there are many other
facial and muscle conditions whose few signs are very similar to those with signs
of TMJ disease, which is why it is considered a complex disorder. That makes this
job a challenging one. Early detection and treatment [4] are most imperative steps
to be taken to decrease the symptoms of TMJ disorder problem and to improve the
quality of life of suffering people. However, there is no procedure of medical test
for detection of TMJ disorder. In dentistry, One of the most common disorder in the
oral cavity that mostly affects women is temporomandibular joint disorder. Popular
Temporomandibular Joint Disorder (TMD) symptoms include,
• Problems with opening your mouth wide as you try.
• Jaws in the open-mouth or closed-mouth role that get “stuck” or “stall”.
• When you open or shut your mouth or chew, cracking, ringing, or grating sounds
in the jaw joint.
• Inflammation or swelling on the side of your face [1–3].
• Dental pain, headaches, neck aches, dizziness, earaches, hearing issues, upper
shoulder pain, and ear ringing (tinnitus) can also be present in patients [4].
The method of diagnosing this is conducted manually with the use of clinical tools
that with improper use becomes unsafe for the patient. The existing clinical methods
are also affluent for or common citizens. Arising technologies aided with applications
through the use of machine learning algorithms. Signal processing is practically
applied in medical diagnosis Digital acquisition techniques with processing of signals
becomes an important option especially in medical diagnosis [3, 5]. Moreover, the
clinical diagnostic procedure for TMJ disorder is costly. In the long run, lack of
awareness or neglecting of early signs and symptoms of TMJ disorder may lead to
even costlier clinical treatment. The general objective of the study is
To detect Temporomandibular Joint Disorder (TMJD) using surface electromyo-
graphy of masticatory muscles.
To introduce current prospects of diagnosis and treatment of Temporomandibular
Joint Disorder.
To mention advantages of using surface electromyography of masticatory
muscles for detection of Temporomandibular Joint Disorder using machine leaning
classification algorithms.
The current paper aims to predict TMJ disorder. This paper is arranged in the
following sections: Sect. 1 gives introduction, Sect. 2 discuss materials and method-
ology. In Sect. 3, experimental results are described. Finally, the conclusions and
future scope of improvement are presented in Sect. 4.
Analysis and Prediction of Temporomandibular … 53

Fig. 1 Block diagram of the proposed work

2 Literature and Methods

This research study is categorized as a form of qualitative research as it determines

the suitability of masticatory muscle surface electromyography for the diagnosis of
TMJ disorder through machine learning techniques [6, 7].

2.1 Data Acquisition and Subjects

The first part in the research process conducted is the data assembly [3, 8]. Literature
survey is carried out by reading of articles and related journals from the internet
and clinical study was conducted to get details about TMJ disorder, its signs and
symptoms with radiographic evaluation through Orthopantomogram, Conversation
with Neurologist and dental surgeons were directed to alleviate the early findings
in selecting the subjects. A consent form is taken from all the subjects to willfully
participate in the study. Following Fig. 1 shows block diagram of proposed work.

2.2 Materials and Methods

The present study was conducted with total 84 subjects. Forty one subjects were
included in the control group and the rest were in experimental group. While
selecting the subjects, those with the history of dysfunction syndrome and orthodontic
management, lip incompetence, soreness in any muscle of mastication and previous
rebuilding in tooth were excluded.
A complete dental history is made for each subject and intraoral photographs
(OPG) are taken to identify the TMJ parameters [6]. Frontal belly of masseter and
frontward belly of temporalis of both sides (which are for convenience mentioned
here as masseter and temporalis, respectively) were studied.
For all the subjects, bipolar surface electrodes were used. The EMG recording
of these muscles was done in both groups at most calm position of jaw i.e., at
54 R. B. Kakkeri and D. S. Bormane

rest position and during maximum clenching at maximum intercuspal position [9–
11]. In both control group and experimental group, during the activity of maximum
clenching and during rest, the recording was done. Before recording, volunteers were
seated comfortably on a normal chair. Position of the head was kept erect and no
undertaking was allowed as jaw muscles respond to change in crown position. The
subject was settled and not able to see the computer screen. Facial areas were prepared
using a Medi-swab. EMG signals were composed using bipolar throwaway surface
electrodes of 6 mm diameter (Blue Sensor A-50-VS, Medico test, lstykke, Denmark)
placed on the volunteer’s temporalis and masseter muscle areas on both sides of
the face, thus giving four data collection channels (temporalis-right, masseter-right,
masseter-left, temporalis-left). Electrode cream (Grass EC2, USA) was functional
to give a flat surface on the electrode face and then the protective support detached.
Two surface electrodes were applied to the muscle, next to each other, the centers
16 mm apart laterally along each muscle, which was located by palpation. Reference
electrode was involved to an ear lobe.
The electrodes were associated via a Polygraph amplifier (Model 7E, Grass Instru-
ment Co., Massachusetts, USA) to an ADC interface comprising 12-bit, 333 kHz
and sampled using Spike2 data collection software (Cambridge Electronic Design
Ltd., Cambridge, UK), sampling at a rate of 1 kHz. The Polygraph 7P3 preamplifier
modules used filters (26 dB) of 3 Hz and 20 kHz, thence to 7DA driver amplifier
modules with output frequency response to 40 kHz (26 dB). The 7DA module had a
notch 50 Hz filter to suppress mains electrical interference. Recording was continued
in each case for up to 1 min.

2.3 Data Pre-processing

Pre-processing of data is a method in which the raw data is changed into a substantial
and practical format. Real-world evidence is typically partial and unreliable since
it includes a lot of mistakes and negligible values [5, 9]. A pre-processed upright
details still vintages to a decent performance. Various types of data pre-processing
[8] are used to adjust inaccurate and conflicting data, such as managing missing
information, minimizing data, etc.

3 Methodology

The next move was to develop the algorithm for the diagnosis of TMJ disorder
with obvious results on how to diagnose TMJ disorder by feature selection [12]. A
series of Computer Vision outlines is referred to and the probability of using them
to create the algorithm was checked. In the development of the algorithm [6, 7],
python libraries were attempted. This system libraries and documents were analyzed
Analysis and Prediction of Temporomandibular … 55

Fig. 2 Final training and testing

on which of these may provide the simplest but most efficient method for obtaining
the outcomes.

3.1 Training and Testing Model

The whole dataset has been split into two parts, i.e. one part is the training dataset
other part is the dataset testing with a ratio of 70:30 respectively. Training data has
once again been broken into two parts for cross-validation purposes. One component
is the preparation dataset, and another component is the 70:30 ratio evaluation dataset.
Fig. 2 shows training & testing.

3.2 Support Vector Machine (SVM)

SVM is an algorithm for supervised machine learning which can be used for problems
of classification or regression [13, 14]. Fig. 3 shows SVM To transform the results,
it utilizes a technique called the kernel trick, and then finds an optimal boundary
between the possible outputs based on these transformations. Simply put, it performs
some highly complex transformations of information, then works out how to isolate
the information depending on the labels or outputs you have identified. We used a
linear model of the kernel [12].
56 R. B. Kakkeri and D. S. Bormane

Fig. 3 Support vector machine

3.3 Logistic Regression

It is used to test discrete values depending on a given collection of independent

variable(s) (Binary values such as 0/1, yes/no, true/false). Logistic regression is
similar to linear regression, but the curve is built using the normal logarithm of the
target variable’s “odds” rather than the likelihood [7]. Fig. 4 shows logistic regression.
K Nearest Neighbor (KNN) KNN is just a broad view of the nearest neighbor’s
algorithm. Include k-nearest neighbors in their place of considering the closest
neighbor from a dataset containing n data points, k communicates how many neigh-
boring neighbors will have an influence on the classification process [13, 14]. The
distance from Euclidean, Manhattan, Minkowski and Hamming may be such distance
functions. The first three functions are used for continuous functions, and the fourth
one (Hamming) is for categorical variables. Based on the square root, the mainstream
labelled data point to which the unlabeled point must be classified is determined by
the number of samples, depending on the K.

Fig. 4 Logistic regression

Analysis and Prediction of Temporomandibular … 57

Fig. 5 Adaboost classifier

3.4 Adaboost

A machine-learning algorithm expressed by Yoav Freund and Robert Schapire is the

Ada Boost (short for “Adaptive boosting”) widget it is shown in Fig. 5. To boost their
presentation, it can be used with other learning algorithms [7]. This is achieved by
the tuning of poor learners. Boosting refers to an all-purpose and demonstrably true
means of creating a very precise classifier by mixing rough and mildly inaccurate
thumb laws.

3.5 Naïve Bayes Algorithm

It is a classification technique with a presumption of freedom between predictors

based on Bayes’ theorem. A Naïve Bayes classifier believes, in basic terms, that
the inclusion of a certain feature in a class is equivalent to the existence of any
other feature. The Bayes theorem provides a way for P(c) posterior likelihood to be
determined from P(c), P(x) and P(x) [7].

4 Result and Discussion

The result is measured in terms of specificity, sensitivity, and accuracy by using the
confusion matrix and classification report. The result depends on how accurate the
model is trained. The most relevant and important features selected were Masseter
and Temporal muscles for left and right during rest and Maximal voluntary clenching
(MVC).

4.1 Performance Evaluation Metrics

Gauging performance is key to check how fine a classification model work to attain
an objective. Performance valuation metrics are used to evaluate the efficacy and
58 R. B. Kakkeri and D. S. Bormane

performance of the classification model on the test dataset. It is important to choose

the exact metrics to evaluate the model performance such as confusion matrix, accu-
racy, specificity, sensitivity, etc. Following formulas are used to find the performance
metrics [13]:
Accuracy—Accuracy is the most innate performance quantity and it is simply a
ratio of acceptably predicted observation to the total observations.
Precision—Precision is the ratio of correctly predicted positive observations to the
total predicted positive observations. High precision relates to the low false positive
rate.
Recall (Sensitivity)—Recall is the ratio of correctly predicted positive observa-
tions to the all observations in actual class—yes.
F1 score—F1 Score is the biased average of Precision and Recall. Therefore,
this score takes both false positives and false negatives into account. Naturally it is
not as easy to understand as accuracy, but F1 is usually more useful than accuracy,
especially if you have an uneven class distribution.
In the biomedical field, classification has acquired a lot of importance in machine
learning. Classification techniques help to learn a model from a set of training data
and to classify one of the classes with test data. The analysis of the few existing clas-
sification algorithms [12, 15] is related to this study and their comparative efficiency
parameters will help to study the new algorithms.
The main kernel used in the analysis of EMG signal features to achieve better clas-
sification accuracy is the selected features. This article’s key purpose is to determine
the consistency of the classification [13, 14]. This article discusses a few classification
strategies for Supervised Machine Learning (ML) [12], compares and determines the
most successful classification algorithm based on the data collection, the number of
instances and variables (features). Five separate machine learning algorithms were
considered: k-Nearest Neighbor (KNN), Decision tree, Naïve Bayes, Logistic regres-
sion and Support Vector Machine. The TMJ disorder data set was used for the clas-
sification to apply the equations with 84 instances with ten characteristics as an
independent variable and one as a dependent variable for the study.
Cross-validation is a method for testing machine learning models by training
many machine learning models on the available input data sub-sets and testing them
on the sub-set of complementary data. Using cross-validation to avoid over fitting,
i.e. failing to generalize a pattern.

4.2 K-Fold Cross-Validation

Cross-validation shown in Fig. 6, is a re-sampling practice used to measure machine

learning models on a partial data set. To achieve cross-validation, we may use the
method of k-fold cross-validation. You have fragmented the input data into k data
subsets (also known as folds) during k-fold cross-validation.
Analysis and Prediction of Temporomandibular … 59

Fig. 6 k-fold cross-validation model

On all but one (k − 1) of the subclasses, train ML model, and then test the model
on the sub-set that was not used for training [14]. This procedure is repeated k times,
each time with a different sub-set intended for evaluation (and omitted from training).

4.3 Results

The final result is shown in following Fig. 7.

Fig. 7 Overall results of models with tenfold cross validation

60 R. B. Kakkeri and D. S. Bormane

5 Conclusion

Temporomandibular disorder is a condition which can put the individual in severe

pain, agony as well as unable to chew food properly. Presently, a proper diagnosis
and a structured treatment plan is not available as the cost of advanced equipment,
technique sensitivity of handling the equipment as well as understanding of multidis-
ciplinary approach is inadequate. It was clear after the literature survey that several
machine learning algorithms can be applied to the extracted and chosen features. Out
of the five supervised training of classification algorithms, the results indicate that
SVM was not introduced to be the most reliable and effective algorithm. Relative
analysis of the classifiers shows that Decision tree, Logistic regression and Random
Forest classification algorithms stood to be the reliable with cross validity of ten
folds.

References

1. Uddin F, Baten RBA, Rita SN, Sadat SA, Chowdhury NM (2017) Management of temporo-
mandibular joint dysfunction syndrome: an overview. J Bangladesh Coll Physicians Surg
35(3):133–141
2. Klasser GD, Okeson JP (2006) Electromyography in the diagnosis and treatment of temporo-
mandibular disorders. Oral Health 137:763–771
3. Mapelli BC, Zanandréa Machado LD, Giglio CS, De Felício CM (2016) Reorganization of
muscle activity in patients with chronic temporomandibular disorders. Arch Oral Biol 72:164–
171
4. Bianchi J et al (2020) Osteoarthritis of the temporomandibular joint can be diagnosed earlier
using biomarkers and machine learning. Sci Rep 10(1):1–14
5. Spiewak C (2018) A comprehensive study on EMG feature extraction and classifiers. Open
Access J Biomed Eng Biosci 1(1):1–10
6. Khezri M, Jahed M (2007) Real-time intelligent pattern recognition algorithm for surface EMG
signals. Biomed Eng Online 6:1–12
7. Sweeney EM et al (2014) A comparison of supervised machine learning algorithms and feature
vectors for MS lesion segmentation using multimodal structural MRI. PLoS ONE 9:4
8. Chowdhury RH, Reaz MBI, Bin Mohd Ali MA, Bakar AAA, Chellappan K, Chang TG
(2013) Surface electromyography signal processing and classification techniques. Sensors
(Switzerland) 13(9):12431–12466
9. Latif R, Sanei S, Shave C, Carter E (2008) Classification of temporomandibular disorder
from electromyography signals via directed transfer function. In: 30th annual international
conference of the ieee engineering in medicine and biology society, vol 2, no 3, pp 2904–2907
10. Pinho JC, Caldas FM, Mora MJ, Santana-PenÍn U (2008) Electromyographic activity in patients
with temporomandibular disorders. J Oral Rehabil 27(11):985–990
11. Suvinen TI, Kemppainen P (2007) Review of clinical EMG studies related to muscle and
occlusal factors in healthy and TMD subjects
12. Khammas BM, Monemi A, Bassi JS, Ismail I, Nor SM, Marsono MN (2015) Feature selection
and machine learning classification for malware detection. J Teknol 77(1):243–250
13. Khan MMR, Arif RB, Siddique AB, Oishe MR (2019) Study and observation of the variation
of accuracies of KNN, SVM, LMNN, ENN algorithms on eleven different datasets from UCI
machine learning repository. In: 4th international conference on electrical engineering infor-
mation and communication technology iCEEiCT 2018, pp 124–129
Analysis and Prediction of Temporomandibular … 61

14. Kucuk H, Tepe C, Eminoglu I (2013) Classification of EMG signals by k-nearest neighbor
algorithm and support vector machine methods, pp 1–4
15. Al-Faiz MZ, Ali AA, Miry AH (2010) A k-nearest neighbor based algorithm for human arm
movements recognition using EMG signals, EPC-IQ01 2010. In: 1st conference energy, power
control, vol 6, no 2, pp 159–167
Machine Learning Approach
in Cooperative Spectrum Sensing
for Cognitive Radio Network: Survey

Vaishali S. Kulkarni, Tanuja S. Dhope(Shendkar), Swagat Karve,

Pranav Chippalkatti, and Akshay Jadhav

Abstract In cognitive radio network some of the important functionalities is spec-

trum sensing. It plays a very vital role for unlicensed system to operate efficiently and
to provide the required improvement in spectrum efficiency. If the spectrum, which is
sensed is in idle state allow the unauthorized users (secondary users) to use the spec-
trum. Machine learning algorithms are used for spectrum sensing in cognitive radio
networks. They are weighted K-nearest neighbor, Support Vector Machine (SVM)
which comes under supervised learning and Gaussian Mixture Model (GMM), K-
means clustering which comes under unsupervised learning-based classification tech-
niques. In this paper rigorous survey is done by using machine learning algorithms
to review various methodologies used in spectrum sensing like K-nearest -neighbor,
GMM, K-means clustering and SVM.

Keywords Machine Learning · Support Vector Machine (SVM) · Primary user ·

K-means clustering · Secondary user · Spectrum sensing · Cognitive radio

1 Introduction

Due to rapid development in wireless network, a technology called Cognitive Radio

(CR) is developed to overcome the spectrum scarcity. Using this technology both
primary users (licensed) and secondary (unlicensed) users can utilize the spectrum.
In cognitive radio networks spectrum utilization increases with the spectrum allo-
cated dynamically. The unlicensed users can transmit in the vacant spectrum already
assigned to licensed users with minimum level of interference. To find the vacant
spectrum it senses and selects the spectrum by which it can meet the QoS. The licensed
users get back the spectrum from the unlicensed users whenever the licensed users
return. In cognitive radio networks spectrum sensing technique is one of the major

V. S. Kulkarni (B) · T. S. Dhope(Shendkar) · P. Chippalkatti

Department of Electronics & Telecommunication, JSPM’s Rajarshi Shahu College of
Engineering, Pune, India
S. Karve · A. Jadhav
SVERI’s College of Engineering, Pandharpur, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 63

P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_7
64 V. S. Kulkarni et al.

Fig. 1 Classification of
spectrum sensing

techniques [1] in which the harmful interference is prevented with licensed users and
the available spectrum is identified to improve the spectrum’s utilization.
In cognitive radio network one of the important techniques is Spectrum sensing
and is also the primary task for establishing a cognitive radio system [1].
Spectrum Sensing is of three types. Interference-based sensing, non-cooperative
sensing and Cooperative sensing.
In cognitive radio networks, machine learning methods of Cooperative spec-
trum sensing is the major leading technique. The capability of decision making can
be enhanced by combining machine learning with cognitive radio networks from
the previous experiences [2]. Compared to traditional methods unstable cognitive
spectrum resources can be used effectively.
Figure 1 shows the classification of Spectrum sensing. Non-cooperative sensing,
interference-based sensing and Cooperative sensing Spectrum sensing are the
classification of spectrum sensing.
The performance of detection of spectrum holes is enhanced by cooperative
sensing in which secondary user senses the spectrum with the collaboration with
primary user. Also known as receiver detection technique.
In non-cooperative spectrum sensing, individual CR scans for primary signal and
based on its detection CR decides the presence or absence of primary user.
Cognitive radio measures the interference environment and their transmissions
are adjusted such that the interference to PU is not above the regularity limit in
interference-based sensing [3].
Cyclo-stationary feature detection, energy detection and match filter detection are
the classification of non-Cooperative sensing.
One of the intelligent cooperative sensing methods is machine learning which has
the special feature of self-adaptation to environment, high sensing.
Table 1 shows the comparative advantages and disadvantages of spectrum sensing
methods [4]. The author has created this table to familiarize spectrum sensing and for
better understanding the state-of-art techniques to the researchers. The performance
of any spectrum sensing algorithm is determined by the probability of detection,
probability of false alarm, the energy detection for cognitive radio network follows
the following formula [5, 13].
Machine Learning Approach in Cooperative Spectrum … 65

Table 1 Spectrum sensing methods

Sr. No Techniques of sensing Pros Cons
1 Energy detection 1. There is no requirement 1. Noise uncertainty is
of pre knowledge of sensitive
primary signal 2. False alarm rate is high
characteristics
2. Implementation is easy
2 Cyclostationary feature 1. Signal and noise can be 1. When the sample size is
detection distinguished large, energy
2. Probability of false alarm consumption is high
decreases at low SNR
3 Matched filter-based 1. Detection is better at low 1. Knowledge of primary
detection SNR region user signal is required in
2. Sensing is optimal advance, hence
impractical
4 Detection based on It is not essential to have the Computation is very
Covariance knowledge of primary complex
signal characteristics in
advance
5 Spectrum sensing based on 1. Complex model can be 1. Techniques are complex
Machine learning used in easy manner 2. Detection rate is affected
2. Delay of detection is by feature selection
minimized

1. Let time of sensing be τs, at the sensing node, when the PU is absent, the signal
that is received is given by

s(u) = σ n(u) (1)

where σn(u) is the noise power at duration d.

2. When the PU is present, signal is given as

s(u) = g(u). r(u) + σ n(u) (2)

where g(u) is gain of the channel, r(u) is primary signal which is received.
3. Assuming some of the parameters, detection of energy is given by

τ s
1
ED = |s(u)|2du (3)
P0
0

where P0 is the spectral density of power τs is the time of sensing.

4. Probability of detection and probability of false alarm using energy detection can
be given by considering N cooperative sensing nodes. Probability of detection
is given by
66 V. S. Kulkarni et al.
 
αj − (2τsCB + βjτsCB)
pj (det)(αj , βj , τs , CB ) = Q 
4τsCB + 4βjτsCB
  ∞   (4)
1 −t2
where Q(u) = √ exp
2π u 2

where pj (det) is the probability of detection and is given by P (β ≥ αj | H1) at j

belonging to 1 to N, αj = threshold of detection, H1-presence of PU.
Probability of false alarm is given by
 
αj − 2τsCB
pj (false)(αj , τs ,CB ) = Q √ (5)
4τsCB

where pj (false) = P (β ≥ αj | H0) at j belonging to 1 to N, h2-absence of PU.

The organization of the paper is as follows. Literature survey of various method-
ologies of cooperative spectrum sensing are described in Sect. 2. Conclusion in Sect. 3
is preceded by a table of comparative methodologies.

2 Literature Survey

Zan Li et al. [6] proposed CSS model of machine learning in which user grouping
concept is shown. CSS framework based on optimization model which had four
modules. Group scheduling module, SVM classification module, user grouping
module and SVM training module are the four modules of grouping concept. Without
the reduction in accuracy of sensing and the functions like grouping of optimized
users, redundant and abnormal are achieved by using three algorithms of grouping.
In the algorithm of first grouping there are groups of two which are normal and
abnormal. Then in second grouping algorithm redundant and non-redundant users
are distinguished and finally optimized model is established. As per the requirement
of cooperative sensing optimization of specific number of groups are done where
the users are divided, selecting each time. Finally, the user group are non-redundant
and secured that involve in the recognition of pattern. Optimization problem can be
solved by Binary particle swarm optimization. By using this grouping algorithms,
the CR networks can be avoided the harm of abnormal and the redundant users.
Assigning silent state for long term or temporarily for many users reduces the unnec-
essary energy consumption. The efficiency of sensing can be improved by increasing
the grouping. This is done by increasing the speed of classification in model of SVM
in machine learning. Centralized CSS model network consisting of fusion center is
used in the simulation with 10 or 20 CRs. Here the simulation result shows that there
is 17.98% rate of improvement for the speed of average classification. Along with the
entire network security they also showed the operational efficiency of SVM model
increased using group algorithm 1. Few non-highly redundant users and highly redun-
dant users are used in grouping algorithm 2 with network of higher signal quality.
Machine Learning Approach in Cooperative Spectrum … 67

SVM achievement 100% accuracy. In algorithm 3, the users of CR are divided into
cooperation groups of two or more which are optimized. Cooperative sensing can be
performed independently by each group. The classification speed increases with the
number of groups being more.
Yingqi Lu et al. [7] proposed machine learning based classification which uses the
probability vector of dimension being low. This method has the time for classifica-
tion shorter and duration for training smaller. So, energy vector with N-dimensions
is not used here. Keeping the same spectrum performance and by converting a high
dimensional feature vector, a feature vector of constant two-dimensional is achieved.
With M secondary users and primary user selecting single SVM and K-means algo-
rithms of machine learning are used. Small scale CRN uses primary user as single
and M = 2SUs and large-scale CRN using single primary user (PU) and M = 9SUs
are considered.
In Small scale CRN, the techniques of machine learning are trained and then
new feature vectors are tested, which gives the result of (PFA) probability of false
alarm and (PD) the probability of detection. They showed that The accuracy of
detection compared to algorithms of AND rule and OR rule which are hard fusion,
classification based on machine learning is better. In terms of energy vector, there is
better performance by SVM based classification using probability vector. With the
highest transmitting power of PU, performance is better.
In large scale CRN, through plot of Receiver operating characteristics curves
accuracy is shown better in both SVM-ploy and SVM-linear when the probability
vector is combined. Here energy vector is replaced by probability vector and proved
the accuracy to be better.
Through simulation they proved that
1. With the equal training samples, the duration of training for K-means clustering
is longest.
2. SVM-ploy takes more training time than SVM-linear as more time is spent on
feature vector mapping by functions of polynomial kernel.
3. SVM-linear is more efficient than SVM-ploy.
4. With the probability vector, the performance is best of SVM linear compared
to various decision-making scheme such as K-means, SVM-ploy. These deci-
sions making schemes have duration of training and classification delay low,
probability of detection high.
Xiaolin Ma et al. [8] proposed Extreme Learning machine (ELM) for numerous
primary users to implement the algorithm of sensing cooperative spectrum with the
network of cognitive radio. ELM which is an algorithm of machine learning that
has three step free tuning process is used to make very accurate the channel sensing
and identification which involves. In this paper initially the identification of multiple
primary users is established by spectrum sensing model. A classification scheme
of channel pattern can be obtained by ELM with the help of energy detection and
channel model. Channel model as well as sensing model, energy model, fusion center
is also implemented.
68 V. S. Kulkarni et al.

In channel model, the primary user which uses the channel detects the signal
consists of an authorized signal and noise which is called Rayleigh channel. Using
multiple primary users, the signals can be detected by secondary user in the network
if the number of primary users is more than one. In energy model, usage of channel
by primary users can be sensed by energy. The fusion center aims at determination
of usage of channel. Based on the transmission of energy vectors from secondary
user channel usage is determined.
Simulation is carried out with multiple primary users and the results are compared
with SVM, it was shown that training time of ELM is shorter. The detection
probability being higher the probability of overall detection will also be better.
SVM and ELM algorithms were proposed to train the samples with 1000 and 2000
as the sample size. It is proved that ELM needs only 92 ms to reach probability of
detection compared to SVM which requires 8 times more time than ELM. Accuracy
of SVM is also less than ELM.
Olusegun Peter Awe et al. [9] worked in cognitive radio network in order to
solve the problem of sensing the spectrum on the condition of multiple primary
users and proposed algorithm based on based multi-class support vector machine
(SVM) which is ECOC (error correcting output codes). Spectrum holes is detected
by Joint spatial–temporal detection which is implemented using this algorithm. To
study the attributes of each state, through one-versus-all (OVA) and one-versus-
one (OVO) Multi-class support vector algorithms are used [10–12]. Error correcting
output codes (ECOC) scheme is used to solve the optimization problem. The multi
class problem can be dealt with ECOC which is a framework that creates multiple
binary classification task by breaking it [9] and [10]. Initially the condition is consid-
ered as detection of multiple class signal in which more than one sub-class are
formed from each class. Then multi-class SVM algorithms checks for performance
of the energy-based features and ECOC (error correcting output codes). Detection
of performance is mainly based on accuracy of classification and curves of receiver
operating characteristics. They proved that unused spatial spectrum can be detected
by using ECOC.
Karaputugala Madushan Thilina et al. [1] proposed technique of machine learning
for cognitive radio in which both supervised and un-supervised methods are used.
weighted K-nearest-neighbor and support vector machine learning is used in super-
vised and gaussian mixture model, K-means clustering is used in unsupervised where
cooperative spectrum sensing is implemented. Initially by using the energy levels of
vector the availability is checked for channel which is considered as feature vector.
The decision is taken by feeding this to the classifier. A classifier categorizes into
two categories the availability of channel with available class and the unavailable
class for each feature vector. The training phase has to be covered by classifier before
the classification being online. Partitioning of training feature vectors into K clus-
ters is done by implementing the algorithm of K-means in which test energy vector
class and primary user state represented by each cluster is determined by classifier.
Mapping of each cluster determines the available and unavailable class of channel.
Training feature vectors are obtained from Gaussian mixture distributions in GMM
which corresponds to a cluster. The SVM and the KNN is also proposed due to the
Machine Learning Approach in Cooperative Spectrum … 69

capability of higher prediction. In the SVM, the margin between separating hyper-
planes and feature vectors are maximized to obtain support vectors. Performance is
determined by evaluating all the classification techniques such as the ROC curve, the
classification delay and training time.
The result shows that
1. Capability of classification type K-nearest neighbor for 1000 energy vectors is
shown by the time taken for uploading the training energy vector to the classifier
as 50 μsas compared to GMM having high training duration of 12,796 s for
1000 samples and SVM having highest training duration of 1. 65,817 s for 1000
samples
2. KNN classifier has high classification delay even though training time is lowest
compared to Fisher linear discriminant which shows lowest classification delay.
Also, no change is observed in classification delay in GMM, K-means clustering
and Fisher linear discriminant classifiers with the training energy vector with
different batch.
3. K-means clustering is better approach compared to another classifier because of
capability PU detection being higher and classification delay and lower training.
4. Compared to other algorithms performance of detection is found highest in
SVM classifier by using Kernel functions like polynomial kernel and linear
kernel which is used to map feature space with higher dimensional space.
They also concluded that the classifiers can be trained by obtaining the energy
vectors one to one by improving the CSS approaches. Comparative Table Method-
ologies.
Table 2 shows the comparative methodologies used by various authors and their
findings which led to a better result.

3 Conclusion

In wireless communication systems, spectrum plays avital role as a valuable resource.

The challenging concept in radio network is Cognitive radio. The cognitive radio
which provides the channel wireless, can be dynamically configured and also be
programmed in its vicinity to avoid interference and congestion.
In this survey different methodologies using ML are presented with their results.
various spectrum sensing methods with their pros and cons are presented in this
paper. Various methods have been studied for Dynamic spectrum sensing like K-
nearest-neighbor, GMM, K-means clustering and SVM. From this literature survey
it is observed by comparing to various methodologies, support vector machine gives
better result. Further work will be carried out with support vector machine algorithm.
70 V. S. Kulkarni et al.

Table 2 Comparative table methodologies

Sr. No Related work Topic Methodology Results
1 Zan Li et al. Improved SVM based CSS User grouping
cooperative spectrum model is used for algorithm for
sensing model based group scheduling multi-band
on ML for CRN and User grouping cooperative sensing
modules
2 Xiaolin Ma et al. CSS using Extreme To obtain channel compared to SVM
learning machine for pattern classification detection probability
cognitive radio ELM (Extreme is higher in ELM
networks with learning machine) is and also training
multiple primary used for energy time is shorter in
users detection model and ELM
Wireless channel
model
3 Yingqi Lu et al. ML techniques with System model and The performance of
probability vector for probability vector for K-means clustering
CSS in CRN SVM and K-means or SVM algorithms
clustering algorithms can be improved by
probability vector
4 Olusegun Peter Cooperative Multi class support Successfully
Awe Spectrum Sensing in vector machine detected unused
Cognitive Radio spatial spectrum
Networks using opportunities
Multi-clss Support
Vector Machine
Algorithms”
5 Karaputugala Machine Learning Unsupervised Unsupervised
Madushan thilina Techniques for classifiers-GMM, classifiers-K-means
et al. Cooperative K-means clustering and Supervised
Spectrum Sensing in Supervised classifiers-SVM are
Cognitive Radio classifiers-K-nearest accurate approaches
Networks neighbor and SVM for CSS

References

1. Thilina KM, Choi KW, Saquib N, Hossain E (2013) Machine learning techniques for
cooperative spectrum sensing in cognitive radio networks. IEEE J Sel Areas Commun
31(11):2209-2221
2. Kim KJ, Kwak KS, Choi BD (2013) Performance analysis of opportunistic spectrum access
protocol for multi-channel cognitive radio networks. J Commun Netw 15(1):77–86
3. Arjoune Y, Kaabouch N (2019) Review-a comprehensive survey on spectrum sensing in cogni-
tive radio networks. In: Recent advances, new challenges, and future research directions,
sensors
4. Bkassiny M, Li Y, Jayaweera SK (2013) A survey on machine-learning techniques in cognitive
radios. IEEE Commun Surv Tutor 15(3):1136–1159 (Third Quarter)
5. Bae S, So J, Kim H (2017) Article on optimal cooperative sensing with energy detection in
cognitive radio
6. Li1 Z, Wu W, Liu X, Qi P (2018) Improved cooperative spectrum sensing model based on
machine learning for cognitive radio networks. J Inst Eng Technol 2485–2492. ISSN 1751–8628
Machine Learning Approach in Cooperative Spectrum … 71

7. Lu Y, Zhu P, Wang D, Fattouche M (2016) Machine learning techniques with probability vector
for cooperative spectrum sensing in cognitive radio networks. In: IEEE wireless conference
and networking conference track 1: PHY and fundamentals, pp 1–6 (WCNC)
8. Ma X, Ning S, Liu X, Kuang H, Hong Y (2018) Cooperative spectrum sensing using extreme
learning machine for cognitive radio networks with multiple primary users. In: 2018 IEEE 3rd
advanced information technology, electronics and automation control conference, pp 536–540
9. Awe OP, Lambotharan S (2015) Cooperative spectrum sensing in cognitive radio networks
using multi-class support vector machine algorithms. In: IEEE 2015, pp 1-7
10. Arthy A, Periyasamy P (2015) Review on spectrum sensing techniques in cognitive radio
network. In: Proceedings of UGC ponsored national conference on advanced networking and
applications, pp 80–83
11. Hsu C, Lin C (2002) A comparison of methods for multiclass support vector machines. Neural
Netw IEEE Trans 13(2):415–425
12. Allwein E, Schapire R, Singer Y (2001) Reducing multiclass to binary: a unifying approach
for margin classifiers. J Mach Learn 1:113–141
13. Yucek T, Arslan H (2009) IEEE communications surveys & tutorials
Study on Detection of Leukemia
in Human Blood Pattern Based
on Microscopic Photographs Using
Image Processing Techniques

Swagat M. Karve, Pravin Kishrsagar, Akshay A. Jadhav,

and M. Aravind Kumar

Abstract For the time being, blood disorders are defined by examination of micro-
scopic images of blood cells. It may lead to identity the blood disorders class of
blood-associated diseases. This paper discusses an initial look at the creation of
Microscopic Blood Pattern Images to detect leukemia forms. Pixel reading can be
very critical since it is possible to detect and classify pixel diseases at an earlier stage.
From there, more steps, such as disease control, surveillance and prevention, can be
completed. Photos are used because they can be fairly priced and do not require
costly check-out and laboratory equipment. White blood cell deficiency, leukemia,
will be identified by the gadget. The computer can use microscopic image capabilities
to look at texture, geometry, colour and statistical evaluation changes. Changes in
these characteristics may be used as feedback to a classifier. A literature review has
been undertaken and it is suggested to categories types of leukemia by reinforcement
learning to know. A brief debate about the problems involved has also been prepared
by researchers.

Keywords White blood cell · Microscopic images · Leukemia · Reinforcement

learning · ALL

1 Introduction

Over the past decade, scientific imaging has proved to be one of the most vital
visualisation and analysis techniques in biology and medicine. A remarkable creation
of new, powerful devices for the identification, storage, transmission, reading and
display of scientific images has occurred at this period. This has contributed to a

S. M. Karve (B) · A. A. Jadhav

SVERIs College of Engineering, Pandharpur, Maharashtra, India
e-mail: [email protected]
P. Kishrsagar
AVN Institute of Engineering & Technology, Hyderabad, India
M. Aravind Kumar
GVVRIT, Bhimavaram, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 73

P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_8
74 S. M. Karve et al.

substantial increase in the utility of virtual photo processing strategies [1] for clinical
problem solving. In advisement of System used for medical field, the toughest part of
medical imaging lies. A close interdisciplinary partnership between physicians and
engineers involves the design, implementation and evaluation of complex clinical
frameworks. Reading via pix is most importantly aimed at accumulating information,
detecting diseases, analyzing diseases, manipulating and remedying, monitoring and
evaluating [2]. Meanwhile, blood disorders are classified by visual examination of
microscopic blood cell snap shots. It could contribute to a group of safe blood-
associated diseases from the detection of blood disorders. Most cancers are one of
the most feared via The Human Disease. Leukemia is a form of blood cancer, and it
can result in death if it is discovered past due. Leukemia takes place when the bone
marrow contains a number of strange white blood cells. If there are lots of exceptional
white blood cells, the blood system’s equilibrium can be interrupted. When the blood
sample is obtained and analyzed by haematologists, the life of ordinary blood may
be identified. Hematologists must visually examine the microscopic pix, and the
procedure is time consuming and tiring [3–5].
The method needs human professional and sensitive to errors due to emotion
Disruption and Course’s human physical capacity have its own limit. In addition, it is
difficult to obtain a steady impact from visible inspection [3]. For additional research,
visual inspection may also have the simplest qualitative effects [3]. Research shows
that all data is used for maximum of recent techniques Approximately blood for e.G.
Red blood cell variety, degree of haemoglobin, stage of hematocrit, mean corpuscle
extent and much greater as the parameter for classifying diseases like thalassaemia,
most cancers, and many others. The early and fast detection of the type of leukemia
greatly facilitates the presentation of the correct treatment for the precise form of
leukemia [6]. The diagnostic strategies currently used depend on the reading of
Immunephenotyping, Fluorescence In Situ Hybridization (FISH), DanCytochem-
istry Cytogenetic Assessment [6, 7]. State-of-the-art and high-priced laboratories are
required on the way to run the diagnostic techniques and a high ratio of misidentifica-
tion was mentioned; as suggested in [8]: “The first-rate laboratories rely on as many
as a dozen distinct, hard-work-intensive technologies, all of which require relatively
educated professionals.” Having said that in relation to the subtype, maybe 50% of
sufferers are misdiagnosed [6]. With this unit, higher images can be processed, study
time reduced, the impact of subjective elements removed and the accuracy of the
identity system at the same time increased [9]. The leukaemia category and inspec-
tion will be focused on the texture, form, length, coloration and white blood cell
statistical analysis. This research is hoped to help boom productivity globally and
can benefit and be a huge contribution in the field of medical and trend popularity at
the same time. The primary aim is to decorate algorithms that can extract information
from human blood Where human blood is the primary source for detecting diseases
at an earlier stage and can save it quickly [10]. This system should be powerful for
the diversity that exists among individuals, protocols for pattern selection, time and
many others [10]. In this article, in classifying leukaemia types, we will advise on
the use of encouragement to learn (RL). As in [11], a variety of medical picture
problems can be solved by Reinforcement Mastering. Among the objects concerned,
Study on Detection of Leukemia in Human Blood … 75

clinical photographs have a very similar degree of grey and texture. Additionally,
segmentation blunders can appear and boom. Any other issue could be the lack of
a sufficient number of samples of education if a Supervised Mastering Method is
hired. Background. Blood is the main source of records that provide an indicator of
fitness changes and specific disease improvements. Adjustments to the amount or
presence of materials that have been produced can direct a person’s health situation.

1.1 Leukemia

Most blood cells are formed from cells called stem cells within the bone marrow. A
soft material contained in the center of any bone can be the bone marrow. Stem cells
are going to expand and become some pretty blood cells. Each blood group has a
role of their own. Blood components are made up of:
(a) Red blood cells (erythrocytes)—It humps oxygen to the tissues and carbon
dioxide back to the lungs.
(b) White blood cells (leukocytes)—To Contaminate-Safeguard the organism.
Several forms of white blood cells are available.
(c) Platelets—helps control bleeding via blood coagulation.
(d) Plasma—The dissolved ion-containing fluid in the blood needed for cell
activity, consisting of sodium, potassium, chloride, hydrogen, magnesium and
iron. The cells will die when the blood cells are old or damaged and born cells
will replace them [12].
Figure 1 demonstrates how damaged cell got replaced with new cells and grown
into multiple blood components. They grow into either somatic myeloid cells or
somatic lymphoid cells. Myeloid stem cells have now matured and become myeloid
blast cells. This explosion would produce a red blood corpuscle, a platelet, and a
number of other white blood corpuscles. Lymphoid stem cells may also develop and
can form a lymphoid blast, and many kinds of white blood cells may ultimately form
this blast. The lymphoid blast differs from the white blood cells in the myeloid blast.
Since the disease is dangerous and may cause death, the research would specialize in
leukaemia. Bone marrow produces irregular white blood corpuscles for someone who
has leukaemia. Abnormal white blood cells will not die until they have to, contrary
to normal cells. Therefore, there are several irregular white blood corpuscles that
conflict with normal white blood cells in order to perform their functions. It constructs
a blood system imbalance within the human body. Leukemia is also clustered to
promote the rapid progression and severity of this disease. Leukemia is either acute
or chronic.
(a) Leukemia-Leukemic cells may perform tasks like normal white blood cells at
an earlier level. They will eventually develop into serious leukaemia.
(b) Leukemia—Leukaemia cells are unable to perform functions such as normal
white blood cells. Leukemia cell count can develop rapidly and become serious
76 S. M. Karve et al.

Fig. 1 Production of blood

cell [12]

within a short period of time. Typically, leukaemia is also classified into four
types [8].
(c) Acute leukemia (ALL)—In children aged 2–10 years, acute leukaemia (ALL)
typically occurs. This is the most common form of leukaemia. In adults, they
still appear.
(d) Figure 2 shows acute chronic myelocytic leukaemia (AML) is common in
children under 1 year of age with this form of leukaemia. It’s very unusual for
adolescents. Yet it’s mostly in Adults who are 40 years of age.

Fig. 2 Acute lymphocytis

Leukemia (All) [8]
Study on Detection of Leukemia in Human Blood … 77

1.2 Blood Cell Research

Some study has been wiped out to automate the process of detection of blood
corpuscles and the patient can be correctly diagnosed next. A lot of them are [13]
to establish a method for the identification and classification of plasmodium by
microscopic images of blood cells. They use the morphology approach, so the key
criteria for improving this technique are the best methods for segmenting images
of blood corpuscles [14]. Identify patients with thalassemia using genetic program-
ming and neural networks. Instead of microscopic pictures, matured blood corpuscle,
platelet and reticulocyte data is used for eg cell percentage, hemoglobin level, hema-
tocrit level, mean volume corpuscle, distribution of hemoglobin distance, etc. to
spot patients with thalassemia, thalassemia characteristics, and usual. The multi-
layer perceptron (MLP) with 2 hidden layers is the result obtained. Training data
outcomes are 99.27% maximum accuracy, 98.16% mean accuracy, and 0.64% vari-
ance. Although the results of data testing are 88.72% overall accuracy, 84.44% mean
accuracy and 2.41% variation. While [4] a method called Leuko is being developed
and textural information is used to increase differences between leukocytes. They
used textural parameters, which are energy, inertia, homogeneity and correlation,
based on grey level occurrence matrices (GLCM). In designing the Leuko method,
Due to tedious process of features selection data reductioncan be achieved in the
view of learning the classification algorithms quickly and accurately. Classifier can
help standardize from available data, it is easier to interpret results as well as reduce
the time [15]. A device was then developed to identify leukemia cells by using
photographs of the bone marrow. Using the Support Vector Machine (SVM) clas-
sifier, the framework was designed to exploit features related to texture, geometry
and statistical analysis in blood cell images. In order to get the best recognition, they
stress on producing and selecting characteristics. They only pick 30 best features
and this generates a 13.07% training data error and 18.71% testing data errors. In
medical imaging, there are many applications that use reinforcement learning [11]
and the In their work, Kubota et al. [16] have used reinforcement learning (RL). They
use RL in order to Beat the medical photos with some issues. Among the objects
concerned, medical images have a very similar degree of grey and texture. Errors
in segmentation can occur and increase. If a supervised learning technique is used,
another concern may also be the lack of a sufficient number of coaching samples. A
minimum training dataset is required using the RL method [11]. Using Q-learning to
segment images from computed tomography (CT). They use images of cranial CT.
They noticed that they were ready to segment a picture into certain distinct regions
at the same time. The images are split into multiple sub images. An intervention
for sub images to differ and change the Qmatrix is selected by the Reinforcement
Learning (RL) agent. There is an evaluator who compares the outcomes and gives a
gift to the RL agent. The precision of segmentation that they achieve is above 95%
[16]. Apply the RL method to segmental prostate ultrasound images. They still use
Q-learning and the findings indicate a high potential for reinforcement learning to
be implemented in medical segmentation. By using an RL agent, their technique is
78 S. M. Karve et al.

to control the local threshold and hence the post-processing parameter [17]. Extract
the area of the kidney as a preprocessor for the diagnosis of renal disease. They use
X-ray CT abdominal images. Inside the rough kidney area, Q-learning is used and
the edge of the kidney contour is detected. However, for an actual contour, there are
a few error margins and it is corrected by the snake process. The chance of success
is a sort of that’s a poor 53%, another application for RL is used by [18].

2 Research Methodology

Research methodology and steps which will be utilized in this research includes
(Fig. 3):
1. Image Acquisition: With efficient magnification, blood images from slides will
be collected from the nearby hospital.
2. Preprocessing: The photos will be acquired during image processing and unnec-
essary staining during you’ll be distracted by noise. As our ROI would be white
blood cells, the history is going to be removed. Image enhancement will be
performed during this preprocess because the contrast enhancement technique
is capable of enhancing the quality of the medical image [19].
3. Segmentation: White blood corpuscle (WBC) segmentation and determination
of the nucleus ROI for WBC only. This is because the cytoplasm is sparse in
leukaemia cell images [7]. So the emphasis will only be on the nucleus of the
WBC. Determination of the WBC forms from the nucleus should be completed.
Only lymphocytes and myelocytes should be considered, and whether or not
they are blast cells or not should be determined. Others should be omitted, such
as neutrophils, basophils and eosinophils.

Fig. 3 Typical steps in

process of automating blood
recognition
Study on Detection of Leukemia in Human Blood … 79

4. Feature Extraction: The most important problem in generation of features of

blood cells that characterize them during a way enabling the popularity of
various blast types with the very best accuracy [15]. For the nucleus of lympho-
cytes and myelocytes, the features to be included are: Geometric characteris-
tics, which include area, radius, perimeter, symmetry, concavity, compactness,
strength, excentricity, elongation, form factor, will be acquired. Texture charac-
teristics, which include homogeneity, energy, correlation, entropy [7], contrast,
second angular momentum, will be obtained.
5. Color Features: The RGB colour spaces will be converted into colour spaces
for HSV or L*a*b. Their average colour values will be collected. Statistical
characteristics: it is important to obtain the mean, variance, skewness, kurtosis of
the histograms of the image matrix and thus the gradient matrix for RGB or HSV
or L*a*b colour space (whatever the case may be). ALL is small based on [19],
blast cells are uniform, cytoplasm is scarce, round and usually includes single
nucleoli within the nucleus. Although the blasts are larger and irregular shape in
AML, and usually with the involvement of Auer rode, several nucleoli. Qiang
and Zhongh [20] said that although red blood tends to be rather darker than the
background, the WBC appears The (RBC) corpuscle appears at an intermediate
intensity. Cseke [21] shows that white cells are the darker components that tend
to be pale in the RBC images. Platelets are much smaller than cells that are
white and red.

3 Classifications

Classification is the process of assigning to a known class an undefined test vector.

A reinforcement learning algorithm will be proposed during this process. The RL
strategy will identify the leukaemia groups into Both, AML, CLL, and CML as in
Fig. 4 [22].
The essential RL model is used. The world receives the agent’s behavior and
changes it to a replacement state. The agent receives the current state of the world
and the environment receives a gift or punishment. The reward/punishment will be
received by the agent from the environment accepted the environmental action. The
agent gets information and learns how to react to the very best reward. Actions that

Fig. 4 Basic model of RL

80 S. M. Karve et al.

aim to expand the general amount of incentive values should also be selected. The
agent would use a strategy that we have called as a policy of action to decide on
an environmental action. Otherwise the trend would be reduced by the mechanism
generating this behavior [22]. Q-learning can be a soft strategy that suggests that its
Q-values, regardless of discovery, approximate the optimum Q-values that are Q(s,a).
There will be Q-values contained in the Q-matrix. Q(s,a) is that r, obtained by taking
action, a from the economy, s, the estimated amount of potential payoffs. The one
with the best Q value is going to be the optimum action. Q(s,a) will be revised to
support the experience as follows:
 
Q(s,a) = (1 − α)Q(s, a) + α r + γmaxQ(s , a )

where an is the rate of learning and 0 < γ < 1 is the discount factor. As a conclusion,
as in Fig. 5, the research methodology can be viewed.

Fig. 5 Proposed research

methodology [11]
Study on Detection of Leukemia in Human Blood … 81

4 Discussion

Some blood cell concerns need to be addressed in order to solve them. One of the
issues is the matter of the blood corpuscle itself. Capell [6] say that their method
fails to classify some of the blood cells in the process of classification. Thanks to
environmental strain, a number of cells are also deformed into an arbitrary shape [3].
The Notes on their algorithms that do not distinguish cells that overlap. Overlapping
cells may also be joined by disease-induced cells [23].
He argued that identification between two neighboring cells in their development
line is the most difficult problem since the cells are very close and thus the boundary
point between two neighbors is not well defined. But overlapping problems are also
solved using the watershed approaches used in their studies [6]. Another challenge
is information collection. Blood sample images should be adequate to ensure that
generalization properties are always expressed and that unseen data can be identified
correctly [6]. Lacking samples means that the info [5] can only be represented by
a few key components [24]. Believes that separate data sets are the perfect route
should be used for each point. However, cross validation or bootstrap sampling can
also be used as it is difficult to collect a sizeable amount of samples. It is hoped that
the approach of RL in the classification phase would reduce the issue of inadequate
data. Once we are to build the method, all the problems posed by scientists have to
be taken into account. By applying effective strategies, we should always strive to
resolve them.

5 Conclusion

This study includes using microscopic blood sample images to classify the forms of
leukaemia. By using features in microscopic images, the device will be constructed by
analysing changes in texture, geometry, colours and statistical analysis as a classifier
input. The system should be effective, accurate, less time interval, smaller error, and
high precision, cheaper cost and robust for individual varieties, sample collection
protocols, time and so on. Knowledge derived from microscopic blood sample images
will support individuals by rapidly predicting, resolving and treating blood disorders
for a particular patient.

References

1. Valencio CR, Tronco MN, ACB Domingos, Domingos CRB (2004) Knowledge extraction
using visualization of hemoglobin parameters to identify thalassemia. In: Proceedings of the
17th IEE symposium on computer based medical systems, pp 1–6
2. Adollah R, Mashor MY, Nasir NFM, Rosline H, Mahsin H, Adilah H (2008) Blood cell image
segmentation: a review. In: Proceedings of biomedical, vol 21, pp 141–144
82 S. M. Karve et al.

3. Ritter N, Cooper J Segmentation and border identification of cells in images of peripheral blood
smear slides. In: 30th Australasian computer science conference, conference in research and
practice in information technology, vol 62, pp 161–169
4. Sabino DMU, Costa LDF, Rizzatti EG, Zago MA (2004) A texture approach to leukocyte
recognition. Real Time Imaging 10:205–206
5. Colunga MC, Siordia OS, Maybank SJ (2009) Leukocyte recognition using EM algorithm.
MICAI, LNAI 5845. Springer Verlag, Berlin Heidelberg, pp 545–555
6. Capell K (2012) Meeting leukemia’s diagnostic challenge. Business Week, The Mc
Graw Hill Companies, London. https://www.businessweek.com/magazine/content/05_36/b39
49007_mz001.htm
7. Halim NHA, Mashor MY, Hassan R (2011) Automatic blasts counting for acute leukemia based
on blood samples. Int J Res Rev Comp Sci 2(4):971–976
8. Bergen T, Steckhan D, Wittenberg T, Zerfab T (2008) Segmentation of leukocytes and erythro-
cytes in blood smear images. In: 30th annual international IEEE EMBS conference, Vancouver,
Canada, pp 3075–3078
9. Shitong W, Min W (2006) A new detection algorithm (NDA) based on fuzzy cellular neural
networks for white blood cell detection. IEEE Trans Inf Technol Biomed 10(1):5–10
10. Shin H, Markey MK (2006) A machine learning perspective on the development of clinical
decision support system utilizing mass spectra of blood samples. J Biomed Inform 39:227–248
11. Chitsaz M, Woo CS (2011) Software agent with reinforcement learning approach for medical
image segmentation. J Comput Sci Technol 26(2):247–255
12. Bergen T, Steckhan D, Wittenberg T, Zerfab T (2008) Segmentation of leukocytes and erythro-
cytes in Blood Smear Images. In: 30th annual international IEEE EMBS conference, Vancouver,
Canada, 20–24 August, pp 3075–3078
13. May P, Ehrlich HC, Steinke T (2006) ZIB structure prediction pipeline: composing a complex
biological workflow through web services. In: Nagel WE, Walter WV, Lehner W (eds) Euro-Par
2006. LNCS, vol 4128. Springer, Heidelberg, pp 1148–1158
14. Foster I, Kesselman C (1999) The grid: blueprint for a new computing infrastructure. Morgan
Kaufmann, San Francisco
15. Piurri V, Scotti F (2004) Morphological classification of blood leukocytes by microscope
images. In: IEEE international conference on computational intelligence for measurement
systems and applications, Boston, MA, USA, pp 103–108
16. Kubota Y, Mitsukura Y, Fukumi M, Akamatsu N, Yasutomo M (2005) Automatic extraction
system of a kidney region based on the Q-learning. Springer-Verlag, Berlin Heidelberg, pp
1261–1267
17. Markiewicz T, Osowski S, Marianska B, Moszczynski L (2005) Automatic recognition of
the blood cells of myelogenous leukemia using SVM. In: Proceedings of international joint
conference on neural networks, Montreal, Canada, pp 2496–2501
18. Sadeghian F, Seman Z, Ramli AR, Kahar BHA, Saripan M (2009) A framework for white blood
cell segmentation in microscopic blood images using digital image processing. Biol Proced
Online 11(1):196–206
19. Sahba F, Tizhoosh HR, Salama MMA (2006) A reinforcement learning framework for medical
image segmentation. In: International Joint Conference on Neural Networks, Vancouver,
Canada, 16–21 July, pp 511–517
20. Qiang W, Zhongli Z (2011) Reinforcement learning, algorithms and its application. In: Inter-
national conference on mechatronic science, electric engineering and computer, Jilin, China,
pp 1143–1146
21. Cseke I (1992) A fast segmentation scheme for white blood cell images. IEEE, pp 530–533
22. Birndorf NI, Pentecost JO, Coakley JR, Spackman KA (1996) An expert system to diagnose
anemia and report results directly on hematology forms. Comput Biomed Res 29:16–26
23. Using visualization of hemoglobin parameters to identify thalassemia (1981) Proceedings of
the 17th IEE symposium on computer based medical systems, pp 1–6. M.S.: 2004 identification
of common molecular subsequences. J Mol Biol 147:195–197
Study on Detection of Leukemia in Human Blood … 83

24. Srinivisan KS, Lakshmi D, Ranganathan H, Gunasekaran N (2006) Non invasive estimation
of hemoglobin in blood using color analysis. In: 1st international conference on industrial and
information system, ICIIS 2006, Sri Lanka, 8–11 August, pp 547–549
Brain Tumor Detection Using Deep
Neural Network

Rajshree B. More and Swati. A. Bhisikar

Abstract Brain tumor identification is an essential task for assessing the tumors
and its classification based on the size of tumor. There are various types of imaging
strategies such as X-rays, MRI, CT-scan used to recognize brain tumors. Computed
Tomography (CT) scan images are used for in this work for Brain tumor Image Iden-
tification. CT-scan images are used, because as it gives size, shape and blood vessels
detailing and is non-invasive technique. CT-scan is commonly utilized because of the
superior quality of image. Deep learning (DL) is the most recent technology which
gives higher efficiency results in recognition, classification. In this paper, the model
is developed by using Convolution neural network to detect the tumor of brain image
from a dataset from Kaggle. The dataset contains near about 1000 images. Tumor is
identified by image processing algorithm using CNN, time complexity is 90 m sec,
and the accuracy of the present system is 97.87%.

Keywords Brain tumor · CT scan images · Deep neural network · Convolution

neural network · Spyder (Python 3.7)

1 Introduction

Brain Tumor can be characterized as an artificial and abnormal increase in brain cells.
The human skull is inflexible, and its size is restricted in this manner any undesired
development in the tumor may influence on human brain activity. Besides, it gets
increased into other body organs and impact human body functionality [1, 2]. The
World Health Organization (WHO), says that the brain malignant growth influences
under 2% of the human population, which causes serious complications and intrica-
cies [3]. A brain tumor can be gathered from multiple points of view, primary and
secondary tumor. The previous findings show 70% of all brain tumor are primary,

R. B. More (B) · Swati. A. Bhisikar

Department of Electronics and Telecommunication Engineering, JSPM’s Rajarshi Shahu College
of Engineering, Tathawade, Pune, India
Swati. A. Bhisikar
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 85

P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_9
86 R. B. More and Swati. A. Bhisikar

while secondary (optional) tumor are the residuals 30%. This categorization of the
primary tumor first originates in the brain at the early stage. On the contrary, tumors
that first emerge in some other pieces of the body and afterwards moved to the cere-
brum are called optional tumors, and most of them are dangerous [4]. A brain tumor
is an aggregation of irregular cells in the brain. Any development inside the skull,
such a restricted space can cause issues. Brain tumors can be harmful (malignant) or
non-harmful (benign). At the point when the harmful tumor is developed, it increases
the pressure inside the skull. This may cause brain damage, and life expectancy is
very low [2]. Various imaging strategies can be utilized to distinguish and arrange a
brain tumor. In imaging modalities, MRI is widely recognized non-invasive methods
[5–7]. CT scan shows a tumor shape, size and location, blood vessels that feed the
tumor. X-ray images show different types of body parts in different shades of black
and white; the X-ray images are widely used to checking fractures (broken bones)
[8].
At an early stage, tumor detection is done by using image processing provides a
fundamental understanding of the presence of tumor and its type which is advanta-
geous for the treatment strategy and diagnosis to save patients life [9]. Moreover, the
detection of a stage of tumor is a difficult task for radiologists and physicians. This
needs specialists to detect tumor by comparing tumor tissues and by areas, to make
Images clear for human vision, image processing filters are used. Thus, it is a require-
ment for Computer-Aided Diagnosis (CAD) framework to early diagnose brain tumor
in reasonably less time to reduce radiologist burden [2]. Machine Learning (ML) is the
science that allows software applications to predict more accurate outcome depend
on training without being specifically programmed [10]. ML algorithms play a vital
role in the medical imaging field as a part of artificial intelligence [11]. Supervised
and unsupervised learning are two types of ML. In supervised learning techniques,
the outcome is based on input variables it analyses the training data and produces
output from the labelled data, in unsupervised learning, the data is not labelled, and
the algorithm is applied on input data without any prior training of data. Artificial
Neural Network (ANN) is based on a collection of connected artificial neurons. It
is used to simulate input database internal weighting system [12], Support Vector
Machine (SVM) is used for characterization problem it uses the kernel to transform
input data based on these transformations it gives optimum boundary between the
possible outputs, and K-Nearest Neighbors (KNN) is the non-parametric approach
used for the classification. Classification is based on nearest neighbour rule [13]. In
comparison, the outcome of the unsupervised technique is dependent only on the
input variables [14] and Self-Organization Map (SOM) [15]. For segmentation of
tumor the image is to be converted into grayscale used this information for detection
of tumor. From the handcrafted features, meaningful information can be extracted
by the expert in the medical field and which requires a considerable time and may
cause an error while handling large database [16].
Feature learning and accuracy is achieved by increasing training samples instead
of machine learning and neural systems which are the main advantages of Machine
learning which promote extremely powerful and exact model [17, 18]. The approach
Brain Tumor Detection Using Deep Neural Network 87

to train and developed the classification of brain tumor using CNN architecture was
developed connected convolution layers and achieved an accuracy of 91.43% [19].
In this paper, the identification of brain tumor is obtained using CNN classifier
the paper is focused on pre-processing, segmentation and feature extraction. For
simulation Python programming language is used. It is used for general purpose,
and it is a high-level programming language, and it’s more efficient than any other
languages. Accuracy of the proposed method is of 97.87%.

2 Methodology

Figure 1 shows the block diagram for Brain tumor identification using Convolution
Neural Network (CNN). Convolution Neural Network (CNN) is used in Medical
image processing field [2]. A four-layer CNN is introduced for tumor detection, as
shown in Fig. 1. Following is the block diagram with an explanation.

2.1 Pre-processing

There is a lot of non-essential noise in the image, so we need to improve the quality of
the images in pre-processing ways. The main function of the processing is to improve
the single-volume ratio, improve image appearance, remove noise and undesirable
backgrounds, smoothing the interior parts and keeping the edges.

Fig. 1 Block diagram of brain tumor identification

88 R. B. More and Swati. A. Bhisikar

Fig. 2 CNN architecture

2.2 Extraction of Features

Extraction of features is the method of gathering higher-level image details such as

structure, texture, color and contrast. Analysis of textures is an important parameter
for human visual perception and machine learning. Texture finding and analysis at
different stages of tumor detection can improve the diagnosis. It is used to enhance
diagnostic system accuracy effectively by selecting statistical characteristics such as
mean, contrast, energy, entropy, standard deviation, kurtosis, etc.

2.3 CNN Architecture

The convolution neural network contains inputs and output layers, as well as many
hidden layers. CNN’s hidden layers usually include a series of convolution layers
that convince duplication or other dot product. Four types of layers available on CNN
as represented in Fig. 2, layers are
(1) Input layer, (2) Convolution Layer, (3) Pooling Layer and (4) Fully Connected
layer.
Input Layer: The input of the network is a 2 Dimensional digital image. The
network has an input layer which takes the image as the input.
Convolution layer: These are the most important blocks used in convolution
neural networks. Convolution is a filter applied to an input that results in an activation
function. Neural Networks consists of a series of hidden layers that transform the
input image. Every layer is a collection of a set of neurons, where each layer is fully
connected to all neurons in the next layer.
Pooling layer: Pooling is the process of extracting the features from images output
of a convolution layer. Maximum pooling operation calculates the maximum value
in every patch of the feature map.
Fully connected layer: It forms the last few layers in a neural network. The input
to the fully connected layer is the output from the final Pooling layer. This process
determines the most accurate weights. The output Eq. 1 below,
Brain Tumor Detection Using Deep Neural Network 89

y1 = x1ω1 + x2ω2 + · · · xnωn (1)

The primary CNN shown in Fig. 2. CNN advantage is it performs well, and it
gives better accuracy. It is covering local and global features. It also learns different
features from images. In algorithm-based image classification, we need to select the
features (local, global) and classifiers. Other learning algorithms or models can also
be used for image identifications and classification.
Convolution Neural Network is used in object recognition and biological image
segmentation which enhances the performance of deep learning methods. In tradi-
tional methods, features are to be extracted first and then these features are fed
into the training model for the desired result of classification. CNN’s automatically
understands complex characteristic features directly from the data given.
Thus, in this paper, the main focus is on CNN architecture design instead of image
processing for extracting features. CNN takes patch image from the entire image as
input and uses convolution filters which are flexible and trainable; also, it makes use
of local subsampling to extract features.
Figure 3 shows the four-layer CNN model gives us a remarkable result for the
detection of the tumor. We observe the performance in two ways based on the sepa-
ration of the dataset. First, we tried to partition data in 70:30 formats that means 70%
data goes to the training and remaining data goes for testing by using this splitting
ratio, the accuracy is 97.5% [20]. And secondly, 80% of the images are used for
training, and some of the images are given for testing where we get 97.87% accuracy
and 98.47% training accuracy. In line with these lines, the proposed model provides
a positive effect on a scale of 80:20. Table 1 shows the above details in the form of
a table.

Fig. 3 A convolution
network architecture

Table 1 Performance of the proposed CNN model

Sr no Total images Training image Testing image Splitting ratio Accuracy (%)
1 1000 700 300 70:30 92.98
2 1000 800 200 80:20 97.87
90 R. B. More and Swati. A. Bhisikar

Table 2 Performance
Methodology Accuracy (%)
Comparison
Seeta et al. [20] 97.5
Hussain [17] 96.87
Proposed CNN model 98.87

Related to testing the performance, as shown in Table 2 of our model database,

we used the database in the Brain Tumor diagnostic field, and that was taken from
Kaggle. It consists of two classes: phase 0 and Phase 1. Class-0 represents non-tumor
images, and class 1 represents Tumor images. 800 and 200 images containing tumor
and non-tumor respectively classified as class 1 and class-0. All images are CT scan
images. We obtained a positive result by dividing the 80 to 20 datasets by going to
a CNN image test.

3 Flowchart

Figure 4 shows the complete process of Image identification of Brain tumor; CT

digital images from the database are loaded and labelled. Further pre-processing,
validation is done on test sets. The database images are split into training, testing
and validation. The digital brain tumor images are given as input to the optimiza-
tion algorithm. Finally, network training is done, and performance computations are
presented.

Fig. 4 Process of image

identification of brain tumor
Brain Tumor Detection Using Deep Neural Network 91

4 Results

In this work, CNN is used, which provides image data classification. The reason why
CNN is best is CNN doesn’t require feature extraction because CNN layers extract
features from an image by itself. CT scan image data set, which is made available
from Kaggle open-source dataset portal.
The system captures the brain images and verifies the brain tumor is present or
not in that particular brain image. Then it will find a brain tumor is present or not.
Figure 5 shows the GUI system of the complete process, in that process first we
Train the model using CNN algorithm after completing the training process select
the image of brain tumor and then press the prediction button. Finally, the system
Display the result in that result display a message whether the tumor is detected or
not.
Figure 6 shows the system GUI. There are 3 button search buttons have a different
functionality and Image Process windows for image preprocessing technique. In that
GUI first train the model after completing training procedure, select brain tumor
image and after that click the prediction button then we get results on above image
Brain tumor is not detected.
Figure 7 shows the result of the proposed system. It shows the detected tumor,
and Fig. 8 shows the non-tumor image detection.
Train Here CNN Module (Fig. 9) will be training with an accuracy of 98.87%.

Fig. 5 GUI of system & image process window

Fig. 6 Tumor detection

92 R. B. More and Swati. A. Bhisikar

Y- Y- Y-

Fig. 7 Tumor image detection

N- N- N-

Fig. 8 Non-tumor image detection

Fig. 9 CNN module

5 Conclusion

The main aim of this research project is to design an effective system for identifying
brain tumors with high accuracy, performance and low complexity. This paper shows
tumor identification. The Kaggle database contains tumor and non-tumor CT scans
Brain Tumor Detection Using Deep Neural Network 93

images. In this work, brain tumor detection is accomplished through a convolution 4-

layer neural network. Editing is done in the python language. The accuracy compared
to other methods, such as the Support Vector Machine (SVM) is 100%. Processing
time is high. For the proposed CNN based on identification and recognition, the
accuracy of the detection is 95%, and the processing time is 90 ms.

6 Future Scope

(a) By using Web Application patient can upload their brain image and get the
result, whether the tumor is present or not.
(b) By developing Mobile Application it’s easy to use them for the evaluation for
all other Kind of diseases.
(c) In Feature, this technique can be developed to classify the tumors based on
Feature Extraction.

References

1. DeAngelis LM (2001) Brain tumors. N Engl J Med 344(2):114–123

2. Sultan HH, Salem NM, Al-Atabany W (2019) Multi-classification of brain tumor images using
deep neural network. Special section on Deep Learning for computer Aided diagnosis. IEEE
Access 7. https://doi.org/10.1109/access.2019.2919122
3. Stewart BW, Wild CP (2014) World cancer report 2014. IARC, Lyon, France
4. Brain, Other CNS and Intra cranial Tumors Statistics. Accessed: May 2019. [Online]. Available:
https://www.cancerresearchuk.org/
5. Hunnur SS, Raut A, Kulkarni S (2017) Implementation of image processing for detec-
tion of brain tumors. In: 2017 international conference on computing methodologies and
communication (ICCMC), Erode, pp 717–722. https://doi.org/10.1109/iccmc.2017.8282559
6. Gurbină M, Lascu M, Lascu D (2019) Tumor detection and classification of MRI brain image
using different wavelet transforms and support vector machines. In: 42nd international confer-
ence on telecommunications and signal processing (TSP), Budapest, Hungary, pp 505–508.
https://doi.org/10.1109/tsp.2019.8769040
7. Parveen, Singh A (2015) Detection of brain tumor in MRI images, using combination of fuzzy
c-means and SVM. In: 2015 2nd international conference on signal processing a integrated
networks (SPIN), Noida, pp 98–102. https://doi.org/10.1109/spin.2015.7095308
8. Behin A, Hoang-Xuan K, Carpentier AF, Delattre J-Y (2003) Primary brain tumors in adults.
Lancet 361(9354):323–331
9. Litjens G, Kooi T, Bejnordi BE, Setio AAA, Ciompi F, Ghafoorian M, van der Laak JAWM,
van Ginneken B, Sánchez CI (2017) A survey on deep learning in medical image analysis. Med
Image Anal 42:60–88
10. Gooden berger ML, Jenkins RB (2012) Genetic so fadult glioma. Cancer Genet 205(12):613–
621
11. Drevelegas A (2002) Imaging of brain tumors with histological correlations. Springer, Berlin,
Germany
94 R. B. More and Swati. A. Bhisikar

12. Louis DN, Perry A, Reifenberger G, von Deimling A, Figarella-Branger D, Cavenee WK,
Ohgaki H, Wiestler OD, Kleihues P, Ellison DW (2016) The 2016 world health organiza-
tion classification of tumors of the central nervous system: a summary. Acta Neuropathol
131(6):803–820
13. Bishop C (2006) Pattern recognition and machine learning. Springer-Verlag, Berlin, Germany
14. Machhale K, Nandpuru HB, Kapur V, Kosta L (2015) MRI brain cancer classification using
hybrid classifier (SVM-KNN). In: Proceedings of international conference on industrial
instrumentation and control (ICIC), pp 60–65
15. Shasidhar M, Raja VS, Kumar BV (2011) MRI brain image segmentation using modified fuzzy
C-means clustering algorithm. In: Proceedings of international conference on communication
systems and network technologies (CSNT), pp 473–478
16. Goswamiand S, Bhaiya LKP (2013) Brain tumour detection using unsupervised learning
based neural network. In Proceedings of international conference on communication
systems and network technologies (CSNT), pp 573–577
17. Hossain T, Shishir FS, Ashraf M, Al Nasim MDA, Shah FM (2019) Brain tumor detection using
convolutional neural network. In: 2019 1st international conference on advances in science,
engineering and robotics technology (ICASERT)
18. Siar M, Teshnehlab M (2019) Brain tumor detection using deep neural network and machine
learning algorithm. In: 2019 9th international conference on computer and knowledge engi-
neering (ICCKE), Mashhad, Iran, 2019, pp 363–368. https://doi.org/10.1109/iccke48569.2019.
8964846
19. Paul J (2017) Deep learning for brain tumor classification. In: 13 March 2017 SPIE 10137,
Medical imaging 2017: biomedical applications in molecular, structural, and functional
imaging, 1013710. doi: https://doi.org/10.1117/12.2254195
20. Seetha J, Selva kumar Raja S (2018) Brain tumor classification using convolutional neural
networks. Biomed Pharmacol J 11:1457–1461. https://doi.org/10.13005/bpj/1511
Design and Simulation of Different
Structures of Micro Strip Patch Antenna
for Wireless Applications

Anil. J. Kokare, Mahesh. S. Mathpati, and Bhagyashri. S. Patil

Abstract The need for multiband, bigger addition and low profile radio wires to help
numerous remote applications prompted the plan of Microstrip reception apparatuses.
Microstrip radio wires because of their little profile configuration take less zone. This
paper presents a straightforward rectangular Microstrip Patch Antenna, E-Shaped,
U-Shaped, +-Shaped radio wires work at 2.2 to 3.8 GHz. The Proposed reception
apparatus will be in lightweight, keen and conservative unit contrast and comprises
of metallic fix and ground between which is a dielectric medium called the substrate.
This various structures of MSA are utilized for military, remote and common appli-
cations. The CADFEKO programming is utilized to register the increase, power,
radiation example and S11 of receiving wire.

Keywords Microstrip · Fractal · Multiband · WLAN · LTE · CADFEKO

1 Introduction

Taking into account the advancement of the ongoing remote correspondence frame-
works and its application, more extensive data transfer capacity, multiband and low
profile radio wires are in incredible interest for both business and military applica-
tions. The quick increment of remote interchanges prompts an enormous interest in
the planning of a multiband radio wire. Expectedly, every reception apparatus works
at single or double recurrence groups, where distinctive receiving wire is utilized for
various applications. The plan of proposed receiving wires is utilized for the fast,

Anil. J. Kokare (B) · Bhagyashri. S. Patil

Electronics & Telecommunication Department, SMSMPITR, Akluj, Maharashtra, India
Mahesh. S. Mathpati · Bhagyashri. S. Patil
GNDEC, Bidar, India
Mahesh. S. Mathpati
Electronics & Telecommunication Department, SVERI’s College of Engineering, Pandharpur,
Maharashtra, India
Bhagyashri. S. Patil
Visvesvaraya Technological University, Jnana Sangama, Belagavi, Karnataka, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 95

P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_10
96 Anil. J. Kokare et al.

versatile correspondence and furthermore advancement of microwave frameworks,

for example, WLANs, WiMAX alongside the conveyance of rapid information.
Various receiving wire plans, for example, E-formed radio wires have been introduced
for such multi-standard portable terminal application. A tale GPS fix reception appa-
ratus with fractal EBG structure utilizing an attractive natural substrate is introduced,
which can meet the prerequisites of scaling down and elite of GPS [1]. Hexagonal
fractal receiving wire configuration begins with the first emphasis of square fix and
is portioned by eliminating the centre square from it. For the second emphasis, the
square is cut into nine harmonious sub squares by 3-by-3 evaluation, and the focal
sub-square is eliminated. A similar system is then applied recursively to the excess
eight sun squares, and for the third emphasis again, we take 33% of second sub
squares [2]. An enhanced and minimized printed double band fractal reception appa-
ratus reasonable for WLAN applications. The proposed reception apparatus takes a
shot at 2.4 and 5.2 GHz [3]. A composite scaled-down fractal radio wire as a mix of
Minkowski and Koch bends is introduced. The structure of the proposed reception
apparatus is after the effect of the alterations made with the essential fractal square
and three-sided bends. The radio wire can be utilized for most handheld gadgets
and subsequently finds wide applications in the field of remote and portable appli-
cations [4]. A test took care of E-formed fractal fix reception apparatus (EFPA)
for heptads band LTE/WWAN (GSM850/900/1800/1900/UMTS/LTE2300/2500)
activity is proposed, and different cycles of this fractal receiving wire are looked
at, and an improved plan is introduced.

2 About Fractal Geometries

The term fractal was begotten by the French mathematician B.B. Mandelbrot during
the 1970s after his spearheading research on a few normally happening sporadic
and divided calculations not contained inside the domains of customary Euclidian
math [5]. The term has its underlying foundations in the Latin word fractus, which
is identified with the action word finger (which means: to break). These calculations
were commonly disposed of as amorphous; however, Mandelbrot found that specific
uncommon highlights can be related to them. He found a typical component in huge
numbers of these apparently sporadic calculations and figured speculations depen-
dent on his discoveries [6]. Two instances of normally happening fractal calculations
are snow-chips and limit of geographic mainlands. The fractal receiving wires are
not the same as customary radio wires since it is fit for working at various frequencies
all at once. The vital favourable circumstances of fractal radio wires are diminished
reception apparatus size, uphold multiband and wideband activity with improved
receiving wire execution. These can be accomplished utilizing fractal calculation
like Hilbert, Sierpinski, Koch and Minkowski are the different kinds of fractal calcu-
lations [7]. All these fractal calculations are utilized to plan a little size multiband
and wideband reception apparatuses.
Design and Simulation of Different Structures … 97

3 Methodology

The patch antenna has been designed for the following dimensions to achieve the
requirements of wireless communication applications.
Resonant frequency (fr ) = 2.4–4.8 GHz, Dielecrtric constant (ε) = 4.4.

3.1 Rectangular and E-Shaped Microstrip Patch Antenna

The radio wire is planned at 2.4 GHz recurrence and built as a fix on the substrate.
Model number one in Fig. 1 shows a rectangular fix receiving wire configuration to
work near 3.8 GHz will be demonstrated. The model is first developed as a fix on
an infinitely huge substrate since it rushes to make and to recreate. Figure 2 shows
an E-formed fix radio wire plan fractal math is in gendered in an iterative style,
prompting self-comparative structure.

Fig. 1 Rectangular patch

antenna

Fig. 2 E-shaped patch

antenna
98 Anil. J. Kokare et al.

3.2 U-Shaped and Plus Shaped MSA

Figure 3 shows Plus-formed structure. This is a reference receiving wire or base

shape reception apparatus. Further, this base shape radio wire is adjusted by embed-
ding’s level spaces on the two sides with the separate focus of fix can be utilized in
remote correspondence applications. Figure 4 shows U-molded structure can work
at 2.5 GHz for some applications in ongoing remote correspondence.

Fig. 3 Plus-shaped patch

antenna

Fig. 4 U-shaped patch

antenna
Design and Simulation of Different Structures … 99

Table 1 Comparison of different structures of microstrip patch antennas

Parameters Rectangular Plus shape U shape E shape
Resonant frequency (GHz) 2.7 3.8 2.5 2.4
Reflection coefficient (dB) −40 −36 −38 −11
VSWR 1.5 1.83 2 1.85
Impedance (ohm) 50 48 49.5 51.2
Return loss (dB) −22 −23 −22 −18

4 Simulation Results

Table 1 shows the comparison of different shapes of the microstrip antenna.

For the examination of proposed Different shape of microstrip patch fractal recep-
tion apparatus, the receiving wire boundaries like reflection coefficient, VSWR,
addition and transmission capacity are recreated utilizing reproduction program-
ming CADFEKO. CADFEKO is a full-wave electromagnetic field test system that
depends on the Method of Moments (MoM). It is a business programming device that
can be utilized for receiving wire plan, radio wire situation examination, RF struc-
ture execution forecast, EMC just as dissipating issues and bio-electromagnetics.
The initial four cycles of the middle feed E shape fractal reception apparatus are
mimicked by utilizing CADFEKO programming, and results are demonstrated as
follows.
The Reflection coefficient for the of the E shape fractal receiving wire is plotted
in Fig. 5. Reflection coefficient esteems—11 dB, at particular thunderous frequen-
cies 2.4 GHz. The E formed fractal radio wire creates a low reflection misfortune
contrasted with the standard qualities needed for portable application at GSM band.
Impedance and VSWR plots are shown in Figs. 6 and 7 with the appropriate value
required for wireless application. The antenna radiation pattern is plotted in a 3D
view, as shown in Fig. 8.

5 Conclusion

It would have been a major bit of leeway to know the CADFEKO 14.0 reproduction
programming already as a lot of estimations might have been applied. Miniature strip
radio wires and their hypothesis get substantially more perplexing as you need to
make more proficiency. The various sorts of reception apparatuses are explored and
effectively reproduced in this paper. The reproduced reflection coefficient, impedance
and radiation design demonstrated well execution. Miniature strip reception appa-
ratuses have become a quickly developing territory of examination. Their poten-
tial applications are boundless, on account of their lightweight, conservative size,
and simplicity of assembling. E-formed micro strip reception apparatus is entirely
100 Anil. J. Kokare et al.

Fig. 5 Reflection coefficient

of the patch antenna

Fig. 6 Impedance of patch

antenna
Design and Simulation of Different Structures … 101

Fig. 7 VSWR of E-shaped patch antenna

Fig. 8 E-shaped patch

antenna radiation

planned with an expanded transmission capacity as compared to rectangular minia-

ture strip receiving wire antenna. Comparison of different shapes of the microstrip
antenna is simulated and presented.

References

1. Ranjan A, Singh M, Sharma MK, Singh N (2015) Analysis and simulation of fractal antenna
for mobile Wi-max. IEEE Trans Future Gener Commun Netw 7(2):57–64
2. Jannani A (2011) Design of E-shape fractal simple multiband patch antenna for S-band LTE
and various mobile standards. IEEE Trans Antennas Propag 8:126119–126126. https://doi.org/
10.1109/access.2020.3006831
102 Anil. J. Kokare et al.

3. Permana CG, Munir A (2011) Printed multiband antenna for mobile and wireless communica-
tions. IEEE Trans Antennas Propag 236–240
4. Sreelaxmi K & Mithun Megha Sarkar TP (2014) Multiband miniaturized fractal antenna for
mobile communications. Int J Res Eng Technol (IMPACT: IJRET) ISSN(E): 2321–8843;
ISSN(P): 2347–4599 2(4)
5. Khidre et al (2012) Presented U slot micro-strip antenna for higher mode applications. IEEE
Access Digit Object Identifier 8:112840–112845. https://doi.org/10.1109/access.2020.3002789
6. Dhillon SS, Marwaha A, Nagpal A (2013) Multiband E-shaped fractal microstrip patch
antenna with DGS for wireless applications. In: Proceedings of international conference on
computational intelligence and communication networks, Mathura, pp 22–26
7. Lotfi P, Azarmanesh M, Soltani S, Bayatmaku N (2016) Design of simple multiband patch
antenna for mobile communication applications using new E-shape fractal. IEEE Antennas
Wirel Propag Lett 10
State Context and Hierarchical Trust
Management in WSN for Intrusion
Detection

Ranjeet B. Kagade and J. Santhosh

Abstract Wireless sensor network is defined as homogeneous or heterogeneous

system containing large number of sensors namely called as nodes used to monitor
different environments in cooperatives. WSN is composed of sensor nodes (SN),
base stations (BS) and cluster head (CH). The popularity of wireless sensor network
has been increased day by day exponentially due to its wide range of application. The
applications of wireless sensor networks are air traffic control, healthcare systems,
home services, military services, industrial & building automations, network commu-
nications, VAN etc. The advantage of WSN is that it is very easy to install in critical
regions where normal network cannot be set. Thus the wide range of applications
attracts attacker. To secure from different types of attacks mainly intruder, intrusion
detection system based on dynamic state context and hierarchical trust in WSNs
(IDSHT) is proposed. The trust evaluation is carried out in hierarchical way. The
trust of sensor nodes is evaluated by cluster head (CH) whereas trust of cluster head
is evaluated by neighbor cluster head or base stations. Hence the content trust, honest
trust and interactive trust is put forward by combining direct evaluation and feed-
back based evaluation in the fixed hop range. In this way the complexity of trust
management is carried in hierarchical manner and trust evaluation overhead is mini-
mized. This proposed work addresses the security issues of wireless sensor network.
A more prominent intrusion detection system based on context level and trust level
is introduced. This mechanism achieves more than 90% accuracy in detection of
routing attack and sinkhole attack. The architecture suggested in this paper is used to
develop two level of trust model. Accuracy of 90% and 95% is expected in intrusion
detection and context text detection respectively.

R. B. Kagade (B)
Department of Computer Science and Engineering, SVERI’s College of Engineering, Pandharpur,
Maharashtra, India
PAH Solapur University, Solapur, Maharashtra, India
R. B. Kagade · J. Santhosh
Department of Computer Science and Engineering, Veltech Rangarajan Dr Sagunthala R & D
Institute of Science and Technology (Deemed to Be University), Avadi, Chennai, Tamilnadu, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 103
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_11
104 R. B. Kagade and J. Santhosh

Keywords Hierarchical trust management · Base station · Cluster head · Intrusion

detection mechanism · P2P network · WSN

1 Introduction

Wireless sensor network (WSN) is collection of sensor nodes that are equipped
with environmental sensors for heat, moisture, humidity, pressure, air/water quality,
weight, sound etc. WSN does not have fixed topology. WSN has wide range of
applications in field of agriculture, health, home, industrial, military and natural
calamities for monitoring and data collection purpose. The advantage of WSN is
that it is very easy to install in critical regions where normal network cannot be set.
Sensor nodes contain five components trans-receiver, processor, battery, hardware
and memory. The wide range of applications can be elaborated as follows:
Wild Habitat Monitoring
Sensors can be used to monitor the conditions of wild animals or plants in wild
habitats. Sensors can be used to collect information in the water and above water.
Disaster Monitoring
Information collection in natural and non-natural disaster areas is very important task.
A normal network cannot be set in such areas. WSN becomes prominent solutions
in such situations.
Warfield Monitoring
Sensors can be deployed in a war field to monitor the presence of forces and vehicles,
and track their movements, enabling close surveillance of opposing forces.
Sensitive Place Protection
Sensor nodes can be deployed around sensitive objects, for example, atomic
plants, strategic bridges, oil and gas pipelines, communication centers, and military
headquarters, for protection purpose.
Vehicle Monitoring
Sensors can be mounted on unmanned robotic vehicles, tanks, fighter planes,
submarines, missiles, or torpedoes to guide them around obstacles to their targets
and lead them to coordinate with one another to accomplish more effective attacks
or defenses.
Remote Sensing
Sensors can be deployed for remote sensing of nuclear, biological, and chemical
weapons, detection of potential terrorist attacks, and reconnaissance [1]. Due to its
vital range of applications it has several ranges of attackers at different levels. In
WSN sensor nodes have limited battery power, limited power to communicate and
State Context and Hierarchical Trust Management … 105

compute. WSN is susceptible to so many attacks, because of broadcasting nature of

network.
An intrusion is defined as a sequence of related actions performed by a malicious
adversary that results in the compromise of a target system. An Intrusion Detection
System specially looks for something wrong operation and events that might be the
outcome of a cause of attack, worm or system expert for pleasure.
Intrusion detection system consists of four methods which detects the system is
described below [2].

1.1 Network Based Intrusion Detection System

NIDS will monitor the traffic, malicious changes that are happening in the network
which causes changes that leads to system degradation.

1.2 Host Based Intrusion Detection System

HIDS may also be able to make out malicious business trade that starts with malware
and is attempting to put out on top for other systems.

1.3 Signature Based Detection System

Signature-based detection systems guide all the small data traversing the network
and make a comparison against a knowledge-base of signatures or given properties
of experienced violent behavior, like antivirus software.

1.4 Anomaly Based Detection System

Anomaly-based detection systems guide network business trade and make a compar-
ison against a started baseline, to come to a decision about what is taken into account
normal for the network with respect to bandwidth, protocols, harbors and other
devices [3, 4].
Trust is the degree of belief that a node can have on another node in the network
based on trust metric or trust rating o the node. Different types of attacks due to
intrusion are illustrated as follows [5].
GTMS: group based trust management scheme GTMS works in two types of
topologies intergroup topology and intra-group topology. In intergroup topology
106 R. B. Kagade and J. Santhosh

centralized trust management is adopted and in intergroup topology distributed trust

management is adopted.
ATRM: agent based trust and reputation management scheme An agent based
trust and reputation management scheme [6] (ATRM) is based on a clustered WSN
with mobile agent system. It requires every node to hold the mobile agent which is
administrating the trust and reputation of the hosting node.
PLUS Parameterized and localized trust management scheme for sensor network
security [7]. It uses distributed approach to adapt to different operational environ-
ments and different applications. The trust is calculated based on either direct or
indirect observations.
RFSN reputation based framework for sensor networks they have proposed a
framework where each node maintains reputation metrics which includes the past
behavior of other sensor nodes of the network and the metrics used for predicting the
future behavior. The values of the trust are evaluated on the basis of that reputation
and for representing the values of reputation, Bayesian formulation is used.
TRGR Trust management scheme for resilient geographic routing Trust manage-
ment scheme for resilient geographic routing [8] (TRGR) is a simple trust manage-
ment scheme which uses resilient geographic routing. Geographic routing consists
of two parts: geographic forwarding and complementary routing [9]. The trust algo-
rithm works in a fully distributed manner, in which each node monitors the behavior
of one hop neighbors. The basic idea of this trust management scheme is to favor
well behaving honest nodes by giving them the credit for each successful packet
forwarding, while penalizing suspicious nodes that doesn’t route packet according
to route.
BRMSN Behavior reputation method for sensor networks Behavior reputation
method for sensor networks [10] (BRMSN) measures the spatial information between
the nodes as reputation measurement. The model is a reflection of the node’s compre-
hensive ability about the actual physical properties and conduct essentially. Nodes in
the network not only include the identity of the trust but also the trust in the behavior
of the node. The model focuses on the local testing.
In this paper Intrusion detection system is introduced which depends on trans-
mission time and relieving time. Two level trust mechanism is proposed to reduce
overhead of base station and cluster head. The number of successful and unsuccessful
transmission of data at WSN decides the trust value of sensor nodes.
This paper is organized as follows: In Sect. 2 the existing systems are discussed
while Sect. 3 introduces the proposed work. Section 4 decides expected results and
Sect. 5 concludes the proposed.

2 Related Work

Hossein Jadidoleslamy proposed hierarchical intrusion detection mechanism [11].

The mechanism is designed and applicable. It works in one or two levels; it is consis-
tent to application domain and required security level. The research is focused on
State Context and Hierarchical Trust Management … 107

clustering of WSN and deploying cluster based intrusion detection system. The mech-
anism works in static and heterogeneous network, hierarchical and clustering struc-
ture. Hierarchical cluster based and leach based routing protocol is used for efficient
communication. The mechanism is proved to be as robustness and fault tolerant
design. The research provides reliable service but is an expensive mechanism. The
intrusion detection architecture supports real-time detection property almost 80.6%.
The mechanism results 80.6% and 55.7% accuracy for content based detection and
context based detection capabilities respectively.
Daojing He proposed a distributed trust evaluation model for medical sensor
networks [5]. The traditional cryptographic methods are not sufficient for trust
evaluation in medical sensor networks. The research work uses transmission rate
and leaving time into trust evaluation to detect malicious nodes. They proposed an
application-independent and distributed trust evaluation model for MSNs. The trust
management is carried out through the use of simple cryptographic techniques [5].
A novel distributed trust evaluation model for MSNs, where each node manages
trust records of other nodes about performing some activities. Centralized malicious
node detection and secure unicast routing is presented. It proved to be as improved
packet delivery with effective malicious node identification mechanism. IDS mech-
anism depends on transmission time, relieving time and packet dropping ratio. 88%
accuracy is shown by researchers with limited number of malicious nodes in WSN.
Fenye Bao proposed hierarchical trust management for WSNs and applied it
to routing and intrusion detection to detect selfish or malicious nodes [12]. The
work focused on multidimensional trust attributes and the trust value was calculated
through social trust and QoS trust, including intimacy, honesty, energy, and unselfish-
ness; meanwhile, subjective trust and objective trust were taken into consideration
to validate the proposed protocol [1]. However, the node with the maximum number
of interactions with neighbors was considered as the most trustworthy in the process
of the calculation of the intimacy trust inspired by social networks. The difference is
the consideration of the reasonable range of the maximum number of interactions,
as interaction that exceeds the range indicates malicious behavior. The mechanism
is proved to have IDS detection capabilities with 90% accuracy with false positive
probability is zero and method is scalable. Trust based geographic routing is used in
WSN.
Xiaoyong Li put forward a lightweight and dependable trust system for clustered
WSNs, which aims at decreasing resource consumption and enhancing the relia-
bility of CHs’ trust evaluation [13]. A self-adaptive weighting mechanism is used
to calculate trust value of CH which is better than subjective weight method. A
series of theoretical proofs were given in the research to verify the effectiveness of
the mechanism. In the process of trust evaluation, only successful and unsuccessful
interactions were taken into consideration, with no other trust evaluation factors taken
into account. The mechanism takes interactive trust, honesty trust and content trust
into account, addressing problems of consuming energy maliciously and tampering
multidimensional observing data with lower resource overhead, which is described in
108 R. B. Kagade and J. Santhosh

the performance evaluation [1]. The research is focused on minimizing memory over-
head and transmission overhead. It works better to protect from garnished attack and
bad mouthing attack. Results show that intrusion detection capabilities are 87.5%.
Shaikh proposed GTMS; a group-based trust management scheme for clustered
WSNs. GTMS evaluates the trust of a group of nodes in contrast to traditional
trust schemes that always focus on the trust values of individual nodes [14]. In this
approach WSN requires small memory to store trust value at each node. The mech-
anism achieves significant reduction of the cost associated with the trust evaluation
of distant nodes. But it depends on a broadcast-based strategy to collect feedback
from the CMs of a cluster, which requires a significant amount of resources and
power. This mechanism worked for wired and wireless mechanism. It focused on
reducing cost of trust evaluation. Trust value is calculated depending on time based
past interaction. Timing window is used to measure the number of successful and
unsuccessful interactions. Trust evaluation cost is minimized by 14.6–15.7%.
Ismail Butun has presented intrusion detection system for mobile Ad Hoc
networks [15]. Agent based distributed and collaborative IDSs are emphasized in
research. Two types of classifiers are used for detection of intrusion; Decision Tree
and support vector machine. Dynamic Source Routing, Ad hoc On-demand Distance
Vector and Destination Sequenced Distance Vector protocol is used for routing data
in WSN.
Limitations of Classical Approaches
Signature-based detection approaches are relatively easy to implement, require no
learning curve. This eliminates the risk of over-training or voluntary deformation
of the profile that can be observed in behavior-based approaches. However, these
approaches require an active maintenance and very frequent updates of the signa-
ture database to integrate any new attack discovered. Indeed, the update cannot be
performed automatically as in the case of behavior-based detection. This fact implies
a higher rate of false negatives.
The problem arises especially with very recent attacks for which signatures have
not been included in the database yet. Also, the absence of a standard pattern descrip-
tion language limits the usefulness of signatures described in a given language since
interoperability between different detectors is probably not possible. If signature
values are too simplified it can lead to detection of false intrude which corresponds
to legitimate actions and therefore to trigger false positives.
Anomaly-based detection approaches have several interesting features. First, as
the hypotheses are made only on the normal behavior of the system and not on
possible attacks, detection is exhaustive. Indeed, “the system allows a prior to detect
all that differs” from established normal behavior. Thus, it becomes possible to
envisage detection of unknown attacks and no specific knowledge about the attacker
is required. All necessary information is collected within the system. On the other
hand, once the learning phase terminates, the IDS does not require particular update.
The definition of normal behavior evolve only slightly if any.
Nevertheless, a high rate of false positives is the main weakness of these
approaches because it is sometimes difficult to define the “normal behavior”. Sudden
State Context and Hierarchical Trust Management … 109

changes in the environment can have an impact on behavior. This sudden change in
behavior will be considered as an anomaly and an alert will be generated. Also,
since the first phase is dedicated solely to the development of the definition of
“normal behavior”, this one is particularly vulnerable to attack. Indeed, the pres-
ence of signals related to an attack in the learning trace will result in skewing the
definition of behavior. Thereafter, any similar attack will be treated as a normal
behavior. The information used during this first phase in the optimal condition must
be totally free from damage. In practice, it is frequently impossible to have such
perfect environment.

3 Proposed Methodologies

Problem Statement: To Develop State Context and Hierarchical Trust Mecha-

nism in Wireless Sensor Networks for Intrusion Detection
Following assumptions are considered [1]:
(i) The WSN is cluster-based, and SNs in a cluster could communicate with
the CH directly, whereas CHs communicate with BSs directly or indirectly
through other CHs.
(ii) Each SN has a unique ID and belongs to a unique cluster, and CHs have more
energy than SNs.
(iii) The data transmission model in a WSN is hybrid, including continuous and
event-driven.
(iv) The states of SNs include hibernation, monitoring and active, and the transition
between monitoring and active is taken into consideration during the trust
evaluation of SNs.
(v) Sensor nodes are deployed densely and redundantly for reliability.

Cluster Creations
In wireless sensor network large numbers of application specific sensor nodes are
connected with each other on ad hoc purpose. If each node starts communication,
computation and routing then energy consumption, collision and congestion can
occur. Hence it may become reason for performance degradation. Node clustering
can solve these issues. In clustered network wireless sensor network is divided into
small units. Cluster head is elected. Sensor nodes in each cluster will communicate
with respective cluster head CH and CH aggregates data and will transfer data to
a central base station. The cluster head is considered to have maximum energy. To
maintain the maximum energy at cluster head re-election after certain interval of
time is best possible solution.
LEACH [16] is the first clustering scheme. It is used for periodical data gathering
in WSNs. It assumes the sensor nodes will remain static and communicate with
each other by single-hop only, and they can transmit data to the sink node directly.
110 R. B. Kagade and J. Santhosh

Its operation is divided into rounds and each round is composed of two phases
[17]. In cluster formation phase, LEACH elects some cluster heads according to the
probability shown below.

k
N −k(r mod N /k)
, Ci (t) = 1
Pi(t) =
0, Ci (t) = 0

where k is the desired number of cluster heads, Ci (t) is the indicator function deter-
mine whether or not node i has been a cluster head in most recent (r mod N/k) rounds.
Pi (t) is the probability for node i to become a cluster head. The rest of sensor nodes
join the proper cluster according to the signal strength from the cluster heads. In
the data transmission phase, the cluster heads aggregate the data from their cluster
members. Since cluster head is chosen by probability in each round, the load is
balanced to certain extent.
Trust Calculations
Interactive trust refers to the trust value computed by the number of interactions
between nodes, and an interaction means a node sending/receiving a packet or a
request to/from another node. The trust value is mapped to the integer number in
the range of [0, 10], where 0 demonstrates the most distrustful, while 10 implies the
most trusted, and 5 is the medium trust [1] shown in Fig. 1.
Sensor nodes trust is evaluated by the CH in a cluster, i.e., CH-to-SN trust,
which considers multidimensional trust, including interactive trust, honesty trust
and content trust, during the procedure of trust calculation.
Interactive trust SITij (t) is calculated by the number of interactions between
node j and its CH i in t. In the proposed method, interaction refers to all commu-
nication behavior including sending and receiving of request and data packets. The
greater the number of interactions of two nodes, the higher is the trust value [18].
However, in WSNs, if the number of interactions exceeds a threshold, the trust value
will decrease because there may exist malicious interactions such as attacks that send
a large amount of packets or requests to exhaust the energy of the node. Therefore,
unlike trust evaluation in social networks, the interactive trust evaluation method in

Fig. 1 Wireless sensor network

State Context and Hierarchical Trust Management … 111

WSNs is put forward. Inspired by Normal Distribution in Statistics, the probability

density function, which is normalized to [0, 1] to calculate the interactive trust, is
adopted when the number of interactions exceeds a threshold.
Interactions between CH and SNs are abstracted as an undirected weighted
graph, the weight of which represents the number of interactions between them.
The interactive trust value of SN j evaluated by CH, SITij (t) can be defined as
 
Wi j
S I T i j(t) = 10 × , j ∈ G, W i j ≤ λμ;
Wi j
  
|W i j − μ|
10 × ex p ex p − j ∈ G, W i ≥ λμ;
θ

where x denotes the largest integer that is equal to or less than x, μ is the mean value
of the number of interactions between CH and SNs in the same state, λμ is taken as
the threshold of the interaction range, in which λ is a parameter used to define the
upper limit of normal interactions, and θ is a significant factor, which values 1, 10
and 100 when wij is a single digit, tens digit or hundreds digit, correspondingly, and
so on.
Example:
The WSN is assumed to contain eight nodes and one cluster head. Cluster head is
selected depending on computation power, energy, honesty and distance between
CH and base station (Fig. 2).

Fig. 2 Trust calculation in

WSN
112 R. B. Kagade and J. Santhosh

Calculation of Trust value

(i) λ is set to be 2.
(ii) The value of θ is 1,10,100 depending on value of wij . If the value of wij is two
digits θ is 10. If wij is three digit number then θ is 100.
(iii) μ is mean of weights.
μ = ((3 + 5+4 + 13 + 5+6 + 4+16)/8) = 7
(iv) Maximum weight max (wij ) = 13 because weight 16 is greater than λμ(14).
The maximum weight is discarded and 13 is considered as maximum weight.
(v) μ = ((3 + 5+4 + 13 + 5+6 + 4+16)/8) = 7
Trust value is calculated as follows SITij (t) =

(a) For node A = 10 ∗ (3/13) =10 ∗ 0.2307

=2.307
=2

(b) For node D = 10 ∗ (13/13) =10 ∗ 1

=10 ∗ 1
=10

(c) For node H. The weight is 16; it is greater than threshold λμ which is 14. As
the number of is greater than 14 it is considered as distrustful. θ is 10 as the
wij is two digit number. The trust value is calculated as follows (Table 1)

SITij (t) =10 × exp(−|16 − 7|/10)

=10 × exp(−|0.9|)
=10 × 0.4065696
=4.06
=4

Table 1 Evaluation results of

Nodes Evaluation results of trust values
examples in Fig. 1
A 2
B 3
C 3
D 10
E 3
F 4
G 3
H 4
State Context and Hierarchical Trust Management … 113

Honesty Trust of SNs

Honesty trust SHTij (t) is calculated by the number of successful and unsuccessful
interactions between CH i and a non hibernating SN j in t. The CH i overhear the
SN j if j does not deliver a packet in t or transmits the packet to another node that is
not in its routing table, or if the packets from j do not reach the CH i, the interaction
between them is considered an unsuccessful interaction.
The number of successful and unsuccessful interactions between active nodes and
CH i in t is denoted as s and f, and the trust value is evaluated using formula given
below.
 
10 × (s + 1)
S I T i j(t) = , when f = 0
(s + f + 2)

10 × (s + 1)
−1 , when f = 0
(s + f + 2) × f 2

When there are no interactions between active members, i.e., s = f = 0, the trust
value is 5. If there are unsuccessful interactions, the honesty trust value will decrease
sharply because of the punishment executed by [13]. For non active members, they
inherit the trust value of their lasno hibernating state.
Content Trust of SNs
Content trust is the trust evaluation based on observing data, which is data-oriented
trust calculated by CH. Content trust is introduced because the WSN is a data-
centric network and the observing data are the factor of most concern for applications.
Tampering attacks often occur in WSNs to interfere with the network and applications
and can be identified by content trust.

C T i j(t) = 10 × ex p(−Di j)

1
dm 2

Di j = (Xik − X jk)2̇
k=1

The overall trust of SN j evaluated by CH i is calculated by formula given below,

which aggregates the interactive trust, honesty trust and content trust.

S O T i j = αS I T i j + β S H T i j + (1 − α − β)SC T i j

Parameters α, β ∈ [0, 1] are weights for each sub trust value. The higher the
weight, the more important that sub trusts is to overall trust and vice versa.
114 R. B. Kagade and J. Santhosh

Intrusion Detection at SN level

Malicious SN detection is executed by the respective CH. The CH c evaluates and
maintains the trust value of SN j in the same cluster and selects a trust threshold TSthi
according to the trust value of SNs in cluster i, which is calculated as (Table 2) :
 
T S thi = avg j∈C Land S O Tcj ≥5{ S O Tcj } ; ∃ j, s.t S O Tcj ≥ 55; other

Intrusion Detection at CH level

 
T C th = avg j∈C H SandC O Tbj ≥5{C O Tbj } ; ∃ j, s.t C O Tbj ≥ 55; other

4 Expected Results and Outcomes

The proposed methodology will be tested on network simulator (ns2.35). Large

numbers of sensor nodes are considered as components of WSN. The WSN is
assumed as clustered WSN.
There are two major types of attacks in wireless sensor network; active attack and
passive attack [19].
Major intrusion attacks can be listed as follows:
(i) Routing attacks
(ii) Selective forwarding
(iii) Sinkhole attack
(iv) Sybill attack
(v) Wormholes attacks.
The proposed mechanism will work efficiently on routing attacks and sinkhole
attack; as the method is depending on transmission time and relieving time.
Detection Rate (DR) of an Intruder attack

Detection rate
It is defined as ratio of total number of attacks detected and total number of
attacks appeared [19].

P K T detect
W S N acc =
P K T detect + F P K T detect

W S N acc = Accuracy of proposed method

P K T detect = Accurately detected packets as intruder.
F P K T detect = False detected packets as intruder.
State Context and Hierarchical Trust Management … 115

False Positive Rate (Frate )

FP
F Pdetect =
FP + T N

FPdetect is the rate at which false packet detection.

FP is number of false packet detected.
TN is true negative packet detected, the no. of legitimate records.

5 Conclusion

This paper addresses the security issues of wireless sensor network. A more promi-
nent intrusion detection system based on context level and trust level is introduced.
This mechanism achieves more than 90% accuracy in detection of routing attack and
sinkhole attack. The architecture suggested in this paper is used to develop two level
of trust model. Where first trust level is sensor node to sensor node communication
and second trust level is between sensor node and cluster head. The overhead of
cluster head and base station is minimized in this way. Architecture will be tested
on simulator ns2. Accuracy of 90% and 95% is expected in intrusion detection and
context text detection respectively (Table 2).

Table 2 Database at WSN

Items Implications
ID The ID of a SN
wij The number of interactions between CH i and SN j
μ The mean value of interactions between a CH and SNs that are at the same state in a
cluster
θ The significant factor based on the value of wij
s The number of successful interactions between i and j
f The number of unsuccessful interactions between i & j
Dij The Euclidean distance of data between average and j
SITij The interactive trust of SN j evaluated by CH i
SHTij The honesty trust of SN j evaluated by CH i
SCTij The content trust of SN j evaluated by CH i
SOTij The overall trust of SN j evaluated by CH i
TSthi The threshold of malicious SN detection
116 R. B. Kagade and J. Santhosh

References

1. Zhang Z, Zhu H, Luo S, Xin Y, Liu X (2017) Intrusion detection based on state context and
hierarchical trust in wireless sensor networks. IEEE Trans Content Min
2. Benisha RB, Raja Ratna Dr S (2018) Prevention of cyber attacks in control systems: a review.
In: Proceedings of the 2nd international conference on trends in electronics and informatics
(ICOEI 2018)
3. Ntalampiras S (2014) Detection of integrity attacks in cyber-physical critical infrastructures
using ensemble modelling. In: IEEE transactions on industrial informatics, vol 11
4. Sadi Y, Ergen SC, Park P (2014) Minimum energy data transmission for wireless networked
control systems. IEEE Trans Wirel Commun 13(4):2163–2175
5. He D, Chen C (2012) A distributed trust evaluation model and its application scenarios for
medical sensor networks. In: IEEE transactions on information technology in biomedicine, vol
16, no 6
6. Boukerche A, Li X, El-Khatib K (2007) Trust-based security for wireless ad hoc and sensor
networks. Comput Commun 30:2413–2427
7. Maddar H, Kammoun W, Youssef H (2017) Trust intrusion detection system based on location
for wireless sensor network. Springer International Publishing AG
8. Liu K, Abu-Ghazaleh N, Kang K-D (2007) Location verification and trust management for
resilient geographic routing. Parallel Distrib Comput 67(2):215–228
9. He T, Stankovic JA, Lu C, Abdelzaher T (2003) SPEED: a stateless protocol for real-time
communication in sensor networks. In: Proceedings of the 23rd international conference
ondistributed computing systems (ICDCS’03), Washington, DC, USA, pp 46–55
10. Zhou M-Z, Zhang Y, Wang J, Zhao S-Y (2009) A reputation model based on behavior trust
in wireless sensor networks. In: Eighth IEEE international conference on scalable computing
and communications
11. Jadidoleslamy H (2011) A hierarchical intrusion detection architecturer for wireless sensor
networks. Int J Netw Secur Its Appl (IJNSA) 3(5)
12. Bao F, Chen I, Chang M, Cho J (2012) Hierarchical trust management for wireless sensor
networks and its applications to trust-based routing and intrusion detection. IEEE Trans Netw
Serv Manage 9(2):169–183
13. Li X, Zhou F, Du J (2013) LDTS: a lightweight and dependable trust system for clustered
wireless sensor networks. IEEE Trans Inf Forensics Secur 8(6):924–935
14. Shaikh RA, Jameel H, d’Auriol BJ, Lee H, Lee S, Song Y (2009) Group-based trust management
scheme for clustered wireless sensor networks. IEEE Trans Parallel Distrib Syst 20(11):1698–
1712
15. Butun I, Morgera SD, Sankar R (2014) A survey of intrusion detection systems in wireless
sensor networks. IEEE Commun Surv Tutor 16(1):234–241
16. Heinzelman WR, Chandrakasan A, Balakrishnan H (2002) An application-specific protocol
architecture for wireless microsensor netw orks. IEEE Trans Wirel Commun 1(4):660–670
17. Ye M, Li C, Chen G, Wu J (2005) An energy efficient clustering scheme in wireless
sensor networks. In: PCCC 2005, 24th IEEE international performance, computing, and
communications conference
18. Hao F, Min G, Lin M, Luo C, Yang L (2014) MobiFuzzyTrust: an efficient fuzzy trust inference
mechanism in mobile social networks. IEEE Trans Parallel Distrib Syst 25(11):2944–2955
19. Kalnoor G, Agarkhed J (2018) Intrusion threats and security solutions in wireless sensor
networks. Int Robot Autom J
Portable Camera Based Assistive Text
and Product Label Reading from Hand
Held Object by Using Android App
for Blind Person

Somnath Thigale and Ranjeet B. Kagade

Abstract We propose a camera-based mechanical man app. This app helps the blind
persons to browse the text on explicit objects. In this system the camera captures the
actual text on the object. Multiple techniques square measure applied to its text.
Such as Optical Character Recognition that supply the operation of scanning and
recognition of text and a few have integrated voice output. From a grayscale image,
thresholding are often accustomed produce binary pictures i.e. image with solely
black or white colors, Filtering are often accustomed cut back the noise of image, Next
image segmentation technique is employed to perform the method of partitioning a
digital image into multiple segments. The goal of segmentation is to modify and/or
amendment the illustration of a picture into one thing that’s a lot of significant and
easier to analyses. Image scaling is the method of resizing a digital image. Next
technique employed in this project is template matching. Temples matching is a way
in the digital image process for locating tiny components of a picture that match
a template image. Also template extraction are often employed in producing as a
vicinity of internal control, some way to navigate a mobile golem or as some way to
notice edges in images then finally voice output are going to be generated then blind
man will simply listen to the text on it explicit object.

Keywords Assistive devices · Blindness · Distribution of edge pixels · Hand-held

objects · Optical character recognition (OCR) · Stroke orientation · Text reading ·
And text region localization

1 Introduction

The 314 million visually impaired folks worldwide, forty five million square measure
blind handicap that was free by “World Health Organization” in ten facts concerning
sightlessness. Even during a developed country just like the US, the 2008 National
Health Interview Survey (NHIS) reportable that AN calculable 25.2 million adult

S. Thigale (B) · R. B. Kagade

SVERI’s College of Engineering, Pandharpur, Maharashra, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 117
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_12
118 S. Thigale and R. B. Kagade

Americans (over 8%) square measure blind [1, 2]. The human baby generation ages
square measure quickly increasing. all over written text square measure obtainable
as an example receipts, bank statements, building menus, room handouts, report,
product packages, directions on medication bottles, room handouts etc. While screen
readers, optical aids and video magnifiers will facilitate blind users and people
with low vision to access documents. The few devices offer sensible access to
common hand-held objects like product packages and objects written with text as an
example prescription medication bottles. The visually handicapped person to scan
written labels and merchandise packages can enhance freelance living and foster
economic and social self-reliance. Today, there square measure already has some
systems for transportable use however they can’t handle product labeling. The trans-
portable Universal Product Code readers designed to assist blind people identify
totally different products in an in depth product info will change the user for visu-
ally handicapped person to access data regarding these products through speech and
Braille. during this system an enormous limitation is that it’s terribly exhausting for
blind users to seek out the position of the Universal Product Code and to properly
purpose the Universal Product Code reader at the Universal Product Code however
Some reading-assistive systems like pen scanners could be used in these and similar
things. This technique integrates OCR computer code used for scanning and recog-
nition of text and a few have integrated voice output. The OCR is optical character
recognition. These systems square measure typically designed and perform best with
document pictures with appropriate backgrounds, commonplace fonts, a tiny low
vary of font sizes, and well-organized characters. This technique is used instead of
business product boxes with multiple decorative patterns. The OCR computer code
cannot directly handle scene pictures with complicated backgrounds and therefore
the document to be scan should be nearly flat, placed on a transparent, dark surface
and contain principally text [3, 4].
Even though variety of reading assistants square measure designed specifically for
the blind folks, to our knowledge, there’s no existing scanning assistant will read text
from the varieties of difficult patterns and Even though variety of reading assistants
square measure designed specifically for the blind folks, to our knowledge, there’s
no existing scanning assistant will read text from the varieties of difficult patterns
and backgrounds that found on many everyday business product. Like text data will
obtainable in varied scales, fonts, colors, and orientations. Images. Initial is Milk
box (Top) and Men lavatory signage (Bottom).
Above Figs. 1 and 2 shows (a) camera captured images. (b) Localized text regions
that show in fig mark blue. (c) Text region scropped from image. (d) Text codes
recognized by OCR (optical character recognition). Text at the top-right corner of
bottom image is shown during a increased callout. Mobile accurately reads black print
text on a white background however there some issues to recognizing colored text
or text on a colored background. It cannot scan text with complicated backgrounds
as a result of they can’t simply detected the text from background. Text printed
on cylinders with crooked or incomplete pictures as an example as soup cans or
medication bottles. These systems require a blind user to manually localize areas
of interest and text regions on the objects. As shown in Fig. 1, such text data will
Portable Camera Based Assistive Text … 119

Fig. 1 Examples of printed text from hand-held objects with multiple colors backgrounds

appear in multiple scales, fonts, colors. to help blind persons to scan text from hand-
held objects, we have planned of a camera-based helpful text reading frame work
to track the item of interest at intervals the camera read and extract print text data
from the item. the prevailing rule will effectively handle complicated background and
multiple patterns and extract text data from each hand-held objects and assemblage as
shown in Fig. 2. The hand-held object extract camera image and we tend to develop
a motion-based technique to obtain region of interest (ROI) of the item. During
this ROI perform the text recognition technique. The localization of text regions
in scene pictures they’re divide in 2 classes: rule-based and learning–based. Rule
primarily based rule to used component level image process to extract the text data
120 S. Thigale and R. B. Kagade

Fig. 2 Two examples of text localization and recognition from camera captured Images. First is
Milk box (Top) and Men bathroom signage (Bottom)

from predefined text layout like character size, ratio, edge density, character structure,
color uniformity of text and learning primarily based rule square measure used model
text structure and extract representative text options to make text Classifiers [5, 6].

2 Existing System

In these days society, there square measure already a couple of systems that square
measure moveable use for blind persons. For instance, moveable code readers. The
blind those who want to access data regarding these product, the moveable code reader
helps the blind individuals to spot totally different product in an intensive product
information. However properly position the code reader at the code is incredibly
arduous task for blind users.
Con jointly there some system that uses camera-based helpful text reading frame-
work. Foremost object of interest inside the camera read is track so written text data
from the item is extracted by this framework. During this framework, the item of
interest is positioned at the middle of camera’s read. This object ought to be posi-
tioned specified, it ought to be seem within the camera read. These wide angles
of this camera accommodate the users with solely particular/approximate aim. The
system extract hand-held object from the camera image. To get the region of interest
(ROI) of the item this framework uses the motion primarily based technique. Then
text recognition is performed just for that region of interest (ROI) [7, 8].
Portable Camera Based Assistive Text … 121

Fig. 3 Region of interest

Sometimes multiple scales, fonts and colors square measure won’t to write the
text characters. Conjointly immense quantity noise is contained within the captured
pictures that contain text instead of industrial product boxes with multiple ornamental
patterns the pictures with straightforward background, commonplace fonts, tiny vary
of font sizes and well organized character square measure used for these systems,
systems provide higher performance however scene pictures with advanced back-
grounds aren’t directly handled some optical character recognition(OCR) software
system shown in Fig. 3.

3 Proposed System

In existing system camera capture the text on image and so process is finished in
portable computer and voice output is created however it’s not possible for visually
handicapped person to hold the portable computer every time. To overcome the
matter of existing system the new system is introduced that’s transportable camera
primarily based golem app to helps the visually handicapped person. We projected
golem app this app facilitate the visually handicapped person to browse the text on
explicit object. In this system camera capture the actual text on image and multiple
techniques square measure applied on it text and finally voice outputs are generated
and so visually handicapped person will simply listen the text on it explicit object.
122 S. Thigale and R. B. Kagade

4 Framework and Algorithm

Gray scale
In Fig. 4 photography and computing, a grayscale or gray scale digital image is a
picture during which the worth of every component could be a single sample, that is,
it carries solely intensity data. Pictures of this type, conjointly referred to as black-
and-white, area unit composed solely of reminder grey, varied from black at the
weakest intensity to white at the strongest.
How does one convert a color image to grayscale? If every color component is
represented by a triple (R, G, B) of intensities for red, green, and blue, however
does one map that to one range giving a grayscale value? There area unit following 3
algorithms. The lightness methodology averages the foremost distinguished and least

Fig. 4 Framework and algorithm

Portable Camera Based Assistive Text … 123

distinguished colors: (max(R, G, B) + min(R, G, B))/2. The average methodology

merely averages the values: (R + G + B)/three.
The luminousness methodology could be a additional refined version of the
common methodology. It conjointly averages the values, however it forms a weighted
average to account for human perception. We’re additional sensitive to inexperienced
than different colors, therefore inexperienced is weighted most heavily.
The formula for luminousness is 0.21 R + 0.71 G + 0.07 B.
The example sunflower images in Fig. 5.
Threshold
Thresholding is that the simplest methodology of image segmentation.

Original image

Lightness

Average

Luminosity

Fig. 5 Flower
124 S. Thigale and R. B. Kagade

• From a grayscale image, thresholding is accustomed produce binary pictures i.e.

image with solely black or white colours.
• It’s typically used for feature extraction wherever needed options of image area
unit regenerate to white and everything else to black. (Or vice versa) refer above
Fig. 5.

5 Conclusion

In this paper we’ve represented a system that scans written text on hand-held objects
for aiding visually handicapped people. So as to resolve the common aiming draw-
back for blind users. We tend to plan mobile apps. The mobile camera captured
specific images and applied the text recognition on the image to extract and localize
the written text. We planned an OCR software package. This OCR software package
performs varied techniques for manufacturing the audio output for the visually hand-
icapped person. Then visually handicapped people simply scan the text from the
object.

References

1. Rajkumar N, Anand MG, Barathiraja, Portable camera-based product label reading for blind
people. Department of electrical and electronics engineering, The Kavery Engineering College,
Mecheri, Tamil Nadu, India
2. Yi C, Tian Y, Arditi A Portable camera-based assistive text and product label reading from
hand-held objects for blind persons
3. Text detection in natural scene images by Stroke Gabor words, Yi C Dept of Computer Science,
The graduate center, City Univ of New York, New York, U.S.A., e-mail: [email protected].
Tian Y Dept of Electrical Engineering, The City College and Graduate Center, City Univ of
New York, New York, U.S.A., e-mail: [email protected]
4. Chen X, Yang J, Zhang J, Waibel A Automatic detection and recognition of signs from natural
scenes
5. Kumar S, Gupta R, Khanna N, Chaudhury S, Joshi SD Text extraction and document image
segmentation using matched wavelets and MRF
6. Yang X, Tian Y, Yi C, Arditi A Context-based indoor object detection as an aid to blind
persons accessing unfamiliar environments. Dept of Electrical Engineering, The graduate Center
Arlene R Gordon Research Institute, the city college of New York. The City University of New
York Lighthouse International NY 10031, USA NY 10016, USA New York, NY 10222, USA
{xyang02,ytian}@ccny.cuny.edu [email protected] [email protected]
7. Experiments with a New Boosting Algorithm., Yoav Freund and Robert E. Schapire,
AT&T Laboratories 600 Mountain Avenue, Murray Hill, NJ 07974-0636, yoav,
[email protected]
8. Assistive Text Reading from Complex Background for Blind Persons, Yi C Dept. of
Computer Science, The Graduate Center, City Univ of New York New York, U.S.A. e-mail:
[email protected], Tian Y Dept of Electrical Engineering. The City College and Graduate
Center, City Univ of New York, New York, U.S.A
Automatic System for Identifying
Cholesterol

Mohua Biswas, Pragtee Tathe, Geeta Unhale, and Papiya Biswas Datta

Abstract At the outer edge a solid white ring circling iris exist which is known
as sodium ring or deposit normally located around iris of an eye which symbolizes
the existence of high level cholesterol in the human body. There is a half-circle of
gray, white and yellow deposits in the outer edge of cornea known as Arcus senilis
or arcus senilis cornea occurred because of fat and cholesterol deposits in people
under 45 years of age. The high cholesterol is also an indication of the presence of
hyperlipidemia which indicates the increased amount of fats in the blood. There is a
risk of developing heart disease which leads to stroke and death. There is a chance to
occur a common type of heart disease known as coronary artery disease (CAD) when
the arteries that supply blood to heart muscle become hardened and narrowed. This
is also due to the buildup of cholesterol. Iridology approach is the different type that
also helps to identify diseases with the help of pattern of an iris. Whenever there is a
excess deposition of the cholesterol occur in the body it will create a whitish sodium
ring around the iris, By analyzing the sodium ring we can identify the existence of
cholesterol in the human body.

Keywords Cholesterol · Coronary · Arcus senilis · Iridology

1 Introduction

In the human body there is a presence of a slippery or waxy substance known as

Cholesterol. It is required to develop cells in human body. But excessive of it may
affect in human body. Normally it occurs in two ways. First the human liver gener-
ates the required amount of cholesterol in human body and the remaining amount of
cholesterol is from food which we digest. Excessive amount of it leads to heart attack.
It is a combination of steroid and alcohol having the chemical formula C27H46O. An

M. Biswas (B) · P. Tathe · G. Unhale

SVERI’s College of Engineering, Pandharpur, Solapur, India
e-mail: [email protected]
P. B. Datta
Devi Mahalaxmi Polytechnic College, Titwala, Kalyan, Thane, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 125
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_13
126 M. Biswas et al.

approach based on laser helps to calculate the cholesterol in human blood through
skin. In order to classify the cholesterol in the skin the infrared (IR) absorption spec-
troscopic is proposed [1]. Generally, there is a presence of cholesterol of approx-
imately 11% in skin because of weight. It gets increased if there is a symptom of
coronary artery disease. In that case for analyzing coronary artery disease palm test
is not so effective as a screening tool. The blood test or lipid profile test is the modern
method to analyze the level of cholesterol. Before lipid profile test there is a need of
12 h fast. This test is tested by observing variations in the pattern of iris which helps
to identify the increase level of cholesterol known as Arcus senilis. Arcus senilis is
a half-circle of gray, white and yellow deposits in the outer edge of cornea. Arcus
senilis is caused due to the lipid deposit in the inner layer of the cornea peripheral.
Because of increase amount of cholesterol there is chance of occurring arcus juve-
nilis problem which indicates discoloration of an eye [2]. Thus for analyzing and
detecting the level of cholesterol in the body of humans’ iris pattern is the alternate
and effective approach known as iridology which was found by Bernard Jensen’s.
Many iridologists admitted that the occurrence of heart disease is due to cornea
having yellowish-white deposit. The high level of cholesterol in the blood is known
as Hypercholesterolemia. It is necessary to check the level of cholesterol regularly
which helps to get rid of the formation of gallstones with bile pigments that normally
occurs because of high level of cholesterol. Protein in human body helps cholesterol
to flow in human body and it is known as lipoprotein. Cholesterol is distinguished in
two ways depend upon the amount of protein present: (i) Low Density Lipoproteins
(LDL) and (ii) High Density Lipoproteins (HDL).

2 Human Iris

The iris of the eye is a thin having 12 mm in diameter and rounded in shape consists of
connective tissue and muscle that encircles the pupil. The Fig. 1 shows the schematic
diagram of a human eye. There is a lens behind the iris of the eye. At the back of the
eye on the inside there are linings covered by cells for light-sensing known as retina
[3]. Retina converts light into electrical impulses. These impulses are transferred to
the brain by optic nerve. The coloring portion of the human eye manages the light to

Fig. 1 Schematic
representation of human eye
Automatic System for Identifying Cholesterol 127

enter. The iris of the eye shut the pupil if the light is bright and because of that less
amount of light passes. If the light is low then iris opens the pupil that allows light to
pass. The size of the pupil is managed by the sphincter and the dilator muscles that
vary from 10% to 80% [4].

3 Iridology

The term iridology which is also referred as iridodiagnosis or iridiagnosis is an impor-

tant approach or method that helps to test iris of a human eye in order to find out
patient’s health. Many experts test their estimation to iris diagram which is distributed
into various regions that correlate to different human body parts. Iridologists consid-
ered human eyes as “windows” that reflect the present condition of human health.
The color of iris is greenish-yellow that encircles the pupil of black color. Sclera is
the white portion of the eye. Usually for examining iris iridologists use torch or hand
held lens for tissue change. The iris of an eye is the basic representation of human
body that speaks about specific transformation in human organs. The iris at the right
hand side as shown in Fig. 2 demonstrates the situation of the organs at the right side
of human body and the left hand side of the iris demonstrates the situation of the
organs at the left side of human body [5].
In order to examine the cholesterol in human body, Iridology is the substitute
for treatment where individual’s health and the condition of organs in the body are
reflected. As per the iridology chart, the detection of cholesterol in human body is
simply by observing the sodium whitish ring present in the patient’s eye. Also, due
to the presence of cholesterol in human body there is a variation in the iris pattern
and known as Arcus Lipoides (Arcus Senilis or Arcus Juvenilis) [6]. Some of them
is shown in Fig. 3.

Fig. 2 Relationship cholesterol presence with arcus senilis

128 M. Biswas et al.

Fig. 3 Example of iris image

4 Methodology

The following the approaches that are supported by the automated system in this
paper are shown in Fig. 4:

A. Image Recognition or Acquisition:

Generally, image recognition or acquisition is the initial phase of vision system.
It is a process by which the image is gathered with the help of photographic
equipment and that images are saved in the database. But the authentic sample
of an eye image is gathered from sufferer at the clinic of cardiology and ophthal-
mology. Database consists of both the cases normal as well as exceptional cases
gathered from clinic or from internet. The existence of cholesterol in the body
is detected by a sodium ring which is normally occurred in corneal part of the
iris [7].

Fig. 4 Methodology
Automatic System for Identifying Cholesterol 129

Fig. 5 Captured eye image

B. Pre-processing:
The second phase is image preprocessing. The Fig. 5 shows captured image for
second phase. As we normally gathered images from various sources there may
be existence of distortion. So the purpose of image pre-processing is to improve
the distorted images to get better quality image with the help of enhancement
techniques. In this phase histogram equalization method is used to enhance the
contrast of an image.
C. Image Localization or Segmentation:
The image localization or image segmentation is the most critical phase. It is
a method of detecting interior and exterior border regions of the iris. By using
subtraction approach circular shaped iris is detected and extracted from it.
D. Image Normalization:
As the phase of segmentation ends the image normalization phase starts where
it reconstruct the image of an iris. In this phase it unwrap the circular region
of the iris to rectangular block with the help of Daugman’s rubber sheet model
method. Evaluation can be done either from top to bottom or from bottom to
top.
E. Region of Interest:
In this phase of extraction the portion of the iris image is extracted which we
want to refine by simply cropping the normalized iris image. The region of
sodium ring is only extracted.
F. Feature Extraction:
After the region of interest is carried out the next phase is to find out the various
features such as Mean, Entropy, Standard Deviation, Smoothness, Kurtosis,
Variance and Homogeneity related to this region. The results are displayed in
two ways; one is the normal conditions and other is the exceptional cases of
cholesterol.
G. Classification: (Keep text style as normal)
Figure 6 show the classifier used for next stage of image processing based on the
result of feature extraction we classify the level as normal level and cholesterol
level which is beyond normal level [8].

5 Result

In Fig. 7, the above result shows the implementation of our methodology tested on
patient suffering from cholesterol. In this two images of an iris are captured; one is
colour image and other is black and white. In the case of localized iris image the
130 M. Biswas et al.

Fig. 6 SVM classification

Fig. 7 Stages for cholesterol detection for Abnormal Iris

white portion i.e. sodium ring is present in brown fiber like area. The sodium ring is
segmented and by analyzing the ROI we get the conclusion that there is a presence
of cholesterol in human body beyond the safe level.
In Fig. 8, there is absence of sodium ring in brown portion which means no
cholesterol in the body.
Automatic System for Identifying Cholesterol 131

Fig. 8 Stages for cholesterol detection for Normal Iris

6 Conclusion

In this paper automatic identification of cholesterol in human body through iris is

presented. This system requires less time to identify the existence of cholesterol in
human body. In biomedical field it is an important application.

References

1. Li L-H, Dutkiewicz EP, Huang Y-C, Zhou H-B, Hsu C-C (2019) Analytical methods for
cholesterol quantification. Sci Direct 27(2)
2. Arora A, Paggowal N, Chawla D, Sharma E (2018) A non-invasive framework for detecting
hypercholesterolemia using computer vision. IJEDR 6(4). ISSN: 2321–9939
3. Songire SG, Joshi M (2016) Automated detection of cholesterol presence using iris recognition
algorithm. Int J Comput Appl
4. Shanker NR, Ezhil A, Archana S (2012) Non-invasive method of detection of cholesterol using
image processing. Int J Med Eng Inform
5. Silambarasan T, Sethupriyan P (2015) Identification and cholesterol estimation of skin using
hand pattern image. Int J Adv Res Electr Electron Instrum Eng 4(8)
6. Mahesh Kumar SV, Gunasundari R (2013) Comparative Analysis and Earlier Detection of
Cholesterol Based on the Measurement of Various Lipid Level Deposits in the Eye. CIIT 5(7)
7. Adi KG, Rao PV (2016) A non-invasive method of detection of cholesterol by level set segmen-
tation and ANN as classifier using MLP-BP and RBF. Int J Adv Electr Comput Eng (IJAECE)
3(3)
8. Adi KG, Rao PV, Adi VK (2015) Analysis and detection of cholesterol by wavelets based and
ann classification. In: 2nd International conference on nanomaterials and technologies (CNT
2014)
Design of an IoT Based System
for Monitoring and Controlling
the Sub-Station Equipment

Pranali Bodke and A. A. Kalage

Abstract In the era of modern digitalization world, it is a simple to monitor and

control the substation equipment remotely using expensive PLC and SCADA system,
but it is desirable to design a system which is cost-effective, smart and reliable. So that
IoT is an effective solution as the real-time capability of IoT is considered as a key
feature for monitoring and control applications of power systems. The IoT is a system
of interrelated computing devices, mechanical and digital machines, objects, animals
or people that are provided with unique identifiers and the ability to transfer data over a
network without requiring human to human or human to computer interaction. These
devices capable of interacting with one another directly or indirectly and also data
collection are performed locally or remotely via centralized servers or cloud-based
applications. This paper aims to design a low-cost energy monitoring and controlling
system using IoT devices. This paper shows a result of the effective use of IoT devices
in a power system to obtain better efficiency with less time. The use of IoT devices
improves the power system performance remotely without any human intervention.
In this project, the prototype of the system using Raspberry pi has been designed. The
use of Raspberry pi reduces manpower and maintenance cost. It performed mainly
two functions such as oil quality and oil level sensing and transformer differential
protection.

Keywords Internet of things · Power system monitoring and controlling · Oil

quality and quantity sensing · IoT based differential protection

1 Introduction

Remote monitoring and controlling the sub-station equipment is an essential issue

for the power/energy management department which is normally done manually or
using an expensive PLC and SCADA system. With the emergence of the internet
and computational era, a smart monitoring and reliable controlling system over the
entire sub-station equipment is highly desirable that can be achieved by introducing

P. Bodke (B) · A. A. Kalage

Sinhgad Institute of Technology, Lonavala, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 133
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_14
134 P. Bodke and A. A. Kalage

the Internet of Things (IoT) technology [1, 2]. IoT is the network of physical devices
embedded with electronics, software, sensors, actuators and network connectivity
which have the ability to identify, collect and exchange the data. Each thing is
uniquely identifiable through its embedded computing system and able to inter-
operate within the existing internet infrastructure [3]. An IoT based network strategy
for monitoring and controlling the sub-station equipment provide efficient time and
resource management [4, 5]. This project mainly focused on to indicate the quality
and present amount of oil in the transformer & oil circuit breaker (OCB) at a remote
location without being physically present so that proper corrective actions must be
taken by the operator. This proposed system represents an alternative way to provide
differential protection where the distance between two sides of the equipment does
not affect system performance.
This system’s objectives are to proposed IoT-based smart monitoring & control-
ling system using Raspberry-Pi module, which will enable the large data storage on
the cloud at a very low cost so that live tracking can be done.

2 Proposed System Block Diagram

Figure 1 shows the block diagram of the proposed system. The system aims to provide
an operator at a remote location with sufficient information to determine the quality
and present amount of oil in the transformer & oil circuit breaker (OCB).
Even though monitoring & controlling, the system represents an alternative way to
provide differential protection where the distance between two sides of the equipment

Fig. 1 Proposed system block diagram

Design of an IoT Based System for Monitoring … 135

has no effect on system performance. An alarm message with what corrective action to
be taken will be provided to the operator at any dangerous condition. Any authorized
person can get mail alert & perform the same tasks by visiting the power system’s
official webpage. It uses the different sensors to sense the transformer oil quality and
quantity as well as the current sensor has been used, and the data is sent to the cloud
server through Wi-Fi network. The advantage of this system is that the data can be
saved and the real-time data can be obtained at any time. Once the data is received
it analyses and checks for the respective action and signal are sent to the actuator.
The major part of the system is as follows:
A. Sensor
The sensors are used here to sense the transformer/OCB oil quality and oil level
also the current sensor to sense the current of a current transformer.
• Transformer or OCB Oil Level Sensor: The HC-SR04 Ultrasonic distance
sensor is used to sense the oil level of the transformer or OCB.
• Transformer oil quality sensor: The TSSP40 infrared sensor is used. An IR
sensor consists of an IR LED and an IR Photodiode. When the IR transmitter
emits radiation, it reaches the object, and some of the radiation reflects back
to the IR receiver.
• Current sensor: The ACS712 is a fully integrated, hall effect-based linear
current sensor with 2.1kVRMS voltage isolation and an integrated low-
resistance current conductor.
B. Actuators
Actuator performs the tasks that are required by the operator at a remote loca-
tion. Servo motor is connected to Raspberry-pi module which performs the
operation as per the instructions received.
C. IoT based differential protection
The differential protection is one of the most commonly used protection
schemes for electrical equipment in power system substation. But due to the
cost of pilot wire, this protection scheme can’t be useful for every system.
One of the major disadvantages of pilot wire differential protection is that as
the length of pilot wire increases the effect of capacitance on system perfor-
mance increases. IoT based differential protection is a smart solution where
two modules send the secondary current data to the webserver. When the data
coming from two different terminals of the protected zone differs from each
other beyond permissible limit, the webserver sends a command to an internet
module which initiates the CB operation to protect the system. So thus, the
problem associated with a pilot wire can be eliminated. To implement practi-
cally IoT based differential protection, two current sensors readings are taken
and the difference of this current is continuously monitoring.
136 P. Bodke and A. A. Kalage

Fig. 2 Proposed system circuit diagram

3 IoT Technology

Internet of things IOT consists of two words Internet and Things. The term things
in IoT refers to various IoT devices having unique identities and have capabilities to
perform remote sensing, actuating and live to monitor of certain sort of data [6, 7].
IoT devices are also enabled to have a live exchange of data with other connected
devices and application either directly or indirectly, or collected data from other
devices and process the data and send the data to various servers.

4 Proposed System Circuit Diagram

In this system, we are using a Raspberry-Pi 3 Model B+ controller with IoT tech-
nology using Wi-Fi 8266 module. From Fig. 2, the ultrasonic sensor and current
sensor, both data’s are stored in the cloud, and these data’s are controlled using
Raspberry-Pi. Finally, the sensors provide the data continuously to the consumer [8].

5 Proposed Methodology

The proposed system is to monitoring and controlling of substation equipment

remotely using IoT. After the successful connection of all the sensors to the rasp-
berry pi module system check for instructions continuously [9]. Once the fault occurs,
instructions are received by mail and corrective action must be provided by actuators.
This module sends the data’s to raspberry-pi through IoT and to operate using Email
or web page. This IoT system monitoring the substation parameters like current,
voltage, and oil level inside the transformer or OCB and provides continuous data to
Design of an IoT Based System for Monitoring … 137

the operator. After getting the correct data from the sensors, they can easily able to
operate substation parameters such as CB and relay [10]. For data exchange MQTT
protocol is used. MQTT is fast becoming one of the main protocols for IoT (Internet
of Things) deployments. The Linux operating system is used.
The proposed system uses different sensors to sense the transformer oil quality
and quantity as well as the current sensor has been used, and the data is sent to the
cloud server through Wi-Fi network. The advantage of this system is that the data
can be saved and the real-time data can be obtained at any time. Once the data is
received, it analyses and checks for the respective action and signal are sent to the
actuator.

6 Results

After the successful connection of communication devices to the server, oil tank
vacancy level and CTs current are sent to the webserver in an automatic manner. At
the violation of their normal values, an alarming message with a corrective step to be
taken to solve the specific problem was sent to the webserver. We can continuously
monitor the substation parameters on ThingSpeak UI.
Figures 3 and 4 show the command Email in order to control the servo motor (CB
state) and to change the tap position of the transformer from the webpage. Based on
the experiment results, it is terminated that the proposed new IoT based sub-station
controlling and monitoring system worked successfully.

Fig. 3 Email alert for current differential protection

138 P. Bodke and A. A. Kalage

Fig. 4 Email alert for low oil level indication

7 Conclusions

In this paper, the IoT based smart system has been developed for remote monitoring
and controlling the entire sub-station equipment, which is very reliable, user-friendly
and low cost as compared to the conventional system. The proposed approach is
an absolutely automatic system that includes self-checking of oil level from the
transformer/oil circuit breaker, continuous sensing of two CTs secondary current,
sending data to the webserver, storing and displaying data in the web page, and
sending a comment to an internet module for performing the specific task such as
initiating the CB operation, tap changing the transformer, and so on.
Also, it created a new pathway to take necessary measures priory in case of any
emergencies that may occur in substations. In future, the system can be enhanced
with additional features for automation of the tasks at the substation.

References

1. Hossain MS, Rahman M, Sarker MT, Haque ME, Jahid A A smart IoT based system for
monitoring and controlling the sub-station equipment
2. Balamurugan S, Saravanakamalam D (2017) Energy monitoring and management using
internet of things. In: Proceedings of the international conference on power and embedded
drive control (ICPEDC), Chennai, pp 208–212
3. Morello R, Capua CD, Fulco G, Mukhopadhyay SC (2017) A smart power meter to monitor
energy flow in smart grids: the role of advanced sensing and IoT in the electric grid of the
future. IEEE Sens J 17(23):7828–7837
4. Hlaing W, Thepphaeng S, Nontaboot V, Tangsunantham N, Sangsuwan T, Pira C (2017)
Implementation of Wi-Fi-based single-phase smart meter for the internet of things (IoT). In:
Design of an IoT Based System for Monitoring … 139

Proceedings of the international electrical engineering congress (IEECON), Pattaya, pp 1–4

5. Arun Chandra P, Mohith Vamsi G, Srimanos Y, Mary GI (2018) Automated energy meter using
Wi-Fi enable raspberry Pi. In: IEEE international conference on recent trends in electronics,
information & communication technology
6. Ramakrishnan R, Gaur L (2016) Smart electricity distribution in residential areas: internet of
things (IoT) based advanced metering infrastructure and cloud analytics. In: Proceedings of
the international conference on internet of things and applications (IOTA), Pune, pp 46–51
7. Pau M, Patti E, Barbierato L, Estebsari A, Pons E, Ponci F, Monti A (2018) A cloud based
smart metering infrastructure for distribution grid services & automation. Sustain Energy Grids
Netw 15
8. John A, Varghese R, Krishnan SS, Thomas S, Swayambu TA, Thasneem P (2017) Automation
of 11 kV substation using raspberry pi. In: Proceedings of the international conference on
circuit, power and computing technologies (ICCPCT), Kollam, pp 1–5
9. Li L, Ota K, Dong M (2017) When weather matters: IoT-based electrical load forecasting for
smart grid. IEEE Commun Mag 55(10):46–51
10. Chooruang K, Meekul K (2018) Design of an IOT energy monitoring system. In: International
conference on ICT & knowledge engineering, November 2018
Implementation of Iridology for Pre
Diagnosis of Brain Tumor

Pragtee Tathe, Mohua Biswas, Anup Vibhute, Geeta Unhale,

Mrunmayi Raut, and Papiya Biswas Datta

Abstract Human body is the magical creation of god. It carries many intercon-
nected systems. Changes in one system replicate changes in another system. Due to
such interconnection of systems, we can analyze one system by observing changes
in another system. Iridology supports the same theory. Iridology tells the relation
between iris and other systems present in the body. Evaluation can be done in the
form of the iris that speaks about the physical condition of different body parts. In
the proposed method by observing different iris images without doing any compli-
cated and time-consuming test, we can perform diagnosis for the brain tumour. This
method can be used as a pre diagnosis tool.

Keywords Iris · Iridology · Pre diagnosis · Brain tumour

1 Introduction

One of the best creations of god is the human body. The human body contains many
organs, and these organs are linked with each other. One of the important organs in
the human body is the eye which contains many parts like sclera which is whitish
outer background, iris which is middle central ring may be brown, black or blue and
the black centre pupil. The branch of science which focuses on the study of the iris is
called Iridology. Iridology, also known as iridodiagnosis, is a technique in which the
iris images are analyzed with different aspects. The technique of iridology is based

P. Tathe (B) · M. Biswas · A. Vibhute · G. Unhale

SVERI’s College of Engineering, Pandharpur, India
e-mail: [email protected]
Solapur University, Solapur, India
G. Unhale
e-mail: [email protected]
M. Raut
FTC COER Sangola, Solapur University, Solapur, India
P. B. Datta
Devi Mahalaxmi Polytechnic College Titwala, Kalyan, Thane, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 141
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_15
142 P. Tathe et al.

Fig. 1 Iris chart for right and left iris

on the theory that different body parts are related to different regions in the iris. The
human iris is categorized into 90 different regions, and each region replicates the
status of different human body parts. Whenever there is a change in the body organs,
it is reflected in the iris. The form of the iris varies according to the condition of body
parts. The right eye iris replicated the status of right side body organs and left eye
iris replicates the status of left body organs [1, 2]. With several nerves connected,
the iris is linked with body parts as shown in the iris chart in Fig. 1. It indicates the
distribution of different organs over the iris circle.
The well solidly organized fibers of iris are the indication of a healthy body. The
unhealthy body has disturbed solidly organized fibers and makes them slackly as
shown in Fig. 2.
A brain tumour is a major disease that has affected over many million people across
the globe; the rate of people getting affected will exponentially increase in the coming
years. There are different methods for detecting the brain tumour like MRI scan, CT
scan etc.; these techniques are time-consuming; it also needs specially trained persons
to handle it. The major thing related to these techniques is that the patients at the initial
stage are not going to look forward towards these techniques. The proposed project
aims to construct a graphical user interface that allows any physician to predict brain
health with user-friendly and fast diagnosis techniques. These techniques integrate
various image processing steps together in order to complete the process of diagnosis
as shown in Fig. 3.
Implementation of Iridology for Pre Diagnosis of Brain Tumor 143

Fig. 2 Schematic representation of normal and abnormal human eye

2 Methodology

A. Capturing of eye images

In this approach, the images of the eye are collected by using a high-resolution
camera and with these images a database is formed as shown in Fig. 4. The database
is segregated into two parts: one part contains images of iris of a healthy person, and

Fig. 3 Methodology for iris

diagnosis 1. Capturing of Eye Image

2. Pre-processing of Eye Image

3. Sectionalisation

4. Standardisation of iris Image

5. Taking ROI

6. Feature Extraction

7. Classification
144 P. Tathe et al.

Fig. 4 Captured eye image

another part contains images of iris of brain tumour suffered person. The database
can also be formed by using images collected from an online database such as CASIA
database, MMU database, and UBIRIS database [1]. Normally, in-hospital ophthal-
mology department is dealing with the iris images. For the proposed work database
is created by taking the total 60 images out of those 30 images are of a normal
person, and 30 images are of persons with a brain tumour. The below figure shows
the captured image of an eye.
B. Pre-processing of eye images
The pre-processing is the next step in the methodology of the proposed system. In this
stage, whatever the captured image is there, it has to pass through various filtering
algorithms. After passing through these filtering algorithms, the image quality gets
improved because this stage scales back the presence of noise within the iris image.
This stage is very important to make end results more relevant than first. For the iris
images, adaptive filtering is used [1, 3–5].
C. Sectionalisation
Sectionalisation is the process to seek out interior and exterior precincts of the iris.
The circular Hough transforms often won’t to comprehend the radius and centre
coordinates of the pupil and iris sections. By subtracting pupil from the sclera, we will
be able to fix the iris. Once the iris region is segmented from an eye, the subsequent
step is to modify the iris region into fixed dimensions. After subtraction, we will get
the iris pattern into a circular shape ring [1, 6, 7].
D. Standardization of the iris image
Daugman’s rubber sheet model (as represented in Fig. 5) is employed for standard-
ization of Iris ring after standardization of iris ring; the ring is transformed into a
rectangular shape were to represent the points on the ring we can use x and y coordi-
nates. The relation between preceding and the following image is specified with the
relation between the coordinate systems [1, 3, 8].

Fig. 5 Daugman’s rubber sheet model

Implementation of Iridology for Pre Diagnosis of Brain Tumor 145

Fig. 6 Taking out ROI from a standardized iris image

E. Taking out ROI

With reference to the point mentioned in the introduction part that different sections
of the iris are related with specific body parts so for performing analysis of brain
tumour we need to crop the area from the normalized iris. As we have divided the
iris area into sectors, we only need to crop the sector, which represents the brain
area. In the circular iris, if we consider a specific point as a starting point then with
respect to that, we have to select coordinates for a brain tumour. Depending upon
which point we have selected as a starting point to convert into a rectangular shape,
the coordinates values are going to change. If we compare the iris with the image of
a clock, then the zone from 11 to 1 represents the brain area. So accordingly, we can
take the region of interest [2, 9] as shown in Fig. 6.
F. Feature extraction
Once we get the significant region subsequently, we can find out different
features for that region. The extracted features are as follows [1, 3].
1. Mean: The average level of intensity of the image.


G−1
Mean: μ = i p(i)
i=0

2. Variance: Variance describes the variation of intensity around the mean


G−1
V ariance: σ 2 = (i − μ)2 p(i)
i=0

3. Skewness: It is the measure of the asymmetry of the probability distribution

of the random variable.
146 P. Tathe et al.


G−1
Skewness: μ3 = σ −3 (i − μ)3 p(i)
i=0

4. Entropy: It is the amount of information present.


G−1
Entr opy: H = − p(i)ln[ p(i)]
i=0

5. Kurtosis: It is the measure of the flatness of the histogram of the image.


G−1
K ur tosis: μ4 = σ −4 [(i − μ)4 p(i)]−3
i=0

6. Energy: It is the measure of the brightness of the histogram of the image


G−1
Energy: E = [ p(i)]2
i=0

G. Classification
Once we have calculated different features for normal as well as abnormal iris ROI
images, the next step is to train the classifier. Nowadays there are different classi-
fiers which we can use, but out of all support vector machines is more preferable
because it has the kernel trick to transform our data and with the help of which an
optimal boundary is detected for preferable outcomes. So nonlinear SVM classified
the images into two classes; one indicates healthy human iris, and other indicates iris
affected by a brain tumour [1, 10].

3 Result

Figure 7 shows the various operations which are mentioned in the methodology on
normal eye images. If we consider the sequence of operations, then in the first step
eye image is taken with the help of the high-resolution camera. Then that image is
converted into grayscale or in black and white format. Then using the segmentation
method iris area is detected. Once the iris ring is detected, then it is converted into
a rectangular shape, and then the specified area for a brain tumour is separated
by the cropping method. Once we get the ROI then the different features listed in
methodology are calculated, Table 1 shows the values of different features.
In Fig. 8 once our features get calculated depending upon the values, classifier
gives the result. Figure 9 shows the result of the classifier for normal Iris images.
Figure 10 shows the various operations which are mentioned in the methodology
Implementation of Iridology for Pre Diagnosis of Brain Tumor 147

Fig. 7 Various operations on input eye image to find out ROI of brain tumour

Table 1 Values of different

Features Values
features calculated for normal
iris Mean 94.9621
Standard deviation 64.7603
Entropy −0.3473
Smoothness −0.0645
Skewness 0.0319
Contrast 1.0000
Kurtosis 0.0508
Homogeneity 0.0145
Variance 0.5039

on brain tumour affected people eye images. If we consider the sequence of opera-
tions, then in the first step eye image is taken with the help of a high-resolution camera.
Then that image is converted into grayscale or in black and white format. Then using
the segmentation method iris area is detected. Once the iris ring is detected, then it is
converted into a rectangular shape, and then the specified area for a brain tumour is
148 P. Tathe et al.

Fig. 8 Colour feature and texture feature distribution for normal iris images

Fig. 9 Decision of classifier

separated by the cropping method. Once we get the ROI then the different features
listed in methodology are calculated, Table 2 shows the values of different features.
Figure 11 shows the plot of values of different features calculated for abnormal
iris ROI image.
Once our features get calculated depending upon the values of features, classifier
gives the result. Figure 12 shows the result of the classifier for brain tumors affected
person’s Iris image.
Once we train the classifier with 30 iris images and we check the system for the
remaining 30 images then out of that for 25 images, we will get accurate results. So,
we can say that for the above system we get about 83% of accuracy.
Implementation of Iridology for Pre Diagnosis of Brain Tumor 149

Fig. 10 Various operations on input eye image to find out ROI of brain tumour

Table 2 Values of different

Features Values
features calculated for brain
tumor affected iris Mean 121.5513
Standard deviation 68.3303
Entropy −0.4531
Smoothness −0.0718
Skewness −0.0317
Contrast 1.0000
Kurtosis 0.1207
Homogeneity 0.0102
Variance 0.5177

4 Conclusion

About brain tumour as pre diagnosing stage in an easy and accurate way. As this
technique is non-invasive and less time consuming, it is user friendly to the patients.
If we train the classifier with different variety of iris images, we can also move
towards further accuracy. The proposed system is developed for the pre diagnosing
150 P. Tathe et al.

Fig. 11 Colour feature and texture feature distribution for brain tumour affected iris image

Fig. 12 Decision of
classifier

of brain tumour using Iris images by using MATLAB simulation software. This
method allows any physician to predict.

References

1. Chaskar UM, Sutaone MS (2012, Apr 3) On a methodology for detecting diabetic presence
from iris image. IEEE Xplore. ISSN 12654029
2. https://www.irisdiagnosis.org. Accesses May 2010
3. Tathe PB, Patil MM (2015) Analysis of health condition based on iris image. IJIERT
ICITDCEME-2015. ISSN: 2394-3696
4. Zheng J (2018) An iris recognition algorithm for identity authentication. IEEE Xplore. https://
doi.org/10.1109/ICITBS.2018.00162
5. Kaur N, Juneja M (2014) A review on iris recognition. IEEE Xplore, INSPEC Accession
Number: 14254396
6. Lozej J, Štepec D (2019) Influence of segmentation on deep iris recognition performance. IEEE
Xplore. https://doi.org/10.1109/IWBF.2019.8739225
Implementation of Iridology for Pre Diagnosis of Brain Tumor 151

7. Jayalakshmi S, Sundaresan M (2013) A survey on iris segmentation methods. In: International

conference on pattern recognition, informatics and mobile engineering ear: 2013. Conference
Paper, Publisher IEEE
8. Mohammed A, Al-Gailani MF (2019) Developing iris recognition system based on enhanced
normalization. IEEE Xplore INSPEC Accession Number: 19029194
9. Shanker NR, Ezhil A, Archana S (2012) Non-invasive method of detection of cholesterol using
image processing. Int J Med Eng Inf IJMEI2012.048384
10. Ali H, Salami MJE (2008) Iris recognition system by using support vector machines. IEEE
Xplore, INSPEC Accession Number: 15385486
Wireless Communication Using Light
Fidelity Network

Nimisha Deval, Prajakta Satarkar, Akshata Jadhav, and Rupali M. Shinde

Abstract In current period, Light fidelity network is becoming very famous and
catching the interest of some of customers with that one advanced technology-based
totally features. It brings a brand new decision inside the Wi-Fi communication
exchange and those are showing so much curiosity to recognize approximately it.
Light fidelity is a wireless communication-primarily based expertise which transmits
records over the community through a light source like led’s as opposed to radio
frequency indicators (RF) with very excessive information charge. The aim of paper
Wireless and optical networks are widely used nowadays because the network perfor-
mance is a very important issue to supply services to a good variety of users whereas
reassuring users’ quality of service necessities. It aims to analysis the wireless and
optical networks performance. The learning toward research the act as an Light
fidelity Network system in wireless communication consumes remained created and
achieve a knowledge speed of 10 k rate over 40 cm distances.

Keywords Li-Fi · Network performance · Optical transmission system ·

Simulink · OFDM

1 Introduction

Inside the period of overloaded (statistics conversation) global, light fidelity stands
an innovative technique of wireless communication exchange like stands usages
managed lighting to transfer information wireless. The communication of statistics
remains one among maximum essential each day accomplishments within imme-
diate developing international. The present wireless communication networks that
is connect North America Country toward net square measure very gradually while
multiple gadgets. Also using amount of tools which get entry to the remaining, the

N. Deval (B) · P. Satarkar · A. Jadhav · R. M. Shinde

SVERI’s College of Engineering, Pandharpur, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 153
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_16
154 N. Deval et al.

availability of secure bandwidth varieties it a good deal greater hard to experience

high records switch fees and to attach a comfy network. Li-fi consumes a miles wider
range for communication compare toward traditional procedure of ratio communica-
tions network that depend on wireless influences. In the simple system over due this
period is that the statistics may be transmitted done controlled light through using
variables light strengths quicker than the social appreciations container observe. This
era makes use of part of the spectrum that is withal not will utilize- the color spec-
trum, in preference to gigahertz wireless communication waves used for proceedings
control.
The knowledge of li-fi became brought aimed at the main time via a German
physicist Harald Hass within the TED Laboratory international speak scheduled
seen wireless communication exchange (VLC) in July 2011, via mentioning toward
the situation as “information through illumination”. They said a standard lamp with
an light to communicate a audio visual of a flourishing bloom that convert then
expected on an show. Now easy relations, li-fi can remain through to light-based
totallyWi-Fii.e.as an alternative of wireless communication waves it makes use of
light to communicate information. Now room of Wi-Fi modems, light fidelity might
us age transmit prepared through controlled able lamp that in addition to transmit
and acquire records [1].
Through including innovative and utilized bandwidth of evident light to the at
present to be consumed wireless communicate impressions for Statistics switch, li-fi
can performance am a in purpose in releasing the light masses which the cutting-
edge wireless appliance is finish. Hence the situation can suggestion extra occurrence
ensemble of the direction of 400 THz as associated to that to be consumed in Radio
Frequency communicate which is around 300 GHz. Similarly, because the li-fi vari-
eties use and observable range, the situation resolve assist improves apprehensions
that the optical can be convey as a waves approaching through wireless fidelity should
to unfavorably have an effect on our health [2].
By means of conversation via visible light communication, li-fi era consumes the
opportunity to exchange in what way charge to the remaining, movement movies,
acquire emails and plenty more. Protection could no longer be a effort as data cannot
be retrieve intimate the absence of light. As per a conclusion consequence, the situ-
ation can be used in excessive safety services areas wherein RF communicate is at
risk of bug somebody’s room.
To aim of paper to analysis the performance optical and wireless network built
proceeding optical transmission method analyze the parameter of performance
network in Light Fidelity [3].
Wireless Communication Using Light Fidelity Network 155

2 Working

2.1 Visible Light Communication

To purpose of paper toward study the presentation optical and wireless network
established on optical transmission method analyze the parameter of performance
network in Light Fidelity as shown in Fig. 1 [3].

2.2 Methodology

Li-Fi primarily built communication method is totally different from light communi-
cation system as a result of Visible Light Communication is simply a point-to-point
communication system. If Li-Fi technology into pictures then by using sender to
receiver through digital transmission [4].
The working principle is simple are using optical communicate to sender and
receiver through messages shown in Fig. 2.

Fig. 1 Light spectrum Wi-Fi or Li-fi

Fig. 2 Optical communicate

to sender and receiver
through messages
156 N. Deval et al.

Fig. 3 Components of light

fidelity

2.3 How Li-Fi Work

There running of li-fi as following. There are two simple components are used:
led light source and light sensor (picture-detector). The light source is picture and
detector on every other end. When led bulb is on it starts off evolved sparkling and
the light sensor senses the light from light source and gets signals now the shape of
binary signals i. e. 1 or 0 [5]. When a few statistics is transmitted over the community
from the net to the consumer tool, it transmits over the network and flashing of led
bulb is a sign of the message after which picture detector feel mild and get hold of
the message and ahead to its vacation spot end by way of displayed in Fig. 3.
• LED, Photo Detector (Light Sensor), Optical Wireless Communication (OWC),
End-user Devices (Laptops, computer), Radio Frequency Signal’s Amplifier, Line
of sight.
More the brighter LED bulb, More frequently data will transmit light signals over
the network and highly reliable.

2.4 Wi-Fi Versus Li-Fi

Comparison of Wi-Fi and Li-Fi [6] is mentioned in Table 1.

Wireless Communication Using Light Fidelity Network 157

Table 1 Parameter
Parameter Wireless fidelity Light-fidelity
Speed 54–250 Mbps 1–3.5 Gbps
Ranges 20–100 m 10 m
IEEE standard 802.11b 802.15.7
Network topologies Point to multipoint Point to point

2.5 Speediness

For Wireless Fidelity, we have rate margins for information transmission Although
Light Fidelity offers exceptionally high-speed of the network, and that we resolve
transfer large records in exactly limited minutes of your period. Quickness for Li-Fi is
10,000 creases quite wireless fidelity abundant more than 1 Gbps are often achieved.

2.6 Security

Radio waves will penetrate the complete manner via walls. These ends in varied
safety worries as they’ll be intercepted while not issue. At an equivalent time as facts
switch aimed at light fidelity is extremely enclosed then innocent. As we to tend
to use seen light-weight not at all sign separates via partitions. Such seen delicate
communication may be used firmly in airplanes without worrying airways indicators.

3 Related Work

The author described this report possesses statistics approximately the visible mild
communicate the use of li-fi technology. This modern-day generation li-fi describes
transmission of “records via illumination” taking the fiber out of fiber optical by way
of sending records through a controlled light bulb that varies in intensity quicker than
the human eye can observe. Discuss of this paper on a Visible Light Communication
(Li-Fi) performed work done by Wate [3].
The author described li-fi is a label for wireless-communication structures using
mild as a service rather of traditional radio frequencies, as in wireless. Li-fi has the
benefit of being able to be used in touchy regions along with in aircraft without
inflicting interference. Discuss of this paper on a The New Era of Transmission and
Communication Technology: Li-Fi (Light Fidelity) LED and TED Based Approach
performed work done by Prakash [7].
The author defined the short boom in devices utilization and application devel-
opments has brought on many researchers to appearance past wireless. The oppor-
tunity of the usage of visible mild for brief variety verbal exchange is explored with
158 N. Deval et al.

the aid of many researchers. Discuss of this paper on a Performance Analysis of

Visible Light Communication System for Free Space Optical Communication Link
performed work done by Shamsudheen [5].
The term li-fi refers to visible mild communicate (vlc) technology that makes use
of as medium to deliver excessive-velocity communicate in a way just like wireless.
Discuss this paper on a Li-Fi–The Future Technology in Wireless Communication
performed work done by Chavan [8].
The author defined this era is quality for high density wireless facts communicate
in restricted area and to overcome radio intervention trouble. In this paper on a Li-
Fi—A Revolution in the field Of Wireless-Communication performed work done by
Parth Vora [1].

4 Li-Fi Network

In light-Fidelity might be a wireless communication expertise, that is uses light to

communicate information then situation among procedures. Now this real-world
passé Li-Fi is a light communiqué method that is skilled and transferring data at
great quickness finished the visible light communication (VLC). On dealings of its
completion users, the expertise are related to Wireless Fidelity the key procedural
transformation existence that Wireless Fidelity uses to radio rate to communicate
information.
The area is struck by number of light attachment, which delivers light for radiance.
Every graceful is focused through a Li-Fi modem before a Li-Fi chip and, then,
similarly works as per a visual improper location otherwise access point (AP). The
optical wireless network positions are related to the core network by great rapidity
compliment networks. The mild fixtures additionally need to include electromagnetic
sensor toward acquire indicators since such stations. The revealing light illuminations
are controlled at in elevation amounts as shown in Fig. 4.

Fig. 4 Light Fidelity environment network

Wireless Communication Using Light Fidelity Network 159

The resultant great radio rate of recurrence flickers which remain abundant devel-
oped than the recharge amount of a laptop display are not evident to the residents of
the area.
Influence as fine as records dismiss remain condition to each light fixture using a
quantity of dissimilar methods.

5 Features of Li-Fi

5.1 Volume

In Wireless communication evidence is communicated over wireless influences that

square measure limited and valued. It’s a restricted information amount, concerning
Li-Fi. By the quickly increasing real world growth of capability like now in era, and
then arranged we have a trend to square measure organization out of range [8].

5.2 Effectiveness

Everywhere stand an out sized variety of cellular wireless improper positions that
put away large amount of drive. Record of the drive is working aimed at conserving
depressed the end position rather than communication. So, effectiveness of such
Radio corrupt positions is extremely little.

5.3 Safety

Numbers appropriation or riding is insignificant compared to Wireless Fidelity

meanwhile the vary of data communication is limited to exact universe and optical.

6 Applications

6.1 RF Range Assistance

Extra ability difficulties of cell systems could also remain fresh burdened to light
fidelity networks wherever offered. That’s especially powerful at the allow wherever
restricted access tend to sit down [9].
160 N. Deval et al.

6.2 Smooth Illumination

Some personal or community light beside road spots may remain used to deliver
li-fi hotspots and then the equal transportations then method foundation could also
remain accustomed show and manipulate illumination and information.

6.3 Harmful Surroundings

Light Fidelity delivers a secure different toward magnetism intervention from

incidence infrastructures in surroundings like sources then organic compound
flowers.

6.4 Clinic and Healthcare

Li-Fi produces no magnetic force intervention consequently doesn’t interfere with

medicinal instruments, neither remains the situation affected through by tomography
scanners [6].

6.5 Subsurface Infrastructures

Since of sturdy sign interest in seawater, roof use is impractical. Audio waves must
extraordinarily small bandwidth then interrupt sea life. Li-fi presents an answer aimed
at quick-variety infrastructures.

6.6 Transportable Connectivity

Workstations, functional phones, doses then different cell devices will interconnect
directly the employment of li-fi. Fast selection hyperlinks provide terribly excessive
records quotes and additionally offers security.
Wireless Communication Using Light Fidelity Network 161

7 Result and Discussion

A sound permit 3 * 3 arrival connected with1 KB existed transferred. In the research

remained accepted out in a 3 m × 3 m × 3 m possibility below usual day bright
situation. The communicated appearance is shown in Fig. 5 [10].
Sender
The Sender contains is a connected with a laptop in distinction to and which statistics
resolve remain communicated. Facts may stand a man and woman otherwise a writing
record or picture reserved at the processor. On the present art effort, rs232 port srem
ain re used to pass the documents since the pc. Rs232 anchorages may remain set
with MATLAB surroundings then may remain second hand to communicate the
information successively. The controlled use exists a silver managed, producing
comprehension now seen range. The varieties method are used is an on–off keying
(ook).
Receiver
Photograph sensors remain recycled aimed at discovered the gray mild, like incor-
porates the communicated statistics. Picture diodes are used as detectors. The sensed
indicator remains before improved the use of transfer impedance amplifier. Inter-
preting the received statistics is executed on the receiver pc the use of MATLAB
2017 programme [11].
Model standon Matlab Simulink R2017b
The whole Model is an carried out in matlab r2017b simulink software program
software. There’s displayed model of optical transmission machine shown in Fig. 6.

Fig. 5 Transmitted image

162 N. Deval et al.

Fig. 6 Model stand on optical transmission system

8 Conclusion

Light fidelity has standard generation in the area of Wi-Fi information transmission.
It’s miles superior techniques of Wi-Fi Communications that use light as facts carrier.
This paper is part of advent of model stand simulation stage of optical spread method
in matlab Simulink model. To investigate maintain and enhance their community
overall performance light fidelity network.

References

1. Vora P, Bhanushali P, Goswam P (2015) Li-Fi—a revolution in the field of wireless-

communication. Int J J Adv Res Field Eng 5
2. Wu X, Basnayaka D, Safari M, Haas H (2016) Two-stage access point selection for hybrid
VLC and RF networks. In: IEEE 27th PIMRC
3. Wate K, Mattani N, Gole A (2013) Visible light communication (Li-Fi). J: IJERT 2(10)
4. Ahriz I, Douin J-M, Lemoine F, Wei A (2018) Performance evaluations in optical and wireless
networks for project 130
5. Shamsudheen P, Kumar S (2016) Performance analysis of visible light communication system
for free space optical communication link. ScienceDirect 24
6. Chatterjee S, Agarwal S, Nath A (2015) Scope and challenges in light fidelity (Li-Fi) technology
in wireless data communication. J: IJIRAE 2(6) (June)
7. Prakash R, Agarwal P (2014) The new era of transmission and communication technology:
Li-Fi (Light Fidelity) LED and TED based approach. J: IJARCET 3(2)
8. Chavan H, Josh A (2016) i: Li-Fi–the future technology in wireless communication. J: ICI2IM
9. Saeed N, Celik A, Al-Naffouri TY, Alouini MS (2019, Nov) Under water optical wireless
communications, networking, and localization: a survey. In: Science Direct, vol 94 Ad hoc
Network
10. Wang Y, Basnayaka DA, Wuand A, Haas H (2017) Optimization of load balancing in hybrid
LiFi/RF networks. IEEE
11. Šalík P, Róka R, Gorazd T (2018) Simulation platform of optical transmission system in matlab
simulink. ScienceDirect
12. Hadi HA (2016) Wireless communication tends to smart technology Li-Fi and its comparison
with Wi-Fi. J: AJER 5(5)
Smart Trolley with Automatic Master
Follower and Billing System

R. Arpita, K. S. Shashidhara, and Veerendra Dakulagi

Abstract The system “Smart Trolley with Automatic Master Follower and Billing
System”, has been developed for shopping malls. Once the client enters the mall, the
trolley follows the client with the help of Bluetooth and Ultrasonic sensors placed
in the cart. All the trolleys are designed with a RFID reader, digital display screen,
Bluetooth and ultrasonic sensors. All the products within the shopping complex
are connected with RFID tags. Once the customer puts any item into the cart, its
distinctive ID range is detected. At the same time, the chosen product name with the
cost, expiry date, discount amount if any available are displayed on the liquid digital
display screen, thereby the value gets added to the whole bill. If the client wants to
take away the item from the cart, the client must press the button which is available
on the trolley and then the client has to scan the product once again and therefore
the cost of that specific product gets subtracted from total quantity bill. Once the
purchase is completed, the customer has to select the upload button and the whole
quantity bill is send to the billing counter. With the assistance of payment app the
payment is completed. Once the bill amount is paid, the data is distributed to the exit
gate and the trolley gets disconnected to the customer and the client can exit the gate.

Keywords Radio frequency identification reader (RFID) · RFID reader · RFID

tags · Atmega 328P microcontroller · Liquid crystal display (LCD) · Mobile
application

1 Introduction

In Indian cities one can see a very big queue at shopping malls on holidays, festi-
vals and weekends. This queue becomes a lot longer throughout large offers and
discounts. As the client enters the mall one must carry the trolley until the end of

R. Arpita · K. S. Shashidhara
Department of ECE, Nitte Meenakshi Institute of Technology, Bangalore, India
e-mail: [email protected]
V. Dakulagi (B)
Department of E&CE, Guru Nanak Dev Engineering College, Bidar, Karnataka, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 163
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_17
164 R. Arpita et al.

the shopping that becomes difficult for older individuals. When purchase of all the
things is completed the client must reach the billing counter for billing purpose. If
the barcode scanner techniques are used to scan the items, then this may be a time
intense method. So this leads to long queues at the billing counter. Therefore this
system provides solutions to all these problems. The main objective of this system
is to help the older individuals and to reduce the time spent by the customer in the
shopping mall.

2 Related Work

The automatic billing system has been implemented with RFID technology [1, 2].
The add-on features such as included product detection, product recommendation,
budget setting and automatic billing [1]. Further in [3], the android application has
been developed, where barcode scanner was used to scan the products. This paper
projected a way to decrease the time of the client. In [4], the aim of this paper was
to cut back the shopping time. The barcode scanner and the touchscreen display
were used and the payment was done using payment app like Paytm, UPI. In the
paper [5], the aim of the paper was to modify and accelerate the method by the
utilization of RFID tags on every product and Android application connected to the
microcontroller employing a Bluetooth module [6, 7].

3 Proposed System

The complete system contains of AT mega 328P small controller, RFID reader, RFID
tags, 16*2 LCD, ultrasonic sensing element, Bluetooth, WIFI module is as shown
in Fig. 1. This projected system works as client in the shopping mall, she/he first
takes a trolley car. Each trolley car has ultrasonic device and Bluetooth, as he/she
comes closer to the trolley car, the trolley car gets connected to explicit client and
also the trolley car follows the client. Each cart is connected with a RFID reader,
a microcontroller and LCD screen. Once the client starts dropping product into
the trolley car, tags are scan by the reader and also the reader sends the data to the
microcontroller. The microcontroller compares the data with the information already
keep in it. If the information matches then the price of that product are displayed on
the LCD screen for user. If the user would like to get rid of any product from the cart
then they’ll subtract that product from trolley car and value of that specific product
are subtracted from the overall quantity in a flash and the product information with
total quantity gets transmitted to the central Billing system through wireless network
module. With the assistance of the payment app, the overall bill quantity is often paid.
Once the payment is completed, the data is shipped to the exiting gate. Therefore the
gate opens and also the trolley car gets disconnected.
Smart Trolley with Automatic Master Follower and Billing System 165

Fig. 1 Block diagram of the Trolley system

Arduino, UNO, Bluetooth HC-05, RFID reader module, RFID tags, Ultrasonic
sensor HC-SR04, Power supply, WIFI module. Arduino IDE, Google Firebase, MIT
Inventor.

4 Design and Implementation

The functionality of the system has been explained with the help of flow chart shown
in Fig. 2. Initialize the system by scanning the RFID tags to the RFID reader. Once
the tag is read, it reads the memory of the tag from the data which is already assigned
to the particular tags. The cost, name, the expiry date and the discount of the products
gets displayed on the LCD. All the selected products are scanned in the similar order
and the respective cost of all the items are added to the total bill amount. If any items
are to be removed from the cart, the particular item should be scanned again and
hence the price of the product gets reduced from the total bill amount. At the last,
the final amount is displayed and the total bill is generated.

5 Experimental Results

The Drone unit and the Dropping Mechanism unit have been developed and the
system is successfully coordinated. Initially the drone without payload as shown in
Fig. 3 has been tested successfully.
166 R. Arpita et al.

Fig. 2 Flowchart of the system

Fig. 3 Scanned details

displayed on the LCD
Smart Trolley with Automatic Master Follower and Billing System 167

The client enters the shopping complex and he/she picks the smart trolley. Each
smart trolley has ultrasonic sensors and Bluetooth placed in it so the smart trolley
follows that individual client. Every product within the shopping mall is hooked up
with the RFID tags and every RFID tags have distinctive number that is employed
to unambiguously establish the merchandise. Once the merchandise placed upon the
RFID reader the tag gets known by the reader once it comes inside the antenna vary
and so the product-ID beside the end date and value is shipped via server. Google
Firebase is employed to form and maintain the information at the server finish.
At the server finish, the information is compared with the RFID ID range that
is placed on the smart trolley and therefore the details of those things like Tag ID,
price, and expiry date and discount percentage are compared. If it matches then
the information gets displayed on the liquid crystal display screen. The information
communicated is saved for a flash within the small controlled memory. The client
is able to choose the things with the assistance of the information displayed on the
screen.
If the client would really like to get rid of any item then he/she must scan the
merchandise once more that removes the actual item. The actual item and its price
are aloof from the whole bill quantity. The ultimate quantity displays on the liquid
crystal display.
After finishing the shopping, the client must press the “upload” button. Once the
button is ironed, the whole bill quantity of all things gets generated. The client must
select the payment technique and so created invoice is transferred to payment web
site and so the payment is finished.
Once the payment is finished, the knowledge is transferred to the exit gate. Thus
the client has completed the shopping while not standing in long queues. When
the system is initialized the first message displayed is “SHOPPING CHART” as its
initialization message. When the RFID tags of the products are read by the reader,
the details of the product gets displayed and the amount of each product gets added
to the total bill as shown in Fig. 3.
The Firebase Database is used for assigning and storing the value of the products.
Google Firebase is used to create and maintain the database at the server end as
depicted from Fig. 4.
With the help of ultrasonic sensors, the calculation of the distance between the
trolley and the Customer is done and the calculated value gets displayed on the LCD
screen.
The Ultrasonic sensor situated on the trolley handles establishes the connection
between client and trolley with the help of Bluetooth module. Further with the help
of ultrasonic sensors, the distance between the trolley and the Customer is calculated
and value gets displayed on the LCD screen as shown in Fig. 5. The trolley follows
the customers when the Bluetooth gets connected to that particular person and the
trolley follows that client till the end of the shopping with the aid of ultrasonic sensor
as shown in Fig. 5. In Fig. 6, the calculated distance value is displayed on LCD.
The mobile application shown in Fig. 7 has been developed can be used by both
the shopkeeper and the Customer. Using this app, the seller can update the details of
any products whereas the customer can use this app to make the payments.
168 R. Arpita et al.

Fig. 4 Database created for the products

Fig. 5 Trolley following the customer

Fig. 6 The calculated

distance value displayed on
LCD
Smart Trolley with Automatic Master Follower and Billing System 169

Fig. 7 The Login page of

the mobile app

The trolley follows the customers when the Bluetooth gets connected to that
particular person and the trolley follows that client till the end of the shopping. As
shown in Fig. 5. The mobile application shown in Fig. 7 has been developed can
be used by both the shopkeeper and the Customer. Using this app, the seller can
update the details of any products whereas the customer can use this app to make the
payments.
Once the purchase of items is completed, then the bill can be paid using the mobile
app developed. The Bill summary is displayed in the app as shown in Fig. 8.
The items being purchased is logged along with details such as Product name,
Quantity and Total cost as shown in Fig. 8.

6 Conclusion

The objective is effectively attained in the prototype model developed. The developed
product is of low cost, amiable to use and does not require any specific practice.
The ability to take a decision can be done in the cart itself which can be used in
the shopping complexes for effortless and clever way of purchasing items to save
vitality, time and money of the customers. The project is evaluated with different
trial cases with distinct items assessed for all the practical trials. The system ought to
170 R. Arpita et al.

Fig. 8 The bill payment

page of the mobile app

be generated that scale back the time of the customers in looking out the placement
of the product. The client simply sorts the name of the item; the trolley ought to
mechanically guide them to the location of the item. The system ought to be hooked
up with buzzer; if the load of trolley exceeds certain weight unit then it will facilitate
the older customers. Additional economical and long vary RFID readers are often
used. Thefts are often avoided by use of robotic ARM for choosing the items. Since,
it’s a prototype it has limited load carrying capacity around 3 kg.

References

1. Prasiddhi K (2017) Innovative shopping cart for smart cities. In: 2017 2nd IEEE interna-
tional conference on recent trends in electronics information and communication technology
(RTEICT), 19–20 May, India
2. Ravindranath KR, Sanjay AI, Prashant CM (2017) RFID based supermarket shopping system.
In: 2017 international conference on big data, IoT and data science (BID) Vishwakarma Institute
of Technology, Pune, 20–22 Dec
3. Wankhede SS, Nikose A, Radke DP, Khadse DB, Tiwari S (2018) Electronic shopping trolley
for shopping mall using android application. In: Proceedings of the international conference on
communication and electronics systems (ICCES 2018)
4. Viswanadha V, Pavan Kumar P, Chiranjeevi Reddy S (2018) Smart Shopping Cart. 2018 IEEE
5. Kumar A, Srivastava S, Gupta U (2019) Internet of things (IoT) based smart shopping centre
using RFID. 2019 IEEE
6. Raghunandan, Mallikarjun, Satheesh Rao (2017) Billing automation for supermarkets. In: Inter-
national conference on current trends in electronics information and communication technology
(ICCTCEEC-2017), India
7. Sutagundar A, Ettinamani M, Attar A (2018) Iot based smart shopping mall. 2018 IEEE
Analysis and Design of E Shaped Dual
Band Microstrip Textile Antenna
for Wireless Communication

Husain Bhaldar, Sanjay Kumar Gowre, Mahesh S. Mathpati,

Ashish A. Jadhav, and Mainaz S. Ustad

Abstract The microstrip antenna play important role in wireless communication,

because microstrip antennas are light in weight and compact in size. Now days,
the wearable microstrip antenna is emerging technology widely used to perform
the communication between the one human body to another human body and this
technology is called Wireless Body Area Network (WBAN). The wearable microstrip
antennas are used in various wireless applications such as medical, military and
satellite etc. The geometry of the designed microstrip textile antenna consist of E
shaped rectangular patch the top and the full ground plane at bottom. The dimension
of designed antenna is of 35 mm × 31 mm by its width and length. These two planes
are separated by the cotton fabric as dielectric material of (Er ) of 1.6 and the thickness
of 1 mm. In this paper the E shaped microstrip textile antenna is designed which is
resonating at dual frequencies of 3.7 and 7.4 GHz with high directivity. The proposed
antenna is used for wireless communication in S and C band. The proposed antenna
has been simulated in CST software and the various antenna parameters such as
return loss, VSWR and directivity have been analyzed.

Keywords C band · Cotton fabric · E shaped textile · S band and WBAN

H. Bhaldar (B) · A. A. Jadhav

Department of Electronics and Communication Engineering, Bheemanna Khandre Institute of
Technology, Bhalki, affiliated to Visvesvaraya Technological University Jnana Sangama,
Belagavi, Karnataka, India
e-mail: [email protected]
H. Bhaldar · M. S. Mathpati · A. A. Jadhav
Department of Electronics and Telecommunication Engineering, College of Engineering, Shri
Vithal Education and Research Institutes, Pandharpur, India
S. K. Gowre
Department of Electronics and Communication Engineering, Bheemanna Khandre Institute of
Technology, Bhalki, India
M. S. Mathpati
GNDCE Bidar Visvesvaraya Technological University Jnana Sangama, Belagavi, Karnataka, India
M. S. Ustad
SVIPE, ICMS Kasegaon, Pandharpur, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 171
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_18
172 H. Bhaldar et al.

1 Introduction

The microstrip antenna is widely used for wireless application due to its lightweight,
compact size. Due to use of textile material as in the development wearable antennas,
wireless body area network (WBAN) has been rapidly increasing in the wearable
devices which are used for wireless communication in the field of medical, sport wear,
civil and satellite applications. The wearable antennas easily fabricated into clothes.
In this paper E shaped microstrip textile antenna is proposed for dual frequencies for
wireless application at 3.7 and 7.4 GHz. The wearable antenna can be designed by
using various textile materials such as jeans, cotton, polyester, leather and PDMS.
The relative permittivity (Er ) of these textile materials varies from 1.25 to 1.7 except
PDMS material. It is easier now days to study the performance characteristics such
as bending and crumpling conditions of antenna, due to use of textile material in
designing of wearable antenna.
The two frequency bands of 3.7 and 7.4 GHz; these frequencies come under S
band of frequency range 2–4 GHz and the C band of frequency range 4–8 GHz in the
electromagnetic spectrum. The application of S band is used for airport surveillance
system and to control the air traffic. The S band also used for satellite communication
as downlink frequency of range 3.4–4.2 GHz from space to earth station. The C
band is mostly used over Ku band for satellite communication. This band is also
used for multiple applications such as weather forecasting radars and Wi-Fi devices
as IEEE802.11a. The medical accelerators are operated at the S band frequencies.
They are designed at high frequencies to reduce the size and weight. The medical
accelerators are used in radiation therapy field and it has been implemented in the
Intra Operated Radiation Therapy (IORT).
The multiband E shaped microstrip antenna has been designed with ground and
defective ground structure. The proposed antenna is operating at 3 frequency bands
of 3.7, 6.7, 7.9 and 8.7 GHz and antenna provided the 120, 495, 225 and 315 MHz
bandwidths at these resonating frequencies [1].
The proposed antenna has been designed with the rectangular shape which is
resonating at 2.45 GHz frequency and provides the high gain of 8.05dBi. The
designed microstrip textile antenna with jeans as dielectric material been used for
the Wi-Fi modem and achieved the return loss of −15.76 dB [2]. TERA founda-
tion has designed the antenna which is resonating at S and C bands. The designed
antenna used for medical accelerators in hydrotherapy which radiates the protons
or light ions. The designed medical accelerators have been used at 3 and 5.7 GHz
frequencies [3].The designed antenna has simulated at 2.45 GHz with various dielec-
tric substrate materials such as Cotton, Polyester, Cordura and Lycra. The values of
return losses at 2.45 GHz frequency for respective textile material have studied −
32, −35, −29 and −31 dB [4, 5].
The multiband microstrip antenna has designed with moon structure patch on top
and proposed antenna is resonating at the frequencies from 5.3 to 10.15 GHz and
provided the bandwidth of 62.78%, return loss of −25 dB at 5.44 GHz and −24 dB
at 8.05 GHz. The author used the moon strip line structure of patch [6]. The Idellyse
Analysis and Design of E Shaped Dual Band Microstrip Textile … 173

and Rama Reddy have proposed the circular polarized textile antenna and U slot
conical antenna which is designed at ISM band of frequency range 2.4–2.45 GHz,
the values of S11 parameter −35 and −20 dB are observed at resonating frequency
[7, 8].
The proposed multiband wearable antenna has been radiated at frequencies of
3.42, 9.73 and 11.17 GHz with directivities of 3.35, 4.23 and 5.19 dBi [9]. The
author has observed the miniature feeding network for aperture coupled antennas
designed at frequency of 2.4–2.483 GHz with gain of 5.6 dBi. The author have
studied various shapes of coupling aperture such as ring, H-shaped, rectangular, E
shaped, cross etc. at various frequencies in ISM band [10].
The proposed microstrip antenna have designed with circle shaped radiator with
tapered microstrip feed line and has rectangular ground plane. The proposed antenna
used at ultra-wide band (UWB) application for frequency range of 2.8–16 GHz [11].
From the literature survey, it is stated that the gain and directivity of proposed
antenna is enhanced in this paper. The proposed antenna can be used for high
directional antenna for S and C band application.

2 Antenna Configuration and Design

The microstrip textile antenna is most preferred antenna over the printed antenna due
to ease in fabrication and light in weight. The microstrip textile antenna is similar
in structure with simple microstrip patch antenna. The microstrip textile antenna
consist of top patch made with copper foil and bottom ground plane made with
same copper foil of thickness 35 micron. These two planes are separated by textile
material as dielectric substrate as shown in Fig. 1. This methodology of antenna
design is demanding for flexible textile (fabric) antennas, in which designed antenna
can easily attached to clothes and antenna can be wore on the body. This technology
introduced the one body to other body communication called Wireless Body Area
Network (WBAN).
In this proposed study, The rectangular E shaped microstrip textile antenna is
designed for 3.7 GHz frequency also the proposed antenna is resonating at 7.4 GHz
additional frequency band due to structure designed E shaped. The inset line feeding

Fig. 1 Microstrip textile

antenna
174 H. Bhaldar et al.

method used proposed antenna and excited with voltage source connected to the port
of an antenna. The microstrip antenna is designed at the frequency band of 3.7 GHz
for S band and 7.4 GHz for C band wireless communication. The top patch of antenna
consists of E shaped rectangular structure of dimension 35 mm × 31 mm of width
and length. The E shaped slot is designed by dividing the patch width in equal 5
slots which is shown in Fig. 2. The bottom ground plane is designed with full ground
and made of copper foil. The cotton shirt is used as substrate material with dielectric
constant of 1.6 with substrate height of 1 mm, so that designed antenna is wearable,
light in weight and compact in size.
The width and length microstrip patch antenna is calculated for resonant frequency
of 3.7 GHz and dielectric constant of 1.6 by using standard equations of rectangular
microstrip patch antenna design [2]. The Table 1 shows that the list of antenna
parameter used design E shaped rectangular microstrip textile antenna.

a b

c d

Fig. 2 Proposed antenna a E shaped patch of simulated antenna b Ground plane of proposed
antenna. c E shaped patch of fabricated antenna d Ground plane of fabricated antenna
Analysis and Design of E Shaped Dual Band Microstrip Textile … 175

Table 1 List of antenna parameter

Antenna parameter Value
Dielectric constant Er cotton shirt 1.6
Length of patch (L) 31 mm
Width of patch (W) 35 mm
Size of E shaped slot 7.2 mm
Length of ground plane (Lg) 70 mm
Width of ground plane (Wg) 70 mm
Substrate height (hs) 1 mm
Substrate thickness (ht) 0.035 mm

The proposed microstrip antenna resonating at two frequency bands 3.8 and
7.49 GHz. These frequency bands are used for S band and C band communica-
tion. The downlink frequency of satellite communication used in this frequency S
band 3.4–4.2 GHz and the C band mostly used for radar communication 4–8 GHz.
The medical accelerators also used in the S band in the radiation filed theory.
The various parameters of antenna has been analyzed at these two frequency
bands such as return loss, VSWR, directivity and bandwidth as shown in Table 2.
The designed antenna is excited with voltage source connected to microstrip feed
line of thickness around 2 mm of impedance 50 . The proposed antenna have been
designed and simulated in CST Microwave Studio the various antenna parameters
have been observed.

Table 2 Simulated results of E shaped microstrip textile antenna

S. No. Parameters Resonating frequency
F1 = 3.86 GHz F2 = 6.47 GHz
1 S11 dB −33.87 −19.65
2 VSWR 1.14 1.28
3 Directivity in dBi 8.74 7.32
4 Radiation efficiency (%) 115 88.44
5 Total efficiency (%) 33 34
6 Bandwidth 90 MHz 120 MHz
7 %fractional BW 2.33 1.60
176 H. Bhaldar et al.

3 Result and Discussion

(I) Return Loss: The return loss should be kept as low as below −10 dB, to achieve
the best radiation. From Fig. 3a, it has been observed that designed antenna
have two frequency bands at 3.86 and 7.49 GHz with return losses of −33.877
and −19.652 dB. The S11 parameter values of proposed antenna has been
improved and compared with paper implemented. The tested results of fabri-
cated antenna provide the two bands at 2.45 and 3.5 GHz, the wide bandwidth
of 1.9 GHz is obtained at 3.5 GHz frequency band which is shown in Fig. 3b.

The fractional bandwidths are also calculated at these frequency bands using Eq. 1.

%Fractional BW = FH −FL /FC (1)

where FH = High cut off frequency, FL = Low cut off frequency and Fc = Centre
frequency.
The % fractional bandwidths have been calculated 2.33 and 1.6 at these
frequencies which is shown in Table 2 of simulated results.

Fig. 3 Return loss, a Simulated and b Fabricated/tested

Analysis and Design of E Shaped Dual Band Microstrip Textile … 177

Fig. 4 3D radiation pattern the antenna at 2 frequency bands: a 3.7 GHz and b 7.4 GHz

(II) 3D Radiation pattern: From the Fig. 4, it has been observed that 3D radi-
ation pattern of theantenna at resonating frequencies provides the directivity
of 8.74 dBi at 3.7 GHz frequency and 7.32 dBi at 7.4 GHz frequency. These
values of directivity have been improved 7.02–8.74 and 6.37–7.4 compared
with paper implemented and these values fulfil the need of wireless commu-
nication. The proposed antenna can be used for high directional antenna for S
and C band application.
(III) 2D Radiation pattern of proposed antenna: From Fig. 5, it has been
concluded that polar plot of designed antenna has the main lobe magnitude of
8.74 dBi and angular bandwidth of 68.8° at 3.7 GHz frequency and with the
main lobe magnitude of 6.02 dBi and angular bandwidth of 44.1° at 7.4 GHz
frequency.
It has also observed from Fig. 5, that the designed antenna also provided the back
radiation of −21.6 and −15.3 dB at these frequencies.
(IV) VSWR of antenna: From Fig. 6, it has been observed that the value of
VSWR for designed microstrip textile antenna lies in between 1 to 2, which
states the proper impedance matching between excited source and the feed-
line. The values of VSWR at resonating frequencies are 1.14 and 1.28 at 3.7

Fig. 5 2D Radiation pattern of the proposed antenna at 2 frequency bands a 3.7 GHz and b 7.4 GHz
178 H. Bhaldar et al.

Fig. 6 VSWR of proposed antenna

and 7.4 GHz. These vaules of VSWR have been compared with the paper
implemented and it is has been observed values are better.

(V) Efficiency: The efficiency of anenna isdefiend as the ratio of radiated powerof
anantenna to the applied power of antenna. The radiation efficiency antenna
differs from the total efficiency of antenna due to loss takes place because of
the impedance mismatching.
The total efficiency is given by Eq. 2.

Et = Zl ∗ Er (2)

where Et = Total Efficiency, Er = Radiation Efficiency and Zl = Impedance loss.

It has been observed that the proposed antenna radiated at 2 frequancy bands. The
radiation efficiency and total efficiency of antenna at 3.7 GHz are 115 and 33%. The
radiation efficiency and total efficiency of antenna at 7.4 GHz are 88.44 and 33%.
From the values of total efficiency it has been concluded that the ohmic losses takes
place due to impedance mismatch.
The Table 2 shows that the comparison of simulated results of proposed antenna
for two frequency bands.

4 Conclusion

In this proposed study, the E shaped microstrip textile antenna has been designed for
3.7 GHz frequency. The proposed prototype is simulated and analyzed for the various
antenna parameters VSWR, Return Loss and Bandwidth. The gain of antenna has
been improved to good scale at resonating frequencies, if we compared with base
paper implemented. The designed antenna is used for S band and C band commu-
nication. From the simulated results, it has been observed that, antenna is radiating
very properly at 3.7 and 7.4 GHz frequency bands. The total efficiency is reduced
due to impedance loss in comparison of radiated efficiency.
Analysis and Design of E Shaped Dual Band Microstrip Textile … 179

Acknowledgements The author would like to thank the review committee and my guide Dr. Sanjay
Kumar Gowre for his valuable guidance and support. The author also thankful to SVERI’s College
of Engineering Pandharpur for providing support to use antenna laboratory and Research center,
Department of Electronics and Communication, BKIT Bhalki for support and motivation.

References

1. Nagpal A, Dillon SS, Marwaha A (2013) Multiband E-shaped fractal microstrip patch antenna
with dgs for wireless applications. Int Conf Comput Intell Commun Netw 978-0-7695-5069-
5/13 $26.00 © 2013 IEEE. https://doi.org/10.1109/CICN.2013.14
2. Bhaldar H, Goware SK, Mathpati MS, Jadhav AA (2020) Design of high gain wearable rect-
angular microstrip textile antenna for wireless application. Int J Innov Technol Exploring Eng
(IJITEE) 9(5). https://doi.org/10.35940/ijitee.C8478.039520 ISSN: 2278-3075 (March 2020)
3. Mendes C, Peixeiro C (2018) On-body transmission performance of a novel dual-mode
wearable microstrip antenna. IEEE Trans Antennas Propag 66(9) (September 2018)
4. Chen SJ, Ranasinghe DC (2018) A robust snap-on button solution for reconfigurable wearable
textile antennas. IEEE Trans Antennas Propag 66(9) (September 2018)
5. Potey PM, Tuckley K (2018) Design of wearable textile antenna with various substrate and
investigation on fabric selection. In: 3rd international conference on microwave and photonics
(ICMAP 2018), 9–11 Feb 2018
6. Saxena A, Singh VK (2018) A moon-strip line antenna for multi-band applications at 5.44 GHz
resonant frequency. In: 4th international conference on advances in electrical, electronics,
information, communication and bio-informatics (AEEICB-18)
7. Martinez I, Werner DH (2018) Circular-polarized textile based antenna for wearable body area
networks. IEEE Trans 978-1-5386-7102-3/18/$31.00© 2018 IEEE
8. Reddy RS, Kumar A (2018) Dual band circular polarized wearable antenna for military appli-
cations. In: 2018 international CET conference on control & communication, July 2018.
978-1-5386-4966-4/18/$31.00 ©2018 IEEE
9. Purohit S, Raval F (2014) Wearable textile patch antenna using jeans as substrate at 2.45 GHz.
Int J Eng Res Technol 3(5). ISSN 2278-0181 (May 2014)
10. Zhang J, Yan S, Vandenbosch GAE (2017) Miniature network for aperture-coupled wearable
antennas. IEEE Trans Antennas Propag 65(5) (May 2017)
11. Sun Y, Cheung SW, Yuk TI (2014) Design of a textile ultra-wideband antenna with stable perfor-
mance for body-centric wireless communications. Published in IET Microwaves, Antennas &
Propagation. Accepted on 3rd July 2014. https://doi.org/10.1049/iet-map.2013.0658
Enhancement Technique for Early Stage
Lung Cancer Detection Using Foldscope
Methodology

Vanita D. Jadhav and Lalit V. Patil

Abstract Due to ever increasing population and growth in pollution there has been
considerable increase in the cancer-causing agents due to this there is a need to
detect cancer at its early stages so that it becomes easier to cure it. The precautions
to be taken to cure or prevent cancer or other tumor is that cause death very less or
completely negligible this is why there is a need to create a system that can suggest
the early stages of cancer and the precautions that will be taken in order to prevent
this cancer from growing more. Cancer as it is cannot be cured hundred percent
when it is in the moderate stages and above as there are very few treatments for
undertaking or curing cancer completely. We have undertaken the job of detecting
lung cancer as it is the major cause for death in Cancer diseases. Lung cancer when
it is in the initial stages is very accurate and is very difficult to detect and it spreads
rapidly and the chances to stop this growth are very less or completely negligible so
to avoid this, we have decided to create a system that can detect lung cancer at its
early stages. Lung cancer comes in stages and to cure this we need to understand
each and every stage precisely the lung cancer has around Four stages. After looking
at the stages of cancer we can clearly see that the first stage is very important in the
recovery process and for cancer treatment. So, if we detect the cancer at its early
stages it becomes easier for us to cure it and the patient suffers very less due to
this cancer. In some cases, the microscopic examination of this cancer shows wrong
results as the cancer is very less and cannot be easily detected even at microscopic
level. If we record such occurrences it becomes easier for us to detect for the next
time and the chances of error reduces. Here comes the role of Technology and our
system. As we are using machine learning for the development of a system which
is Computer aided design for early lung cancer detection it becomes easier to train
the machine for intangible occurrences of results which cannot be neglected as life
of each and every patient is precious. Our system is completely based on the data
analysis of multiple cases of cancer where lung cancer differs from first stage to

V. D. Jadhav (B)
CSE Department, SVERIs COE, Pandharpur, India
e-mail: [email protected]
L. V. Patil
CSE Department, SKN COE, Sinhgad Road Vadgaon Bu, Pune, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 181
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_19
182 V. D. Jadhav and L. V. Patil

last stage. By considering the occurrences of results we can analyze and design the
pattern for this disease and can easily predict whether the underlying patient being
monitored is suffering from cancer. We are using machine learning based on Python
programming language as it becomes easier to monitor data as well as the detection
process of cancer. Python makes it easier to use machine learning and we can train
the system for the future. Python has multiple directories which we can use for our
project and which makes our project very simple and accurate as we are going to
take the data set which includes arrays of data.

Keywords Image processing · Image enhancement · Foldscope

1 Introduction

Machine learning is a branch of science that deals with training of the machine for
future occurrences. For some cases the cancer detection process gives wrong results
as the cancer occurrence is very rare and for this, we use machine learning which
helps us to detect the cancer for such occurrences in future proceedings. Even for the
microscopy detection of Cancer lungs we need some predetermined data set which
helps us in detection of that cancer so that we can clearly say that it is a normal cell
or paranormal cell. We have taken images of multiple cancer as well as normal cells
and taken the data into consideration of a data set. Due to this multiple use of data
we can clearly say that we have covered multiple cases as the cancer which we are
going to detect is intangible to the system. We have covered around 1000 different
cases where cancer is in early stages.
The data is it is completely unusable when it is not properly managed and precisely
developed. We have taken search algorithms that are very precise in itself and create
very accurate data that can be used easily and precisely to take tests under serious
conditions. This data from the data set is used for matching the occurrences of given
sample images for the system which will detect the occurrences of Cancer if present
or absent. Before creating the detected the images collected of the cells need to be
processed at, they need some sort of image segmentation and image processing as
it contains noise as well as lots of error which are due to the microscopic detection
using human help. Machine learning makes it easier to calculate the situation and
monitor the rare case executions and enter it into the data set as the rare cases should
be properly managed in the future occurrences. In some cases, the cancer in the cells
will not be detected due to its slightly lesser intense present in the cancer cells. These
rare occurrences will be recorded and sent to the data set so that it can store these
occurrences for further convenience and avoid for the inconvenience due to absence
of the data for calculation [1]. Our system will record every occurrence of the result
and send it to the data set so that the system for every next calculation becomes more
pieces and continuously gets more Precision for this testing process. This system will
work fine and will definitely be a boon to the medical world as it is more precise and
can detect cancer at its early stage and make it easier to cure the patient and treat the
Enhancement Technique for Early Stage Lung Cancer Detection … 183

patient easily. This is the reason we have selected this task and problem statement
for making the world a better place.
Foldscope is an origami based paper microscope which can assemble from
punched sheets of card-stock and lens [2]. The foldscope is firstly printed on a
card sheet and according to the marking on the card sheet the foldscope is assembled
the same way as in origami craft works. The lens is also assembled in the fold-
scope. It is used for the magnification required for getting a microscopic view of
the subject or sample being tested by the foldscope user. There are also magnetic
couplers used for connecting a sophisticated camera or other imaging device which
is used for capturing images of videos of the sample. The magnification of the fold-
scope ranges from 140X to 2000X and 2 micron resolution. We made samples of
specimens of different test subjects for the purpose of microscopic testing of the
subjects under foldscope. The created specimen samples were loaded on slide and
then were mounted on foldscope for getting a microscopic view of the samples.
The magnification is 140x+ and resolution is 2 microns [3]. We can actually see and
examine microscopic cellular activities of the test specimens. By this we can propose
the actual underlying abnormality of the cells of the specimens.

1.1 Computerized System for Lung Cancer Detection

Lung cancer remains the leading cause of Cancer related problems in the world.
Early diagnosis can improve the effectiveness of treatment and increase the patient’s
chance of survival. Positron emission tomography (PET), computed tomography
(CT) [4], low dose computed tomography (LDCT), and contrast-enhanced computed
tomography (CE-CT) are the most common noninvasive imaging modularities for
detecting and diagnosing lung nodules [5]. Pet CT scans can be done at various
centers available in the country but due to machinery requirement and cost of the
Machines, very few centers are available [6, 7].
Due to the cost of the machinery and the required technology used for the detection
the cost of this testing is very high and cannot be done by a normal middle class family.
With high cost level how the current systems available make this system completely
invisible to every common man living in the country. Is there a need to make this
system somewhat cheaper so that every citizen in the country being rich or poor can
use this system for their help. Also the reach of the centers is limited to cities and
metropolitan areas. Due to this the people from rural areas and the people living in
the villages and sub urban areas are not benefited by these centers. The average cost
of the pet CT scan at any Centre across India ranges from 15,000 to 20,000 rs for
tests and this does not assure the complete detection of lung cancer [8]. This is just a
preliminary examination done in order to check whether the person’s lung is affected
with cancer or not. Further is the test negative in searching for lung cancer or in case
of unavoidable circumstances which lead to inconsistent results the patient is sent
for other scans which cost lakhs and lakhs of rupees. Even if the patient manages to
collect this money the result would never be different if cases of cancer are very rare.
184 V. D. Jadhav and L. V. Patil

For such cases the money as well as the time that a patient should be treated is wasted
which time could prove to be crucial in the recovery of the patient. To avoid such
circumstances, we have decided to build projects in such a way that every person
living in this country would be benefited by this system. Power system is not much
cost efficient as it requires just one cellular Level Examination of the patient and the
rest of the work will be done by the machine. The system which we are going to
develop will tell you the result of the comparison of the cellular Level examination
of the patient with the predefined data set we have developed in order to detect if
the patient is suffering from lung cancer or not. The cost of the system will not be
more and it will be beneficiary for every patient had the patient will require no direct
visiting to the center of examination [9].
Because the patient can visit the nearby facilities in order to take the samples of
the cells microscopic level and directly send this to the centers so that we can check
the samples and tell whether the patient samples of specimens are suffering from
cancer or affected with any other cases.
This could even be a remote management system for cancer detection and can
prove to be a simple and efficient way for the people living in rural areas or remote
areas where small medical facilities are also not available for cancer detection. The
fight against cancer will become easier and more widely available due to this system.
The reach of cancer awareness and recovery will increase day by day, if this system
comes into existence this is the need for the development of the system.

1.2 Objectives

(a) To implement methodology that works with cellular level detection and
determination of cancer.
(b) To predict the possibility of lung cancer using the data-set facilitated by
machine learning [10].
(c) To compare between cancerous and non cancerous nodules [11].
(d) To create awareness among the masses for early screening of the disease.
(e) To reduce the need for higher-level machinery implementation in order to
develop low cost systems with high output.

2 Methodology

With reference to Fig. 1, we are going to provide the solution to resolve the above
stated problem by designing and developing a CAD system for early lung cancer
detection by inculcating python integrated with machine learning as a key technique
for evaluation of the collected data [12]. Our system has different modules for clear
segmentation of the detection of cells specimen which are stated below:
Enhancement Technique for Early Stage Lung Cancer Detection … 185

Fig. 1 Architecture of
proposed system

A. Search Module

This module is specifically designed and propagated to locate the testing specimen
which is going to be submitted to the system for further evaluation procedure. This
module is a key benefit which facilitates the user to easily locate the test image and
submit it to the system.
B. Image Processing Module

The image of the cell specimen that is passed by the user through the Search Module
is later entered into this module for further processing. The image that is passed by
the search module is not ready for evaluation and needs precise processing. In this
module the image is assessed for noise as well as other impairments and are removed
from the image set. The image is later contrast adjusted for sharpening and better
clearance. Later on this processed image is passed to the next module.
C. Thresholding Module

The image that is received from the previous module is parsed by Otsu’s thresholding
algorithm. Image segmentation takes place in this algorithm and now the image is
ready to be examined and compared with the data-set.
D. Classification Module

Here, the processing ready image is examined and tallied with the data-set. The
image is completely compared with all the existing data-set images and the result
is sent back to the data-set in order to increase the accuracy for further evaluation.
In case, if the result is not found satisfactory then this situation is recorded so that
186 V. D. Jadhav and L. V. Patil

it does not repeat for the next time. In this way the module never makes the same
mistake twice.
Details of the Proposed Implementation
• Search Module

This module is specifically designed and propagated to locate the testing specimen
which is going to be submitted to the system for further evaluation procedure.
• Image Processing Module

In this module the image is assessed for noise as well as other impairments and is
removed from the image set. The image is later contrast adjusted for sharpening and
better clearance.
• Thresholding Module

The image that is received from the previous module is parsed by Otsu’s thresholding
algorithm. Image segmentation takes place in this algorithm and now the image is
ready to be examined and compared with the data-set.
• Classification Module

The image is completely compared with all the existing data-set images and the result
is sent back to the data-set in order to increase the accuracy for further evaluation.
Working
The user needs to browse through the files and select the target image that is to be
entered in the system for the evaluation process. The image that is taken initially
has certain agents of error that need to be processed in order to get a more feasible
image for the testing process. The image also needs to be in a required format as the
prescribed data-set is built of.
Later, when the image is brought in the system then it needs to be further processed
in order to remove the noise and the errors. The noise may be of different sorts like
salt noise, smoke, etc. This noise is removed by using certain techniques that enhance
the image by removing such noise and the errors. The next parameter that is to be
enhanced is the sharpening and the contrast that is to be maintained and enhanced
accordingly in order to make the image more reliable and perfect for evaluation.
After the image processing the, image is next taken into the algorithmic evaluation
so as to increase its compatibility with the system. Otsu’s Thresholding technique is
used for making the image more and more suitable to work with the system. This
algorithm does significant scaling and other required jobs that make the image more
precise and more efficient in terms of details.
This processed image is then taken to evaluate and determine the cancer causing
agents presence in the sampling image. If the given image has some features that are
similar to that of any of the details in the data-set then the system records that data
Enhancement Technique for Early Stage Lung Cancer Detection … 187

and enters it into the data-set for next successive evaluation. All the results that are
evaluated by the system are successively sent to the data-set so as to improve the
quality of the data-set and increase the precision of the system. The more results we
evaluate the more the system works effectively.
The entire development process has been subdivided into two: the front end devel-
opment and the backend development. The front end comprises the visually visible
parts such as the Browse Button, Image Path, Predict Button and the Result. The
back end contains the data-set and its interaction with the trained data.
Front End Development
The front end is coded using PAGE Python GUI Generator. PAGE is a cross-platform
drag-and-drop GUI generator, bearing a resemblance to Visual Basic. It allows one
to easily create GUI windows containing a selection of Tk and ttk widgets. Required
are Tcl/Tk 8.6 and Python 2.7+. I am actually using Tcl/Tk 8.6 and Python 2.7. PAGE
springs from Virtual Tcl, a Tcl/Tk program, forked to generate Python modules that
realize the desired GUI. Tcl is required for running PAGE but is not required for
executing the generated Python code.
PAGE is not an end-all, be-all tool, but rather one that attempts to ease the burden
on the Python programmer. It is aimed at the user who will put up with a less than
completely general GUI capability in order to get an easily generated GUI. A helper
and learning tool, it does not build an entire application but rather is aimed at building
a single GUI class and the boilerplate code in Python necessary for getting the GUI
on the screen.
Back-end Development
The back-end development is totally done with Python. Some of the image pre-
processing techniques like Noise Removal, Gray Scaling and Otsu’s Thresholding
are used. One of the fundamental challenges in image processing and computer
vision is image denoising. What denoising does is to estimate the original image by
suppressing noise from the image. Image noise may be caused by different sources
(from sensors or from environment) which are often not possible to avoid in practical
situations. Therefore, image denoising plays an important role in a wide range of
applications such as image restoration, visual tracking, image registration, and image
segmentation. While many algorithms have been proposed for the purpose of image
denoising, the problem of image noise suppression remains an open challenge, espe-
cially in situations where the images are acquired under poor conditions where the
noise level is very high.
In global thresholding, we used an arbitrary value for threshold value. So, how
can we know if a value we selected is good or not? Answer is trial and error method.
But consider a bimodal image (In simple words, bimodal image is an image whose
histogram has two peaks). For that image, we can approximately take a value in
the middle of those peaks as a threshold value. That is what Otsu binarization does.
So in simple words, it automatically calculates a threshold value from an image
histogram for a bimodal image (For images which are not bimodal, binarization
won’t be accurate.). For this, our cv2.threshold () function is used, but passes an
188 V. D. Jadhav and L. V. Patil

extra flag, cv2.THRESH OTSU. For threshold value, simply pass zero. Then the
algorithm finds the optimal threshold value and returns you as the second output,
retVal. If Otsu thresholding is not used, retVal is the same as the threshold value you
used.

3 Results and Discussion

The above Fig. 2 is the unprocessed image that is selected by the user. The image
contains sources of error and noise and needs to be processed.
The above Fig. 3 is the processed image after noise removal. Salt-noise and other
types of noise are removed from the image. The image is also sharpened so as to
make it compatible with the data-set.
The above Fig. 4 is gray-scaled to hide the uncertain parameters and con-
concentrate on the required part of the image. This image comes under 2-D image
spectroscopy of array data.
The Fig. 5 shown above is the processed image by the algorithm we have used
known as Otsu’s Thresholding Technique. The algorithm enhances the image details
and makes the image more accurate and precise [13]. This algorithm does image
enhancement.

4 Conclusion

Our enhancement technique effectively implements methodology that works with

cellular level detection and determination of cancer. It predicts the possibility of
lung cancer as cancerous or non cancerous using the data-set facilitated by machine

Fig. 2 Original image taken using foldscope

Enhancement Technique for Early Stage Lung Cancer Detection … 189

Fig. 3 Image after noise removal

Fig. 4 Processed image

learning. Helps to create awareness among the masses for early screening of the
disease. Reduces need for higher-level machinery implementation in order to develop
low cost systems with high output.
190 V. D. Jadhav and L. V. Patil

Fig. 5 Image after applying Otsu’s Thresholding algo

References

1. Gonzalez Rafael C, Woods Richar E, Eddins Steven L Digital image processing using MATLAB
2. Cybulski JS, Clements J, Prakash M (2014) Foldscope: origami-based paper microscope. PLoS
ONE 9(6):e98781. https://doi.org/10.1371/journal.pone.0098781
3. Ferlay J, Soerjomataram I, Ervik M, Dikshit R, Eser S, Mathers C, Rebelo M, Parkin DM,
Forman D, Forman D (2015) [online]. Available: www.wcrf.org
4. Grove O, Berglund AE, Schabath MB, HJWL Aerts, Dekker A, Wang H, Gillies RJ (2015)
Data from: quantitative computed tomographic descriptors associate tumor shape complexity
and intratumor heterogeneity with prognosis in lung adenocarcinoma. The Cancer Imaging
Archive [online]. Available: https://doi.org/10.7937/K9/TCIA.2015.A6V7JIWX
5. Grove O, Berglund AE, Schabath MB, Aerts HJWL, Dekker A, Wang H, Gillies RJ, Muoz-
Barrutia A (2015) Quantitative computed tomographic descriptors associate tumor shape
complexity and intratumor heterogeneity with prognosis in lung adenocarcinoma. PLOS ONE.
Public Library of Science (PLoS)
6. Lee JH et al (2017) Anatomically and functionally distinct lung mesenchymal populations
marked by Lgr5 and Lgr6. Cell 170(11491163):e1112
7. Storvall H, Ramskold D, Sandberg R (2013) Efficient and comprehensive representation of
uniqueness for next-generation sequencing by minimum unique length analyses. PLoS ONE 8
8. Clark K, Vendt B, Smith K, Freymann J, Kirby J, Koppel P, Moore S, Phillips S, Maffitt D,
Pringle M, Tarbox L, Prior F (2013) The cancer imaging archive (TCIA): maintaining and
operating a public information repository. J Digital Imaging 26(6):1045–1057
9. Armulik A, Genove G, Betsholtz C (2011) Pericytes: developmental, physiological, and
pathological perspectives, problems, and promises. Dev Cell 21:193215
10. Simon H (1994) Neural networks. Mac Millen College Pub co, New York
11. Marques S et al (2016) Oligodendrocyte heterogeneity in the mouse juvenile and adult central
nervous system. Science 352:13261329
12. Zepp JA et al (2017) Distinct mesenchymal lineages and niches promote epthelial self-renewal
and myofibrogenesis in the lung. Cell 170(11341148):e1110
13. Vanlandewijck M et al (2011) A molecular atlas of cell types and zonation in the brain
vasculature. Nature 554, 475480
Foldscope to Detect the Growth
of Microorganisms on Various Materials
and Vessels
Vanita D. Jadhav, Richa Tamhane, Kiran Kedar, Shruti Kawade,
and Aboli Gaikwad

Abstract Dr. Manu Prakash Professor, Department of Biotechnology, Stanford

University invented foldscope known to be paper microscope. A low cost, portable
and foldable microscope. It is ultra-affordable and durable and quality is almost
similar to conventional microscopes. It has various applications in various fields.
However, its importance is still obscured by many. Foldscope can be used as an
efficient tool to study pollen viability and stomata, to visualize cells and detect seed
viability was demonstrated. Foldscopes are used in projects by students to detect,
visualize and learn real life examples. In this paper we will use Foldscope to detect the
growth of microorganisms on various materials and on vessels. It can be a preventive
measure and various diseases can hence be prevented.

Keywords Foldscope · Research tool · Biology · Image-processing ·

Microorganisms · Detection

1 Introduction

The easiness of Foldscope attracts the attention of many of us. The educationalists
to researchers and the common people to scientists. The thing that pulls them is this
modern origami based microscope with its wide applications. Foldscope, however,
has a big hand in solving the microbial problems. During the twenty-first century,
Manu and his team put forth a modern hand microscope—Foldscope and it helps
lots of young minds to solve the mysteries and puzzles of the microbial world [1].
This modern world microscope is sure to take up other dimensions of the microbial
world with its applications [2]. In this paper, we shall discuss the use of foldscope
to detect the growth of microorganisms on different construction materials and on
vessels [3].

V. D. Jadhav (B) · R. Tamhane · K. Kedar · S. Kawade · A. Gaikwad

CSE, SVERI’s College of Engineering, Pandharpur, India
e-mail: [email protected]
R. Tamhane
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 191
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_20
192 V. D. Jadhav et al.

Foldscope to detect the growth of microorganisms on construction material

Corrosion is a natural process that converts a refined metal into a more chemically
stable form such as oxide, hydroxide, or sulfide. It is the gradual destruction of mate-
rials by chemical and/or electrochemical reaction with their environment. Corrosion
is one of the leading economic losses in the constructive world [4]. At construction
sites, though there are many factors in retardation of metals, microbes are found to be
a major factor to corrode the metals. A scientific report says that almost 20% of the
reason for corrosion is due to microbes [5]. The microorganisms which have been
associated with corrosion involve many genera and species. They may be divided
into three groups: bacteria, fungi, and algae. The most important bacteria that play a
significant role in the corrosion process are those involved in the sulfur cycle [6].
The corrosion starts with the formation of biofilm on the surface of the metal. A
biofilm comprises any syntrophic consortium of microorganisms in which cells stick
to each other and often also to a surface [7]. A microenvironment is created above
the surface and it can be different from bulk with properties such as pH or dissolved
oxygen. Biofilm is a complex structure. It is a tridimensional structure consisting of
nutrient gradients [8].
When a multispecies of biofilm is caused, it creates the most destructive corrosion.
These biofilms are caused due to the interaction between different species and a
metabolic product formed leads to corrosion. Also, there are several methods to
detect microbes on metals [9].

1.1 Process to Detect Microbes

This study consists of the samples from Cement, Wood, Iron plates, Gravel Asbestos
sheets, brick, glass and tiles. The considered sample was diluted one by one using
sterile distilled water. The diluted sample is now treated with agar medium by
pour plate technique and is incubated for 24–48 h. Petri plates were considered
for enumeration and finally, foldscope was used to detect the bacteria on the sample.

1.2 Detection of Microbes

1432 microbes were found enumerated among 174 samples. From the considered
sample material, 6 bacterial genus and 3 fungal genus were found isolated. Foldscope
images were later compared with the microscopic images and it was found that it
was the best alternative, cheap and durable. Detection of microbes on construction
materials had always been a challenging task in the science out there. But foldscope
has now proved to be the best possible solution for such problems.
Foldscope to Detect the Growth of Microorganisms on Various … 193

Table 1 Food and the

Grains and grain products Number of mycotoxin types
bacteria that affects it
Milk and milk products Aflatoxin
Peanuts, nuts and pulses Aflatoxin
Fruits and vegetables Patulin

2 Foldscope to Detect the Growth of Microorganisms

on Cooking Vessels

Microbes can be found almost everywhere in nature including air, water and land.
They are ubiquitous in nature. In the other cause, microbes are also important in our
day to day life and culture in ways to treat sewage, produce fuels, enzymes and even to
ferment the food. They play a role as an essential biological compound. On the other
hand, microbes are those pathogens responsible for spreading diseases and polluting
the hygiene [2]. Many disease causing pathogenic microorganisms are present in
the environment and they are easily exposed to the food products by handling and
growing on the surface of vessels used for food preparation [10].
Foldscope can be used to detect different types of microorganisms and bacteria
but can’t help to detect the viruses. The percentage of bacteria is highest with respect
to other reasons to cause illness and is up to 66%. Table 1 shows some of the bacterias
affecting food.
Moulds like Aspergillus produce aflatoxin, ochratoxin, citrinin and patulin are
vulnerable to humans. However, these are even difficult to get rid of. To prevent the
growth of such microbes, cleansing and sanitizing of utensils, vessels and equipment
must be done from time to time to avoid consequences [4].
Detection of microbes
As the studies have clearly proved that foldscope has a major role in the application to
detect microbes in cooking vessels. Beyond utensils, growth of microbes is found in
kitchen sponges and dish-clothes too. If it is not cleaned timely, it may lead to cross-
contamination and thus contaminate the food too. This study consists of different
types of cooking vessels such as copper, brass, stainless steel, tin, plastic.

3 Procedure

3.1 Preparation of Starch

Starch of 1000 ml was prepared and sterilized. The sterilized starch solution was
poured onto the vessels 100 ml each and kept for incubation at room temperature 27
°C. This starch has been applied over all the cooking vessels.
194 V. D. Jadhav et al.

3.2 Preparation of Nutrient Agar

It was made of sterile distilled water with the composition of nutrient agar 14, 2.5 g in
500 ml water. A plate was kept over it and it was incubated for about 24 h. Hence, the
presence and absence of colonies above it was monitored. This showed the growth
of microbes on the plate.

3.3 Gram Staining of Bacteria

The slides were prepared by simple staining of selected colonies. The cell wall
composition made it clear about the bacteria whether it was Gram positive or gram
negative. Thus, gram staining was done.

3.4 Isolation and Characterization of Microbes

Staining was performed at various time intervals to differentiate between gram posi-
tive and gram negative. The slide was placed in the paper based microscope interfaced
with a mobile phone. After staining the cells were observed under foldscope, the
recorded images were further analyzed using a microscope. Morphological obser-
vations were done upon simple staining methods and microbes were seen using
foldscope. Foldscope images clearly have shown the presence of bacteria and fungi
from the collected samples (Refer Figs. 1 and 2).

4 Conclusions

We came to know that foldscope unlike other microscopes is durable, affordable and
easy to carry anywhere. It is a best choice to detect microbes and other micro particles
for educational or research purposes. It gives best possible results and proves to be
an ideal tool to detect and monitor them. Experiments clearly showed that foldscope
gives upto 140X magnifications. Low cost and high outputs, made this tool a best
choice among the society. It has various applications and its specific advantages in
each new application.
Foldscope to Detect the Growth of Microorganisms on Various … 195

Fig. 1 Microbial growth on nutrient agar plates

Fig. 2 Microbes observed using foldscope

196 V. D. Jadhav et al.

References

1. Jadhav VD (2018) Foldscope: a low cost magnification device and its applications in various
fields
2. Rathod D. Foldscope: a educational cum research tool used in identification of microorganisms
3. Cybulski JS. Foldscope: origami based microscope
4. Calder R. Preliminary studies in the use of foldscope paper microscope for diagnostics analysis
of crystals in urine
5. Muthukumaran E. An experimental investigation on the quality of air and growth of microor-
ganisms in various construction materials at various conditions and in various cooking
vessels
6. Ephraim RKD. Diagnosis of Schistosoma haematobium infection with a mobile phone-mounted
foldscope and a reversed-lens CellScope in Ghana
7. Sharmila S. Anatomical Characterization of medicinal plants using foldscope: a paper based
origami microscope
8. https://en.wikipedia.org/wiki/Foldscope
9. Raghavamma STV. Applications of paper microscope for the early detection of parasitic
pathogens in blood smear
10. Banerjee S. Foldscope, the frugal innovation and its application in food microscopy—a review
Implementation of A* Algorithm
for Real-Time Evacuation During Fire
Situation

Shilpa K. Rudrawar, Pallavi Ghorpade, and Dipti Y. Sakhare

Abstract Building architectures are growing towards increased complexity, with

countless people moving through them. Not all amongst the crowd could possibly
be familiar with the building to escape a fire danger zone. Even if the infrastructure
complies to safety standards, decision making for fire evacuation, while ensuring
safety, is utmost critical. Tailoring to these constraints, it is essential to protect lives
by efficient and complete evacuation. For fire emergency, the proposed evacuation
routing system is inputted by a group of wireless sensor nodes present across the
considered floor plan; a MATLAB based central server to find/calculate better safe
evacuation routes for the imperiled people, at a remote location in the building; a Wi-
Fi based network that communicates this calculated route from the sensor network
to server and server to the occupant, on evacuee’s cell phone. The information from
the sensors is transmitted by a Wi-Fi network and is aggregated by the Thingspeak
server. The real-time evacuation route is calculated by the server, towards the nearest
and safest exit door from the occupant’s instantaneous location, by deploying A*
algorithm for route optimization, along with data from sensor network that informs
about origin and fire spread regarding hazard’s location. The server transmits the route
information to the occupants through Wi-Fi connectivity. The endangered evacuees
aretherebyenabled to view and follow this information of dynamic and real time active
maps using a Smartphone. The proposed framework is prototyped and analyzed for
their future inclusion into existing fire evacuation systems.

Keywords Fire emergency · Real-time evacuation · Routing · A* algorithm ·

Wi-Fi · MATLAB

S. K. Rudrawar (B) · P. Ghorpade · D. Y. Sakhare

Department of Electronics and Telecommunication Engineering, MIT Academy of Engineering,
Alandi (D), Pune, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 197
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_21
198 S. K. Rudrawar et al.

1 Introduction

Living in skyscrapers has caused significant issues and difficult and unpleasant chal-
lenges in occupant evacuation during fires or any other emergencies. Predefined
evacuation maps are static in nature, because they showcase some main evacuation
routes which could be prone to blockage as a fire hazard evolves [1]. Evacuation
management regards the exit choice to be critical; some prominent evacuation strate-
gies generally consider egress by shortest distance, or balanced use of exits/stairs for
proper crowd dispersion, or prioritization of Fire floor or Fire layer evacuation [2].
To alleviate casualties, it is important to have emergency evacuation plans in place
before fire outbreaks occur. However, there considerably too many factors which
involve uncertainties and complexities in evacuations. Previous studies unfold that
amongst others, prominently, building structures, fire scenarios and evacuees share
a common physical and social interaction during evacuation activities for fire inci-
dents. Some factors why these components indulge in the evacuation are listed in
Table 1.
Smoke inhalation kills many, compared to the fire. The prominent dangers are
smoke and high temperatures, too that accompany fires, potentially of causing large
scale casualties. Therefore, critical signaling to the evacuees needs ensuring that
evacuation routes are safe and free from smoke to the maximum possible extent [3].
Hence the main aim of this paper is to inculcate such a mechanism which ultimately
is a real-time fire-safe pathfinder that helps improving the efficiency of evacuation.
Many projects in the past limit the real-time delivery of fire incident information
to fire services providers and first responders only.

Table 1 Components that interact during fire evacuation incidents

Interactional components for evacuation Reasons why the basic components obstruct the
systems safe evacuation during fire situation
Building structures 1. For high rise buildings, inoperable lifts or
escalators
2. Behavior of construction materials and building
structures
3. Inadequate/limited way outs or exits for escape
Fire scenarios 1. Insufficient evacuation practice through
evacuation drills
2. Impact of fire scenarios different in nature even
for same evacuation strategy can be different
3. Different fire materials can possibly produce
various smoke layers and thermal radiation
Evacuees 1. Unguided, self-directed evacuation might jam
the exits
2. Movement only towards known emergency
exits
Implementation of A* Algorithm for Real-Time Evacuation … 199

But the efficiency of evacuation can be improved if evacuees are enabled to take
the correct decision on time through digital signage/mobile devices. It is aimed to
be a guidance system for fire evacuation that will employ smart phones to provide
appropriate as well as timely information. This approach will be more reliable than
just some static data on walls shown by guide boards or say any announcements.

2 Literature Survey and Related Work

2.1 Market Survey

There is advanced automatic fire protection equipment available now such as

dampers, control doors, fans, and fire suppression pumps meant for fire extinction
in buildings. The market also has multi-sensor detectors responsive to ranges of
fires and with high stability to skip occurrences of unwanted alarming [4]. There
are also IOT enabled detectors offered by suppliers like Nest Labs, Kidde Systems,
Roost, First Alert, Leeo, etc. [5]. But many of these fire detection, protection and
extinguisher systems fall short of evacuation intelligence.

2.2 Literature Survey

‘Egress models’ mean computer simulation models. They basically anticipate

required time for building inhabitants to evacuate. Egress models could be similar to
zone models, and thereby determine time to the outset of illogical conditions arising
at a building [6]. Egress models are reliably utilized in design analyses (performance-
based) for code compliance towards alternative design. Egress models are also
put to use for figuring out where during egress, probably, the zones of blockages
would develop. PATHFINDER, EESCAPE, EGRESS, ELVAC, building EXODUS,
EVACNET4, EVACS, Simulex, EXIT89, EXITT, are some examples developed for in
the past. Many studies have attempted to analyze fire evacuation modeling by intense
study of Building Information Modeling (BIM), Agent-based modeling, Behavior
based modeling of crowd and Network-optimization based algorithms [7].
Amongst different others, A* is efficient routing algorithm [8]. The A* path-
calculation algorithm is just extended Dijkstra algorithm, so it similarly uses distance
as a shortest path calculation metric. It simplexes the area of search by its tile-
segmentation, calculating scores for then possible paths from a starting point to a
destination, circumventing each obstacle during this [9]. The shortest path is then
constructed, backtracking from the final point via intermediate tiles to starting point.
A* algorithm has been majorly used in computer games navigation, robot navigation,
and in that of Google’s self-driving car too, because of its efficiency in time [10].
Comparison of Prominent Path Search algorithms is shown in Table 2.
200 S. K. Rudrawar et al.

Table 2 Comparison of prominent path search algorithms

Criteria Genetic algorithm Artificial neural A* algorithm
network
Time complexity Greatest time More time needed than Efficient in chaotic
A* algorithm, but is paths too, ensuring no
best if input size is collisions
more
Optimality Overall path travelled Travels longer distance Ensures shortest path
is less than ANN always
Space complexity Memory requirements Better than Genetic Average memory
are very high Algorithm requirements
Reasons for failure Poorest performance if Might Fail for highly Might take too long
the path is like a Maize chaotic environment time to calculate moves
applied for too large input is
and exceeds the limits
of memory

A revised system is needed that can reorganize the existing safety infrastructure
of large buildings to become effective in saving people, reliably, by fast response.
The proposed system is not meant to replace existing fire-safety systems, but
rather it can incorporate evacuation intelligence into them. And has better future
scope due to its efficiency.

3 Methodology

3.1 Sensor Nodes

Fire sensing nodes have Node MCU to command and are prototyped as a module
consisting of 3 main sensors: a fire-detecting flame sensor, a smoke and gas MQ-2
sensor and aDHT-11 temperature sensor.
Congestion monitoring nodes also have Node MCU to command are additionally
consisting of 4th main sensors: IR (infrared) sensor for congestion detection, along
with others: a fire-detecting flame sensor, a smoke and gas MQ-2 sensor, and a
DHT-11 temperature sensor. Hence respective sensors are used at each node for data
acquisition by the Node MCU [11].

3.2 Central Hub and Path Planning

An algorithm was to be developed that would majorly take 2 primary inputs: Building
Map and the Fire situation. The system needed to be triggered by fire situation from
Implementation of A* Algorithm for Real-Time Evacuation … 201

the network meshed; and should reliably identify location of fire based sensor data.
Then using strategies in path planning, we aimed to develop evacuation maps starting
at various points as needed from inside the building. Hence sensor inputs have been
routed to MATLAB via Thingspeak server and A* is used for sensor data optimization
in evacuation path planning [12].
A* algorithm calculates the cost function f(n), as in Eq. 1 where h(n) is the distance
between the start position and node n and g(n) is the distance between n and target.

f(n) = h(n) + g(n) (1)

3.3 User Interface

So Thunkable App platform is used to create an Android Application—a fast, yet

informative app for finding most appropriate evacuation-route.

3.4 Prototype Implementation and Testing

We applied this approach to a close indoor floor map; which is of a single-floor

building and we observed effect of this infrastructure that provides evacuation routing
information to occupants. A sensor node in network covers a considered area of floor
plan. Detecting if that area is blocked/transitable. The sensor acquired data is carried
by a WiFi, collected at the central server. Here calculation of evacuation route, in real
time, to the closest exit to the occupant’s location begins. When done computing,
server sends this assessed info to the occupant through the communication network.
The user may follow this suggested path when he queries help for himself by the
Mobile App. Testing was verified by whether a flame brought close to the IR Flame
Sensor or not, sends a digital 1 or 0 to the Node MCU. Power-save/sleep mode
is exited by Node MCU if flame comes near IR sensor. Then interpreted data is
packaged and sent to the ESP8266 Wi-Fi transceiver by the Node MCU.

3.5 Algorithm Implementation in MATLAB

The experimental stages of algorithm Implementation in MATLAB are shown in

the following figures. Targets, starting point and obstacles were assumed in different
cases provide an acceptable output for the shortest path. The basic A star algorithm
implemented was capable of finding the shortest route, irrespective of the starting
point and the target.
202 S. K. Rudrawar et al.

Fig. 1 Prototypes sensor nodes

To have the input of Sensor Node onto the on-line server Thingspeak Platform, a
bread-board implementation of the sensors is connected to a Wi-Fi. This is depicted
by Fig. 1.

3.6 Input of Sensor Node onto the On-Line Server

Thingspeak Platform

Once Wi-Fi connection is sensed by the nodes, the monitored data is transmitted to
the Thingspeak Server. Hence, these sensor nodes continuously send the data online
to Thingspeak. This is shown in Fig. 2.
This data is further imported in MATLAB and incorporated in to the A* algorithm.
For instance, data captured by Node1, Node2 and Node3 is imported in MATLAB
Script by the following lines of code from the respective thingspeak channel.

node1 = thingSpeakRead(723592, ‘Fields’, 2)

node2 = thingSpeakRead(723592, ‘Fields’, 5)
node3 = thingSpeakRead(723592, ‘Fields’, 8)
Implementation of A* Algorithm for Real-Time Evacuation … 203

Fig. 2 Data sent to a channel on Thinkspeak platform

3.7 Android Application Development

To have the output of calculated evacuation route as a recommendation on Android

application Platform, the basic buttons & app connectivity is accomplished as shown
in Fig. 3. The various instances of evacuation map or images, provided timely at
hand for the end user are displayed in Fig. 4.

4 Results and Discussion

Presence of fire (flame) brought nearer to the Flame Sensor, sends a digital 1 or
0 to Node MCU. It in turn analyses this data by appropriate thresholding, for the

Fig. 3 Android
application—front page
204 S. K. Rudrawar et al.

Fig. 4 Android application

changed reading, packages this data and sends it to Thingspeak channel. Node MCU
by its ESP8266 Wi-Fi transceiver sends this data wirelessly out to the Wi-Fi network
aligned field-wise to the Thingspeak Channels with its information. Hence fire is
imitated by candle flame as sensed by fire sensor. For the prolonged presence of fire,
it is simulated to aggravate, and this spreading of fire is assumed as the dispersion
Wi-Fi specifications over which the nodes were tested is described below:
• Security: WPA2 PSK
• Frequency: 2.4 Ghz band
• Identification: Pixel Graphics
• SSID: uitech333.

Test Case 1:
Path calculated in case of absence of fire danger: In this instance, there was no fire
present and hence no fire detected by the sensor nodes. So the evacuee was provided
with the shortest path to escape from the room which he queried over the app from.
MATLAB calculation of optimized route is shown in Fig. 5a and b.
Test Case 2:
Decision of Correct Choice of Available Exits: In this instance, fire was present
at one door of a room. Even if that path was shortest, but it was overridden by a
comparatively safer path. So the evacuee was provided with the shortest path to
escape from the room which he queried over the app from. MATLAB calculation of
optimized route is shown in Fig. 6a and b.
In this instance, fire was now detected at another door of the same room. The
second door had no fire. Hence the shortest and safest path was suggested by opti-
mization. So the evacuee was provided with the shortest path to escape from the
Implementation of A* Algorithm for Real-Time Evacuation … 205

Fig. 5 a Path calculated in case of absence of fire danger. b Real time Output on Mobile App

Fig. 6 a Decision of correct choice of available exits. b Real time output on mobile app

room which he queried over the app from. MATLAB calculation of optimized route
is shown in Fig. 7a and b

4.1 Execution Time

Execution time of modified A* Algorithm implemented in MATLAB for one query

from user or evacuee is pictured in Fig. 8.
206 S. K. Rudrawar et al.

Fig. 7 a Decision of correct choice of available exits. b Real time output on mobile app

Fig. 8 Profile of M-script mapped from MATLAB

5 Conclusion

Proposed work gives a framework for calculating better evacuation routes for occu-
pants of an indoor scenario threatened by a fire caused, and a case of single occu-
pant considered. The A* algorithm via MATLAB uses in-formation collected from
the sensor network. Information from the sensors indicates door area blockage. A
Implementation of A* Algorithm for Real-Time Evacuation … 207

network provides indication on the direction to exit to the occupant for a quick evac-
uation. We tested a version of routing using the information about transit able spaces
and fire location and another that adds a distance as safety margin between the occu-
pant path and fire to keep the occupant far from the fire. The prototype results were
satisfactory, as aimed through the design framework.

6 Future Research Plan

Each sensor node can be enabled to be decisive. The end user interface can be
developed further as active Exit Signs/digital Sign Boards spread across facility that
consider themselves as starting point, and detect dangerous areas around them in real
time. Such a UI can direct evacuees to the safest evacuation path from that room to the
nearest exist. Eventually data of both, the spread of the hazard and of the safe-egress
accomplished, can be shared to the Hub computer for calibrating the management
strategy decisions taken up each time by the nodes, thereby avoiding congestion.

References

1. ISO and Construction (2017) ISBN 978-92-67-10779-0

2. Gershon RRM, Magda LA, Riley HEM, Sherman MF (2012) The world trade center evacuation
study: factors associated with initiation and length of time for evacuation. Fire Mater 36:481–
500
3. Roan T-R, Haklay M, Ellul C (2011) Modified navigation algorithms in agent-based modelling
for fire evacuation simulation. In: The 11th international conference on geo computation,
London, pp 43–49
4. Fire alarm systems market driven by rapid growth in developing nations. ARC Advi-
sory Group. https://www.arcweb.com/press/fire-alarm-systems-market-driven-rapid-growth-
developing-nations
5. Wang B, Li H, Rezgui Y, Bradley A, Ong HN (2014) BIM based virtual environment for fire
emergency evacuation. Sci World J 2014:1–22
6. Richardson O, Jalba A, Muntean A (2018) Effects of environment knowledge in evacuation
scenarios involving fire and smoke: a multiscale modelling and simulation approach. Fire
Technol 55(2):415–436
7. Barnes M, Leather H, Arvind DK (2007), Emergency evacuation using wireless sensor
networks. In: The 32nd IEEE conference on local computer networks (LCN 2007). IEEE,
Dublin, Ireland, pp 851–857
8. Tabirca T, Brown KN, Sreenan CJ (2009) A dynamic model for fire emergency evacuation
based on wireless sensor networks. In: 2009 eighth international symposium on parallel and
distributed computing. IEEE, Lisbon, Portugal, pp 29–36
9. Ma J, Jia W, Zhang J (2017) Research of building evacuation path to guide based on BIM.
In: The 29th Chinese controls & decision conference (CCDC). IEEE Press, Chongqing, pp
1814–1818
10. Bakar NAA, Adam K, Majid MA, Allegra M (2017) A simulation model for crowd evacuation of
fire emergency scenario. In: (ICIT) The 8th international conference on information technology,
IEEE, Amman, Jordan, pp 361–368
208 S. K. Rudrawar et al.

11. Ronchi E, Uriz FN, Criel X, Reilly P (2016) Modeling large-scale evacuation of music festivals.
Case Stud Fire Saf 5:11–19 (Elsevier)
12. Radianti J, Granmo O-C, Sarshar P, Goodwin M, Dugdale J, Gonzalez JJ (2015) A spatiotem-
poral probabilistic model of hazard and crowd dynamics for evacuation planning in disasters.
Appl Intell 42:3–23
Involuntary Traffic Control System

Shriniwas V. Darshane, Ranjeet B. Kagade, and Somnath B.Thigale

Abstract In an automated manner, In an automated manner, holistic traffic manage-

ment is essential to enhance management in metro cities and even in two-tier cities.
Detection of Vehicle flow is deemed to be crucial in the management of Traffic. In
fact, flow of the Traffic shows the state of the Traffic in a definite amount and helps to
rectify situations leading to traffic jam. Particularly this project intends to elucidate
a traffic TV for non-chaotic vehicular traffic management. The fundamental idea
includes five steps: subtraction of background, detection of the blob, blob analysis,
pursuit of blob and reckoning of vehicle. Ideally, a vehicle is considered as asso-
ciate rectangular patch and classified via blob analysis. After analyzing the blob of
vehicles, the pertinent choices unit of mensuration extracted. The pursuit of moving
targets is achieved by examination the extracted choices and activity. The exper-
imental results show that the projected system can give a huge amount of useful
information for traffic investigation.

Keywords Image processing · Distributed system · Background subtraction ·

Intelligent systems

1 Introduction

Development of an intelligent remote control system for street light and traffic signal
control system needed because of present traffic light controllers based on old micro-
controllers such as AT89C51 which has very less internal memory and no in-built
ADC. These systems have limitation because it will use the predefined program that
does not have the flexibility of modification on real-time application. The present
traffic system has a fixed time interval for a green and red signal, which does not
provide the flexibility to the system and street lighting system public sector are design

S. V. Darshane · R. B. Kagade (B) · S. B.Thigale

SVERI’s College of Engineering, Pandharpur, Maharashtra, India
e-mail: [email protected]
S. V. Darshane
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 209
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_22
210 S. V. Darshane et al.

according to the old standards. The intelligent remote control system for street light
and traffic signal control system consists of high-performance, low cost, low power.
The system will deal with two basic problems: (i) Detection of traffic volume by
using genetic algorithm (ii) automatic control of street light using the sensor. The
traffic signal management is a very important facet in electronic equipment town
traffic system as we have a tendency to all glorious, traffic systems area unit time-
varying, random system. Thus lots of standard strategies for traffic signal manage-
ment primarily based precise models fail to deal efficiently with the advanced and
ranging traffic things. During all amongst in every of the most options of contem-
porary cities is that the permanent growth of population in a comparatively little
space. The consequence of this truth is that the increase within the range of cars and
additionally the requirement of movement and transport of individuals and product
in urban town networks. Traffic congestion in main road networks is one amongst the
most problems to be addressed by today’s traffic management schemes. Automation
combined with the increasing penetration of on-line communication, navigation,
and advanced driver help systems can ultimately lead to intelligent vehicle main
road systems (IVHS) that distribute intelligence between margin infrastructure and
vehicles which above all on the long term are one amongst the foremost promising
solutions to the traffic congestion drawback. During this paper, we have a tendency
to gift a survey on traffic management and management frameworks for IVHS. First,
we have a tendency to provide a short summary of the most presently used traffic
management strategies for freeways. Next, we have a tendency to discuss IVHS
based traffic management measures. Then, varied traffic management architectures
for IVHS like PATH, Dolphin, Auto21 CDS, etc. are mentioned, and a compar-
ison of the varied frameworks is conferred. Finally, we have a tendency to sketch;
however, existing traffic management methodologies may fit in Associate in Nursing
IVHS-based traffic management set-up.
Fast transportation system and rapid transit system are nerves of economic devel-
opment for the nation. All developed countries have a well-developed transportation
system with efficient traffic control on the road in, rail, and air transportation of good,
industrial products, human resources, and machinery are the key factors which influ-
ence the industrial development. Mismanagement and traffic congestion result in
long waiting time loss of fuel and money. It is therefore utmost necessary to have a
fast, economical and efficienttraffic control system for nation development.

2 Literature Review

In current system traffic get collected at particular places. Now days there are lack
of traffic analysis which results into heavy Traffic. Many time emergency services
are trapped. We refer various papers.
The author Mu et al. [1] describes A camera-based rule for period of time durable
stoplight detection and recognition was planned. This rule is supposed chiefly for
autonomous vehicles. Experiments show that our rule performs well in accurately
Involuntary Traffic Control System 211

investigating targets and in determinative the gap and time to those targets. However,
the current methodology planned here can have some drawbacks. First, the maneuver
performs well within the daytime but not additionally within the dead of night. The
warning rate can increase within the dead of night as results of lots of light-weight
interference. Whereas the maneuver can discover every circulartraffic light and other
people with arrows, exclusively the classical suspended, vertical traffic lights were
detected. Detection and recognition of lots of types of traffic lights will meet an
important house for future work.
Shu-Chung [2] proposed driver assistant system design supported image process
techniques. A camera is mounted on the vehicle front window to sight the road
lane markings and confirm the vehicle’s position with regard to the lane lines. A
changed approach is projected to accelerate the HT method during a computationally
economical manner, thereby creating it appropriate for time period lane detection.
The no heritable image sequences are analyzed and processed by the projected system
that mechanically detects the lane lines. The experimental results show that the system
works with success for lane line detection and lane departure prediction.
Alcantarilla [3] presents degree automatic road traffic management and looking
system for daytime sequences using a B&W camera. Necessary road traffic data
like mean speed, dimension and vehicles numeration are obtained practice laptop
computer vision methods. Firstly, moving objects are extracted from the scene by
suggests that of a frame-differencing algorithm and texture data supported grey scale
intensity. However, shadows of moving objects belong to boot to the foreground.
Shadows are far away from the foreground objects practice silk hat transformations
and morphological operators. Finally, objects are tracked in AN extremely Kalman
filtering technique, and parameters like position, dimensions, distance and speed of
moving objects are measured. Then, per these parameters moving objects are classi-
fied as vehicles (trucks or cars) or nuisance artifacts. For results mental representation,
a 3D model is projected onto vehicles among the image plane. Some experimental
results practice real outside sequences of images ar shown. These results demon-
strate the accuracy of the planned system to a lower place daytime interurban traffic
conditions.
Hussian [4] proposed system involves use of Wireless sensing element network
technology to sense presence of Traffic close to any circle or junction and so able to
route the Traffic supported Traffic handiness or we will say density in want direction.
this method doesn’t need any system in vehicles therefore are often enforced in
any Traffic system quite simply with less time and fewer pricey additionally. this
method uses Wireless sensing element networks Technology to sense vehicles and
a microcontroller based mostly routing formula programmed for wonderful Traffic
management [12].
Sinhmar [5] proposed system records vehicle count in its memory at user prede-
fined recording interval on real time basis. This recorded vehicle count information
is utilized in future to research traffic condition at various traffic lights connected to
the system. For acceptable analysis, the recorded information is downloaded to pc
through communication between microcontroller and also the computer. Adminis-
trator sitting on laptop will command system (microcontroller) to transfer recorded
212 S. V. Darshane et al.

information, update lightweight delays, erase memory, etc. therefore administrator

on a central station laptop will access traffic conditions on Associate in Nursing
approachable traffic lights and close roads to scale back traffic congestions to
an extent. In future this technique is accustomed inform individuals regarding
completely different places traffic condition [6].
In this system the traffic lights will be controlled mechanically. it’s not needed to
expressly set a time or amendment the Traffic lightweight manually.. The planned
algorithmic rule consists of five steps: background subtraction, blob detection,
blob analysis, blob trailing and vehicle investigation. A vehicle is modeled as an
oblong patch and classified via blob analysis. By analyzing the blob of vehicles,
the purposeful options square measure extracted. trailing moving targets is achieved
by comparison the extracted options and mensuration the minimal distance between
consecutive frame. The experimental results show that the planned system will give
real-time and helpful info for traffic police work [7–11].

3 System Architecture

Vehicle flow detection seems to be a very important half in closed-circuit television.

The traffic flow shows the traffic state in fixed amount and helps to manage and
management particularly once there’s a traffic jam. During this project, we have a
tendency to propose a traffic closed-circuit televisionfor vehicle.investigation. The
planned algorithmic rule consists of five steps: background subtraction, blob detec-
tion, blob analysis, blob following and vehicle investigation. A vehicle is sculptured
as an oblong patch and classified via blob analysis. By analyzing the blob of vehi-
cles, the purposeful options square measure extracted. Following moving targets is
achieved by scrutiny the extracted options and activity the lowest distance between
consecutive frame.shown in Fig. 1.

4 Proposed Algorithm

A. Set timer for switching camera

B. Start signal rotation
C. Capture image data
D. Process captured data
E. Match for traffic rules breaking
F. Send image to admin
G. Stop.

The above algorithm shown in Fig. 2.

Involuntary Traffic Control System 213

Fig. 1 System architecture

Fig. 2 Flow diagram of

proposed algorithm
214 S. V. Darshane et al.

5 Conclusion

Thus in this we have made an analysis on Automatic Traffic Control Signal in which
our motive is to reduce the Traffic on road in peak hours and make a way for emer-
gency situations.The method presented in this paper is simple and there is no need
to use sensors that have been commonly used to detect Traffic in the past. However,
one of the most important disadvantages of this method is extreme sensitivity to
light. For example, when installed in the road, changes in sun light potentially cause
interference with the camera. This problem can be overcome by using specific filters
during Image Processing or changes in Matlab code. With some improvements, this
method can be used to detect road accidents and identify violations of the spiral
movements of cars.

References

1. Mu G, Xinyu Z, Deyi L, Tianlei Z, Lifeng A (2015) Traffic light detection and recognition for
autonomous vehicles 22(1):50–56
2. Shu-Chung Y-C, Chang C-H (2015) A lane detection approach based on intelligent vision.
Comput Electr Eng 42:23–29
3. Alcantarilla PF, Sotelo MA, Bergasa LM (2008) Automatic daytime road traffic control and
monitoring system
4. Hussian R (2013) WSN applications: automated ıntelligent traffic control system using sensors.
J IJSC E 3(3)
5. Sinhmar P (2012) Intelligent traffic light and density control using Ir sensors and microcon-
troller 2(2)
6. Teodorescu H et al (1999) Fuzzy and neuro-fuzzy systems in medicine. CRC Press, Boca
Raton, FL, USA
7. Danescu R, Nedevschi S, Meinecke M, Graf T (2007) Stereo vision based vehicle tracking
in urban traffic environments. In: IEEE conference on ıntelligent transportation system, pp
400–404
8. Ferrier NJ, Rowe SM, Blake A (1994) Real-time traffic monitoring. In: Proceeding of the
second IEEE workshop on applications of computer vision, pp 81–88
9. Vu A, Barth M (2019) Catadiopetric omnidirectional vision system integration for vehicle-
based sensing. In: Proceeding of IEEE ıntelligent transportation system conference
10. Cao M, Vu A, Barth M. A novel omni-directional vision sensing technique for traffic
surveillance. In: Proceeding of IEEE
11. Jin B, Zou D, Gan Y (2010) Research and design of traffic detection based on GPRS. In: IEEE
conference on advanced computer control (ICACC)
12. https://www.researchgate.net/publication/337086901_Energy_Efficient_Routing_Prot
ocol_for_IOT_Based_Application
Early Detection of Diabetic Retinopathy
Using Machine Learning

Vishal V. Bandgar, Shardul Bewoor, Gopika A. Fattepurkar,

and Prasad B. Chaudhary

Abstract Early detection of Diabetic Retinopathy shields patients from losing their
vision because Diabetic Retinopathy may be a typical eye disorder in diabetic
patients. The elemental explanation for a visual deficiency within the populace. Thus,
this paper proposes an automated method for image-based classification of diabetic
retinopathy. The technique is separated into three phases: image processing, feature
extraction, and image classification. The target is to naturally group the evaluation
of non-proliferative diabetic retinopathy at any retinal image. For that, an underlying
image preparing stage separates blood vessels, microaneurysms, and hard exudates,
so on extricate highlights utilized by a calculation to make sense of the retinopathy
grade.

Keywords Machine learning · Image processing · Diabetic retinopathy

1 Introduction

Diabetic Retinopathy (DR) may be a standout among the foremost successive reasons
for visual debilitation in created nations. The primary source of the latest instances of
visual deficiency within the working-age populace. By and huge, almost 75 individ-
uals go dazzled every day as an outcome of DR. A viable treatment for DR requires
early finding and consistent checking of diabetic patients, however, this can be a
testing undertaking because the malady indicates few manifestations until it’s past
the purpose where it’s possible to allow treatment.
As shown in above Fig. 1, Diabetic retinopathy is an eye issue that can cause
Visual deficiency—little veins within the back of the attention called retinal veins

V. V. Bandgar · S. Bewoor (B) · G. A. Fattepurkar

S.K.N.C.O.E. Vadgaon, Pune, India
e-mail: [email protected]
V. V. Bandgar
e-mail: [email protected]
P. B. Chaudhary
V. I. T. Pune, Pune, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 215
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_23
216 V. V. Bandgar et al.

Fig. 1 A typical colour

fundus image with main
fundus structures and early
lesions marked

[1]. Indications of Diabetic Retinopathy are gliding spot in vision, obscured vision
and blocked vision. At the purpose when the sugar level in the blood builds, blood
vessels within the back of the attention finally end up frail, and seeable of this vessel
releases the blood and lipoproteins liquid, at the moment liquid winds up skimming
spot in vision with the goal that Diabetic patient cannot see anything totally through
the vision, within the event that we do not do the treatment of this ailment on the
time then it’d be conceivable of complete vision misfortune or visual deficiency [2].
On the off chance that we distinguished early the indication of Diabetic Retinopathy,
it’s conceivable to stay extra loss of vision.

1.1 Why is Machine Learning for DR?

As a value effective thanks to handling the healthcare resources, systematic screening

for DR has been identified. A vital screening tool for early DR detection is that the
emergence of automatic retinal image analysis. This could save both cost and time
because it reduces the manual workload of grading still as diagnostic cost and time.
Take, for example, the Netherlands is stated to own approximately 500,000 persons
affected with diabetes, and this number is expected to extend to over 700,000 by 2030.
The patients will need to undergo retinal examinations [3]. This will consequently
result in a huge amount of images that would need to be reviewed. As a result, ophthal-
mologists will have an infinite burden and also cause an increase within the roster
comprising quality of healthcare. Automated, highly accurate screening methods
have the potential to help doctors in evaluating more patients and quickly routing
those that need help to a specialist Machine learning may be a family of computational
methods that enables an algorithm to program itself by learning from an oversized set
of examples that demonstrate the specified behaviour, removing the requirements to
specify rules explicitly [4]. Using established dataset and using multiple classifying
algorithms to detect whether diabetic retinopathy is present or not, would bring relief
Early Detection of Diabetic Retinopathy Using Machine Learning 217

to both patients further because of the resources of the healthcare system. The accu-
racy, sensitivity and specificity of the algorithm for detecting diabetic retinopathy
(DR) can help ophthalmologists, and physicians raise the red flag and thus provide
early treatment to patients and convey in exceedingly more preventive care which
might bring down the burden on a healthcare Machine learning can thus help the
old adage-prevention better than cure, by predicting who is more vulnerable to be in
danger of DR or not [5].
Four stages of Diabetic Retinopathy are as follows
The first stage is known as Mild Non-Proliferative Diabetic Retinopathy (Mild
NPDR). In this stage, there will expand like swelling in the veins in the retina and
little inflatable like swelling in the veins known as Microaneurysms. The second stage
is known as Moderate Nonproliferative Diabetic Retinopathy (Moderate NPDR). In
this stage, a portion of the veins in the retina will end up blocked. The third stage is
known as Severe Non-Proliferative Diabetic. This method detects and classifies the
diabetic retinopathy. Preliminary results show that k-nearest neighbours obtained the
best result with 68.7% for the dataset with different resolutions.
Advantages
Perform automated classification of Diabetic Retinopathy and component analysis
in less time.
Disadvantages
The paper does not include testing the methods with larger data sets and classifying
the subtypes of the retinopathy.

2 History and Background

1. Akara Sophark, Bunyarit Uyyanonvara and Sarah

Retinopathy Diagnosis using Image Mining:

The author has mainly specialized in the detection of Glaucoma and Diabetic
Retinopathy. Glaucoma may be detected by the cup to disc ratio (CDR). Diabetic
retinopathy may be detected by Exudates, Hemorrhages, Microneurysms and plant
fibre Spots. RGB images are converted into YCbCr. Y plane is employed for detec-
tion of blood vessels, point and exudates. After candy edge detection, the image will
be converted into binary to perform Skeletonization operation. DCT is employed
for feature extraction. The author has proposed DCT (Discrete Cosine Transform)
for feature extraction. Employed for detection of blood vessels, point and exudates.
After candy edge detection, the image will be converted into binary to perform
Skeletonization operation. DCT is employed for feature extraction. The author has
proposed DCT (Discrete Cosine Transform) for feature extraction.
218 V. V. Bandgar et al.

The author has mainly focused on the detection of Glaucoma and Diabetic
Retinopathy. Glaucoma can be detected by the cup to disc ratio (CDR). Diabetic
retinopathy can be detected by Exudates, Hemorrhages, Microneurysms and Cotton
Wool Spots. RGB images are converted into YCbCr. Y plane is used for detection of
blood vessels, optic disc and exudates. After candy edge detection, the image will be
converted into binary to perform Skeletonization operation. DCT is used for feature
extraction. The author has proposed DCT (Discrete Cosine Transform) for feature
extraction [6].
Baraman, “Automatic Exudate Detection from Non-dilated Extracted feature goes
to SVM classifier. After that Extracted feature goes to SVM classifier, after that
Diabetic Retinopathy-Retinal Images using Fuzzy C-means Clustering” [7].

3 System Architecture

There are two categories of diabetic retinopathy: non-proliferative diabetic

retinopathy (NPDR) and proliferative diabetic retinopathy, where the NPDR can be
subdivided into mild, moderate and severe [8]. In fact, NPDR is the most commonly
diabetic retinopathy, representing 80% of all cases. The retinopathy grade diagnoses
are normally provided by medical experts based on 0. Normal (µA = 0) and (H =
0).
1. Mild NPDR (0 < µA ≤ 5) and (H = 0)
2. Moderate NPDR (5 < µA < 15 or 0 < H < 5) and (NV = 0)
3. Severe NPDR (µ ≥ 15) or (H ≥ 5) or (NV = 1).
where µA is the number of microaneurysms, H the quantity of haemorrhages and
NV the presence of neovascularization.
The above Fig. 2 shows the overall structure of the proposed system. It includes
all the modules which we are going to implement in this project. The architecture
gives an idea of input to the system, processing on that input and what will be the
output of the project.

Fig.2 Purposed systems for detection and classification of different stages of diabetic retinopathy
Early Detection of Diabetic Retinopathy Using Machine Learning 219

3.1 Image Database

The Messidor database [9] includes 1200 eye fundus shading numerical photos of the
back shaft obtained by 3 ophthalmologic workplaces using a shading video 3CCD
camera on a Topcon TRC NW6 non-mydriatic retinograph with a 45° field of view.
The pictures were caught implementing 8 bits for each shading plane at 1440 × 960,
2240 × 1488 or 2304 × 1536 pixels.
800 images were acquired with pupil dilation (one drop of Tropicamide at
0.5%) and 400 without dilation. The 1200 pictures are bundled in 3 sets, one for
every ophthalmologic division, utilizing the TIFF position. What’s more, an Exceed
expectations document with therapeutic judgments for each picture is given.
In this work, we utilize the pictures of only one ophthalmologic division containing
152 pictures without retinopathy (grade 0), 30 with mellow NPDR (grade 1), 69 with
moderate NPDR (grade 2), and 149 with serious NPDR (grade 3).

3.2 Flow Chart

The above Fig. 3 shows the flow of the project. It is divided into two parts Training
Phase and Testing Phase. It shows the steps performed during the implementation.
The image is preprocessed first, then from that feature extraction is done, and on the
basis of that, the image is classified. The same steps are performed for the test image.

3.3 Activity Diagram

The above Fig. 4 shows the steps and methods of the project. The activities performed
during implementation are Preprocessing, feature extraction and classification [10].

4 Results and Analysis

4.1 Extracted Features Blood Vessels

See Fig. 5.

4.2 Microaneurysms

See Fig. 6.
220 V. V. Bandgar et al.

Fig. 3 Flow chart

4.3 Hard Exudates

See Fig. 7.

5 Conclusion

An efficient method for the detection of microaneurysms, hard exudate sand blood
veins has been presented. The classifier gives an average accuracy of 88%. We
conclude that image processing plays an important role in the diagnosis of DR. Future
works are to detect soft exudate to improve the accuracy of retinopathy detector.
The sensitivity and specificity of binary classification are 0.8930 and 0.9089,
respectively, which could be a satisfactory result. Furthermore, we developed an
automatic inspection app that may be utilized in both personal examination and
remote treatment. With more image data collected, we expect the accuracy is often
even more enhanced, further improving our system.
Early Detection of Diabetic Retinopathy Using Machine Learning 221

Fig. 4 Activity diagram

Fig. 5 Extraction of blood

vessels (morphological
operations applied)
222 V. V. Bandgar et al.

Fig. 6 Extraction of microaneurysms (disc-based dilation operation is applied)

Fig. 7 Extraction of exudates (conversion from RGB to CMY and binarization operation)

References

1. https://algoanalytics.com/diabetic-retinopathy-machine-learning/
2. Panse ND, Ghorpade T, Jethani V (2007) Glaucoma and diabetic retinopathy diagnosis using
image mining. Int J Computer Appl 5 (May 2015). MPI Forum: Message Passing Interface.
https://www.mpi-forum.org
3. Anisur Rahman Khan (2013) 3.2 million people in Bangladesh suffer from diabetes,
Arrkhan.blogspot.com [Online]. Available: https://arrkhan.blogspot.com/2013/10/32-million-
people-in-bangladesh-suffer.html. Accessed: 01 Apr 2016
4. Decenci ere E, Zhang X, Cazuguel G, Lay B, Cochener B, Trone C, Gain P, Ordonez R, Massin
P, Erginay A, Charton B, Klein J-C (2014, Aug) Feedback on a publicly distributed database:
the Messidor database. Image Anal Stereol 33(3):231–234
5. Ong G, Ripley L, Newsom R, Cooper M, Casswell A (2004) Screening for sight-threatening
diabetic retinopathy: comparison of fundus photography with automated colour contrast
threshold test. Am J Ophthalmol 137(3):445–452
6. Sophark A, Uyyanonvara B, Baraman S (2007) In automatic exudate detection from non-
dilated diabetic retinopathy—retinal images using Fuzzy C-means clustering; Barney B (2007)
Introduction to parallel computing. Lawrence Livermore National Laboratory
Early Detection of Diabetic Retinopathy Using Machine Learning 223

7. Gurudath N, Celenk M, Riley HB. Machine learning identification of diabetic retinopathy from
fundus images. School of Electrical Engineering and Computer Science Stocker Center, Ohio
University Athens, OH 45701USA OpenMP, The OpenMP ARB. https://www.OpenMP.org
8. Zhang F (2010) Research on parallel computing performance visualization based on MPI.
International conference. IEEE explorer
9. Gandhi M, Dhanasekaran D (2013) Diagnosis of diabetic retinopathy using morphological
process and SVM classifier. Int Conf Commun Signal Process
10. Walter T, Klein JC, Massin P, Erginay A (2002, Oct) A contribution of image processing to
the diagnosis of diabetic retinopathy–detection of exudates in colour fundus images of the
human retina. IEEE Trans Med Imaging 21(10):1236–1243. In: Klepacki D, Watson TJ (eds)
Mixed-mode programming. Research Center presentations, IBM
The Effective Use of Deep Learning
Network with Software Framework
for Medical Healthcare

Padmanjali A Hagargi

Abstract We live in an era full of unprecedented opportunities, and deep learning

technology can help us achieve new breakthroughs. Deep learning plays a pivotal
role in the exploration of exponents, the development of new drugs, the diagnosis of
diseases, and the detection of subatomic particles. It can fundamentally enhance our
understanding of biology (including genomics, proteomics, metabolomics, immuno-
hisics, etc.).This era of our lives is also facing severe challenges. Climate change
threatens food production, and may even one day explode because of limited
resources. The challenge of environmental change will also be further exacerbated
by the growing population, with a global population expected to reach 9 billion
by 2050. Coupled with the ever-evolving ability of biological neural networks to
process visual information, vision provides animals with a map of their surround-
ings, improving their ability to perceive the outside world. Today, the combination
of artificial eye cameras and neural networks that can handle the visual informa-
tion captured by these artificial eyes detonates the explosion of data-driven artificial
intelligence applications. Just as vision plays a key role in the evolution of Earth’s
life, deep learning and neural networks will enhance the capabilities of robots. The
ability of robots to understand the surrounding environment will become stronger
and stronger, and they can make decisions on their own, collaborate with humans,
and enhance human capabilities.

Keywords Medical images · Healthcare · Deep learning · Languages ·

Technology · Framework · Mathematical model

1 Introduction

Humans learn from experience. The more experience you have, the more knowledge
you can learn. In the field of deep learning in the field of artificial intelligence (AI),
this principle is also the same, that is, machines powered by artificial intelligence

P. A. Hagargi (B)
Shri Vithal Education & Research Institute’s College of Engineering, Pandharpur, Maharashtra,
India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 225
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_24
226 P. A. Hagargi

software and hardware to learn knowledge from experience. These experiences for
the machine to learn from it are determined by the data collected by the machine,
and the amount and quality of the data determines the amount of knowledge that
the machine can learn [1]. Deep learning is a branch of machine learning. Many
traditional machine learning algorithms have limited learning ability. The increase
in data volume does not continuously increase the amount of knowledge learned.
The deep learning system can improve performance by accessing more data, that is,
the machine pronoun of “more experience.” Once the machine has gained enough
experience through deep learning, it can be used for specific tasks such as driving a
car, identifying weeds between field crops, confirming a disease, detecting machine
malfunctions, and more. The deep learning network learns by discovering the intricate
structure of empirical data. By building a computational model that contains multiple
processing layers, a deep learning network can create multiple levels of abstraction
layers to represent data. For example, a convolutional neural network deep learning
model can be trained using a large number (e.g., millions) of images, such as images
of these cats. This type of neural network typically learns from the pixels contained
in the acquired image. It can group the physical features of the cat in the image
(such as paws, ears, and eyes) and classify the pixels that represent these physical
features into groups. There is a fundamental difference between deep learning and
traditional machine learning. In this example, the domain expert takes a considerable
amount of time to engineer the traditional machine learning system to detect the
physical characteristics of a cat. For deep learning, only a very large number of cat
images need to be provided to the system, and the system can learn to form the
physical characteristics of the cat [2]. For many tasks (e.g., computer vision, speech
recognition, machine translation, androbotics), deep learning systems outperform
traditional machine learning systems. This is not to say that building a deep learning
system is much easier than building a traditional machine learning system. Although
feature recognition is performed autonomously in deep learning, we still need to
adjust thousands of hyper parameters (buttons) to ensure the validity of the deep
learning model [3].

2 Literature

2.1 Medical Imaging and Healthcare

Deep learning is effective in medical imaging because of the high quality of data
and the ability to classify images by convolutional neural networks. For example,
deep learning is comparable to dermatologists in skin cancer classification, and even
better. Several vendors have obtained FDA approval for the use of deep learning
algorithms for diagnostic purposes, including imaging analysis for oncology and
retinal diseases. By predicting medical events from electronic medical record data,
The Effective Use of Deep Learning Network with Software … 227

in-depth learning has helped to make significant progress in improving the quality
of healthcare [4, 5].

2.2 Deep Learning in Medical Image Field

In the field of medical imaging, when doctors or researchers perform quantitative

analysis, real-time monitoring, and treatment planning for a particular internal tissue,
it is often necessary to understand some details of the tissue in order to make the right
treatment decisions. Therefore, biomedical imaging has become an indispensable
component of disease diagnosis and treatment, and is increasingly important.
Medical imaging techniques such as magnetic resonance imaging (MRI), Positron
emission tomography (PET), computed tomography (CT), cone beam CT, and 3D
ultrasound imaging have been widely used. For clinical examination, diagnosis,
treatment and decision making. How to make full use of artificial intelligence deep
learning method to analyze and process these large-scale medical image big data,
provide scientific methods for screening, diagnosis and evaluation of various major
diseases in clinical medicine, which is a major science urgently needed in the field
of medical image analysis. Problems and key technologies in cutting-edge medical
imaging [6].
Medical image analysis initially used edge detection, texture features, morpholog-
ical filtering, and the construction of shape models and template matching methods.
This type of analysis is usually designed for a specific task and is called a manual
custom design approach. Deep learning is a data-driven analysis task that automat-
ically learns relevant model features and data characteristics from large data sets
of specific problems [7]. Different from the manual design of models for specific
problems, the deep learning method can implicitly and automatically learn medical
image features directly from the data samples. The learning process is essentially a
solution process for optimization problems. Through learning, the model selects the
right features from the training data to make the right decisions when testing new
data. Therefore, deep learning plays a crucial role in medical image analysis.
In recent years, deep learning has continued to make significant progress, mainly
due to increasing computing power and the ever-increasing amount of available data,
as well as continuous improvement of deep learning models and their algorithms.
The essence is to build a multi-hidden layer machine learning model, use massive
sample data training, learn more accurate features, and ultimately improve the accu-
racy of classification or prediction [8]. Deep learning, which learns the characteristics
of hierarchical features from data, makes it very suitable for discovering complex
structures in high-dimensional data. At present, deep learning has been applied to
image recognition, speech recognition, natural language processing, weather predic-
tion, gene expression, and content recommendation. In other fields and in various
challenges.
The great success of deep learning in the field of computer vision has inspired
many scholars at home and abroad to apply it to medical image analysis. In recent
228 P. A. Hagargi

years, many experts have summarized, reviewed and discussed the research status and
problems of deep learning in medical image analysis. Recently, the review published
in Medical Image Analysis provides a comprehensive summary of deep learning in
medical image classification, detection and segmentation, registration and retrieval
[9].

3 Methodology

Neural networks are an important machine learning technique in the field of machine
learning and cognition. A mathematical model or computational model that mimics
the structure and function of a biological neural network to estimate or approximate
a function. The neural network is calculated by a large number of artificial neuronal
connections. In most cases, artificial neural networks can change the internal struc-
ture based on external information. It is an adaptive system in general, it has a learning
function [10]. A typical neural network has the following three parts: Architecture:
Structures specify variables in the network and their topological relationships. For
example, the variables in the neural network can be the weights of the neuron connec-
tions and the activities of the neurons. Activity Rule: Most neural network models
have a short time scale dynamics rule that defines how neurons change their stimulus
values based on the activity of other neurons. The general incentive function depends
on the weight in the network Learning Rule: Learning rules specify how weights in
the network adjust over time. This is generally seen as a dynamic rule for a long time
scale. In general, the learning rules depend on the excitation values of the neurons.
Usually a simple neural network consists of three parts: the input layer, the hidden
layer, and the output layer. Generally, the nodes of the input layer and the output layer
are fixed, and the hidden layer can be freely specified in Fig. 1 is a schematic diagram
of a simple neural network structure. The circles in the figure represent neurons, and
the arrows represent the flow of data. Corresponding to a different weight, the weight
is obtained by the network by learning itself [11].

Fig. 1 Neural network

structures
The Effective Use of Deep Learning Network with Software … 229

Fig. 2 Neuron structure diagram

Each neuron contains three parts: input, output, and computational model. The
neuron can be regarded as a calculation and storage unit. The calculation is the
calculation function of the input of the neuron. Storage is a neuron that temporarily
stores the result of the calculation and passes it to the next layer. Figure 2 shows a
typical neuron model with 3 inputs, 1 output, and 1 computational function [12].
Where x is the input, w is the weight, and a directed arrow indicating the connection
can be understood as follows: at the beginning, the transmitted signal is still a, and
there is a weighting parameter w in the middle, and the weighted signal becomes x.
w, so at the end of the connection, the size of the signal becomes x w. The weighted
signal summation at the neuron yields the output Z through the function f, as follows:


n
w i · xi
i=1

In fact, in order to achieve better results for neurons, each input is usually weighted
with an offset b, which is calculated as follows: The function f is also called the acti-
vation function. The common activation functions are sigmoid function, Hyperbolic
Tangent (tanh) function and ReLU function.

3.1 Neural Network Framework (Models)

Deep Neural Network (DNN) DNN is a fully connected network with a very large
number of layers. The number of layers determines the ability of the neural network
to characterize data using fewer neurons per layer to fit more complex functions. Note
that there is no fixed definition of “depth” here. In the field of speech recognition,
the 4th layer is considered to be relatively deep; and the 20-layer model in the field
of image recognition is not uncommon.
230 P. A. Hagargi

Convolutional Neural Network (CNN) The convolutional neural network

changes the structure of the neural network, is no longer a fully connected structure,
and greatly reduces the parameters of the network; at the same time, the network
parameters are further reduced by parameter sharing. It takes into account the spatial
structure and local features, and is very suitable for the field of image processing.
Currently, convolutional neural networks are the most widely used in the field of
medical image analysis. For example, when segmenting a medical image, the CNN
can simply classify each pixel in the image by extracting patches around a particular
pixel [13].
Cyclic Neural Network (RNN) In a normal fully connected network or CNN,
the signals of each layer of neurons can only propagate upwards, and the processing
of samples is independent at each moment, so it is also called feedforward neural
network. In RNN, the output of the neuron can be directly applied to itself at the
next timestamp; that is, the input of the i-th layer of neurons at time m contains the
output of the i-1 layer at that time and its own time at m-1.output. Based on this, a
long-term and short-term memory LSTM network has been developed [13].
Deep confidence network DBN Deep belief networks (DBN) is a probabilistic
generation model with multiple hidden units. It can be regarded as a composite model
composed of multiple layers of simple learning models. It can be used as a pre-training
part of the deep neural network and provide initial weight to the network, and then
use backpropagation or other decision algorithms as a means of tuning. Although
we can classify the various methods of deep learning, in a broad sense, NN or DNN
contains other variants. In actual use, it is often a fusion of multiple structures.

3.2 Common Types of Medical Images Required

MRI image A magnetic resonance image (MRI), which is a measure of the magnitude
of a magnetic resonance signal generated by a hydrogen atomic nucleus in human
tissues and organs and under an external strong magnetic field, and is received by a
computer for an in-.vitro nuclear magnetic resonance signal detector. The information
data is subjected to 3Dimage reconstruction. It provides very clear images of human
soft tissue anatomy and lesions as shown in Fig. 3 [14].
CT images computed tomography (CT) scans a section of a certain thickness of a
human body with a precisely collimated X-ray beam, and receives X-rays transmitted
through the section by a detector rotating with the ray beam, and finally the computer
reconstructs a 3D image of the corresponding human body section based on the X-
ray signal data received by the detector. It has a sub-millimeter spatial resolution
and provides clear human anatomical structure and lesion imaging [9]. It has been
widely used in a variety of clinical disease examinations and auxiliary diagnosis as
shown in Fig. 4 [16].
Medical X-ray images are electron density metric images of different tissues and
organs of the human body. X-ray based imaging including 2D computer radiography,
digital X-ray photography, digital subtraction angiography and mammography, and
The Effective Use of Deep Learning Network with Software … 231

Fig. 3 MR image

Fig. 4 Computed
tomography

3D spiral computed tomography, etc., have been widely used in orthopedics, lungs
Clinical disease detection and auxiliary diagnosis such as breast and cardiovascular,
but 2DX ray images cannot provide three-dimensional information of human tissues
and organs as shown in Fig. 5 [16].
Ultrasound imaging The ultrasound beam is used to scan the human body, and
the image of the internal organs is obtained by receiving and processing the reflected

Fig. 5 X-ray images

232 P. A. Hagargi

signals. In recent years, ultrasound imaging technology has been continuously devel-
oped, and new ultrasound imaging technologies such as 3D color ultrasound, ultra-
sound holography, intracavitary ultrasound imaging, color Doppler imaging and
ultrasound biological microscopy have emerged as shown in Fig. 6 [16].
Positron emission tomography (PET) uses positron emission information
emitted by the tracer labeled with radioactive elements such as F18 to image [16].
Therefore, PET image is a measure of the activity of the corresponding tracer and can
provide tumor biology. Information (such as glucose metabolism, hypoxia, prolifer-
ation, etc.) information, its standard intake value can be used to clinically identify
the tumor good/malignant. PET provides more intuitive and accurate visualization
of biological and radiobiological properties than CT and MRI as shown in Fig. 7.

Fig. 6 Ultrasound imaging

Fig. 7 Positron emission tomography

The Effective Use of Deep Learning Network with Software … 233

4 Results and Discussion

4.1 Using Types of Framework for Medical Image Analysis

CAFFE is an open source software framework, which provides a basic program-

ming framework, or a template framework, to implement deep convolutional neural
networks under GPU parallel architecture, Deep Learning and other algorithms. Fully
known as Convolutional Architecture for Fast Feature Embedding, Caffe is a widely
used open source deep learning framework (mostly the GitHub star project in deep
learning before TensorFlow) [11]. For example, the convolution of the Layer, its
input is the entire pixel point of the picture, the internal operations are the convolu-
tion operation of various pixel values and Layer arguments, and the final output is
the result of all convolution kernel filters. Each Layer needs to define two kinds of
operations, one is the forward operation, that is, the output result is calculated from
the input data, that is, the prediction process of the model; the other is the backward
operation, which is solved from the gradient of the output. Caffe also provides the
Python language interface pycaffe, which can be used to simplify operations when
approaching new tasks and designing new networks using its Python interface [12].
Theano Born in 2008 and developed and maintained by the University of
Montreal’s Lisa Lab team, Theano is a high-performance symbolic and deep learning
library, Because of its early appearance, it can be regarded as one of the ancestors of
such libraries, and was once considered to be one of the important criteria for deep
learning research and application. At the heart of Theano is a mathematical represen-
tation compiler designed to handle large-scale neural network training calculations. It
compiles various user-defined calculations into efficient underlying code and links to
various libraries that can be accelerated, such as BLAS, CUDA, and more. Theano
is easy to use is a very important feature, which is the value of other upper-level
package libraries: you don’t need to always design the network from the most basic
tensor granularity, but design the network from the upper layer of the granularity
[13, 16].
Torch positioning is an efficient scientific computing library on LuaJIT that
supports a large number of machine learning algorithms while taking precedence
over GP computing. Torch’s history is very long, but the real development is after
Facebook has open sourced its deep learning components, including Google, Twitter,
NYU, IDIAP, Purdue and other organizations have used Torch. Torch’s goal is to
make designing scientific computing algorithms easy. It contains a large library of
machine learning, computer vision, signal processing, parallel computing, video,
video, audio, and network processing, similar to Caffe, Torch [14, 15].
Keras is a minimalist; highly modular neural network library implemented in
Python and can be executed on both TensorFlow and Theano. It is designed to allow
users to perform the fastest prototype experiments, and the process of turning ideas
into results is the shortest. Keras’s programming model design is very similar to that
of Torch, but compared to Torch, Keras is built on Python and has a complete set
of scientific computing tools, and Torch’s programming language Lua does not have
234 P. A. Hagargi

such a scientific computing tool chain. Regardless of the number of people in the
community or the level of activity, Keras is currently growing faster than Torch.
MXNet is an open source, lightweight, portable and flexible deep learning library
developed by DMLC (Distributed Machine Learning Community), which allows
users to mix symbolic programming patterns and instructional programming patterns
to maximize Efficiency and flexibility are now the deep learning framework recom-
mended by AWS. Many of MXNet’s authors are Chinese, and their biggest contribu-
tion is Baidu. At the same time, many authors come from deep learning projects such
as cxxnet, minerva and purine2. It is the first to support multiple GPUs and decen-
tralized in each framework, while its decentralized performance is also very high.
Its upper computational graph optimization algorithm allows symbolic computation
to be performed very quickly, and saves memory. Turning on the mirror mode saves
memory, and even training other frameworks on some small memory GPUs because
of insufficient video memory [15]. MXNet’s various levels of system architecture
(below the hardware and operating system bottom layer, layer by layer is more and
more abstract interface) as shown in Fig. 8.
AlexNet is a deep network that was first applied to ImageNet, and its accu-
racy is greatly improved compared to traditional methods. It starts with 5 convo-
lutional layers, followed by 3 fully connected layers, as shown in the Fig. 9: Alex
Krizhevs proposed AlexNet to use the ReLU activation function instead of the anh

Fig. 8 MXNet system

architecture

Fig. 9 AlexNet structure

The Effective Use of Deep Learning Network with Software … 235

Fig. 10 Fully connected layer and dropout layer

or Sigmoid activation function used in the early days of traditional neural networks.
ReLU mathematics is expressed as:

f (x) = max(0, x)

The advantage of ReLU over Sigmoid is that it is faster to train because the
derivative of Sigmoid is very small in the stable zone, so the weight is basically
no longer updated [16]. This is the problem of gradient disappearance. So AlexNet
uses ReLU behind both the convolutional layer and the fully connected layer [10].
Another feature of AlexNet is that it reduces the overfitting of the model by adding
a dropout layer.
Dropout layer behind each fully connected layer [11]. The Dropout layer randomly
closes the neuron activation values in the current layer with a certain probability, as
shown in the Fig. 10.
DenseNet The structure has been developed by linking each layer to all other
layers in the style of forward feed, adopting the fact that there are shorter connections
between the intersecting networks near the input and those close to the output, which
can be deeper, more accurate and efficient to train. For each layer, the property
maps of all previous layers as input, their feature maps are also used as input to
all subsequent layers. DenseNets achieves significant improvements to the latest
technology in most areas, requiring less memory and calculation to achieve higher
performance [6]. In the study, the pretrained models were tested by making them
suitable for the classification task. Each model was operated with a maximum of
1812 iterations and a single GPU [11, 12]. The operating performances of the models
are as in Table I. Validation saccuracy as a model performance criterion has been
shown. According to the results, the highest validation accuracy was realized with
the DenseNet201 network. When the operating times are considered, DenseNet201
has the longest working time due to the number of parameters [16].
236 P. A. Hagargi

Deeplearning4j (hereafter referred to as DL4J) is an open source, distributed

deep learning project in Java and Scala environments. DL4J integrates Hadoop and
Spark and is designed to run on distributed GPUs and CPUs. It was led by Skymind,
a San Francisco-based business intelligence and enterprise software company. Team
members include data experts, deep learning experts, Java system engineers, and
perceptual robots. The book published by O’Reilly for DL4J is already available at
Amazon. Everyone has heard about big data. By publishing open source deep learning
tools, we hope to bring business to the era of intelligent data [14]. This is why we
chose to implement this framework in the Java environment. By using these tools,
data experts can recognize the power of scalable deep learning algorithms. These
algorithms have long been used in companies with dedicated deep learning teams
such as Google, Facebook, Microsoft, Baidu, and Netflix. Through deep learning
algorithms, unsupervised learning can be achieved, allowing data experts to avoid a
lot of tedious work in the feature extraction process, and thus focus more on more
interesting tasks. More broadly, we believe that these algorithms can help build a
safer, smarter, more transparent and more efficient social environment [12].
Chainer is a deep learning framework released by the Japanese company Preferred
Networks in June 2015 [15]. Chainer describes its characteristics as follows.
Powerful: Supports CUDA calculations, which can be GPU-accelerated with just a
few strokes of code, and can be executed on multiple GPUs with a few changes.
Flexible: Supports a variety of feedforward neural networks, including convolu-
tional networks, loop networks, and recursive networks, supporting Define-by-Run
in execution.
Intuitive: Feedforward calculations can introduce various control flows from Python,
while being back-spread without interference, simplifying the difficulty of debugging
errors.
The vast majority of deep learning frameworks are based on “Define-and-Run”,
which means that you need to define a network first and then feed the data to the
network (mini- batch). Because the network is pre-statically defined, all control logic
needs to be inserted into the network as data, including defining network structure
files like Caffe, or defining the network using programming languages like Theano,
Torch, and TensorFlow. In contrast, Chainer, the network is defined in the actual
implementation, Chainer stores the results of the historical execution, rather than the
network’s structural logic, so that you can easily use the control flow in Python, so
you can directly Use conditional controls and loops in the network [13].

4.2 Using Programming Language in Deep Learning

Python programming language that has the most popularity in recent times is a
Python. Python, in 1991 by Guido van Rossum, it is high level, an open source
general aim programming language. This dynamic programming language is hold
The Effective Use of Deep Learning Network with Software … 237

up with object oriented, vital important, procedural development paradigms and

functional. The most programming language that used in deep learning is a Python.
Python programming language has a range of libraries and instruments to help
and Supporting of deep learning. TensorFlow and Scikit are two famous Python
developers deep learning libraries. There are several libraries used in C++ such as:
Theano is a Python library for defining and evaluating mathematical expressions
using sequences. It makes it easier to write deep learning algorithms in Python.
Many other libraries are developed based on Theano: Keras is a streamlined, highly
modular neural network library similar to Torch [14].
C++ is a widely used programming language. Among the many C++ developers,
most people write code in an object-oriented way: the C++ projects we touch every
day are basically organized in this style; almost every C++ tutorial will be discussed
in most of the space. Object-oriented; every programmer with years of C++ develop-
ment experience will have his own insights into object-orientation. Object-oriented
has become a mainstream in the C++ development circle, so that in some people’s
view, C++ is similar to many programming languages, just an object-oriented dialect.
But in fact, C++ supports not only the object-oriented programming style. It also
supports another programming style: generic, and thus derived from a program-
ming method, i.e., calculation and compile programming element. C++ is one of
the most common and oldest programming languages. Including TensorFlow, most
deep learning systems support C++. TensorFlow’s C++ API provides mechanisms
for constructing and executing a data flow graph. There are several libraries used in
C++ such as: Caffe, CXXNET, DeepLearning, Eblearn, Mocha.
C# language was made by Anders Hejlsberg at Microsoft and propelled in 2000.
C# is a basic, present day, adaptable, object oriented, safe, and open source program-
ming language. C# is a standout amongst the most adaptable programming languages
on the planet. C# enables designers to construct all sort of utilizations including
Windows customers, supports, Web applications, portable applications, and backend
frameworks. C# can be used for deep learning applications via a .NET Core deep
learning platform, CNTK. CNTK is a framework used in deep learning for training
a model the CNTK liberary is can used to evaluate the model in the application.
CNTK supports model evaluation from C++, Python, C#/.NET, and Java. Starting
from v2.1, CNTK also supports Universal Windows Platform (UWP).
R language is a dynamic programming language, based on array, object
oriented, indispensable, functional, procedural, and reflective computer program-
ming language. R language appeared in 1993 but has become famous in past few
years among data scientists and deep learning developers for its statistical and func-
tional algorithm features. Ross Ihaka and Robert Gentleman at the University of
Auckland, New Zealand was created R language. R language is open source and
available on r-project.org and Github. Currently R is managed and developed under
the R Foundation and the R Development Core Team. The current version of R
language is 3.5.2 that was released on Dec 20, 2018. R language is one of the most
popular programming languages among data scientists and statistical engineers. R
238 P. A. Hagargi

supports Linux, OS X, and Windows operating systems. The R interface to Tensor-

Flow lets you work productively using the high-level Keras and Estimator APIs
[15].
Java and JavaScript Java is the most famous programming language in the world.
Java was developed by James Gosling at Sun Microsystems that later acquired by
Oracle. Java Script is the most popular Web scripting programming language. There
are several deep learning libraries and frameworks that support Java and JavaScript,
for instance Deeplearning4j and it is an open source library.
Julia, Go, Shell, Prolog, Lisp, Ada, TypeScript, and Scala Several other
languages which provide deep learning support and usage include Julia, Go, Shell,
Prolog, Lisp, Ada, TypeScript, and Scala. Deep learning functionality depends on
the frameworks and libraries available to developers. Two of the most popular deep
learning frameworks are TensorFlow and scikit-learn [16].

5 Conclusion

Using the best combination of programming languages and images attempts are made
to get best results for diagnosis. The field of medical image analysis has begun to
pay attention to the development of deep learning techniques on these key issues.
However, the transition from an artificial-based system to a system that learns features
from data are gradual reviewed the application of deep learning in medical image
analysis. Although they have done a lot of work, it seems to be flawed only from a
computer perspective. In recent years, new advances in deep learning techniques have
provided new insights into medical image analysis that allow only the morphological
and/or texture patterns in images to be found from the data. Although deep learning
techniques seem to have achieved the most advanced performance in many medical
application studies, few have shown that it can go beyond traditional methods. At
the same time, the current research on deep learning in the field of medical image
analysis is mainly based on technology, and the evaluation indicators used are also
evaluation indicators in the computer field. For any medical application, we would
prefer to see that the relevant technology is evaluated according to medical rules, for
example, through multi-center, randomized, controlled research methods to prove
that the technology has a more significant advantage than the previous technology.
Furthermore, an important challenge in medical image training is that the number of
training samples for most depth models is related to the number of learning param-
eters. Therefore, how to reduce over-fitting has always been a problem. When deep
learning training results are sent to new central applications, the models need to
migrate learning to maintain performance, which undoubtedly leads to low repro-
ducibility of clinical applications. Stability and reproducibility are the basic premise
that the technology can be widely used in clinical practice. Therefore, researchers
in the field of deep learning should pay attention to how the repeatability of the
algorithm in the prospective sample is guaranteed. In addition, the quality of the data
used for training is another cause of catastrophic results.
The Effective Use of Deep Learning Network with Software … 239

The problem of random noise is easier to solve, it can improve performance

through some parameter settings (technically called label smoothing or soft labeling).
Structural noise is different; it adds a truly different signal that will actually affect the
model learning. The problem is more serious when the noise comes from the same
source as the actual data, because the model will confuse the class, and the black box
problem of the current deep learning method will be infinitely magnified.
In addition to these technical problems, the theoretical obstacles seem to be more
troublesome. Current methods of deep learning do not have causal logic, only corre-
lation calculations, which may have inherent limitations on the cognitive tasks they
can perform. The value of the resources that need to be consumed to achieve our
desired performance goals should also be carefully assessed to avoid falling into
the trap of non- polynomial time issues. In summary, deep learning technology
brings a new method for the most important feature extraction of machine learning.
If deep learning technology can perform well on all problems, it will bring great
help to medical image analysis and processing. However, deep learning technology
is not the ultimate algorithm. It is only a representative of the connected school
in several schools of artificial intelligence, and its performance limits need to be
reasonably evaluated. Deep learning relies too much on high-quality big data, and
its economic effects may not be appropriate for medical images. Care must be taken
in the application of deep learning techniques for application area.

References

1. Arevalo J, González FA, Ramos-Pollán R, Oliveira JL, Lopez MA G.( 2016) Representation
learning for mammography mass lesion classification with convolutional neural networks.
Comput Methods Programs Biomed 127:248−257
2. Gao XT, Lin S, Wong TY (2015) Automatic feature learning to grade nuclear cataracts based
on deep learning. IEEE Trans Biomed Eng 62(11):2693–2701
3. Hagargi PA, Kumbhar S (2020) Combined effects of IoTs and medical sensors for effective
application of smart health care system. J Eng Comput Archit 10(4):11–19. ISSN NO: 1934-
7197
4. Hagargi PA, Danake PV, Gaikwad SN, Mahajan DS, Thite KG, Supekar PS (2020) Internate of
things based patient health monitoring system. J Aegaeum 8(4):221–228. ISSN NO: 0776-3808
5. Hagaragi PA, Shubhangi DC (2018) Brain tumor MR image fusion using most dominant
features extraction from wavelet and curvelet transforms. J Emerg Trends Innov Res (JETIR)
5(5):3–38. ISSN-2349-5162
6. Hagaragi PA, Shubhangi DC (20181) An efficient approach for brain tumor detection and
classification in multi-model MR brain images using RPCA and QT decomposition fusion and
ANN classification technique. Int Res J Eng Technol (IRJET) 5(5):26–40, e-ISSN: 2395-0056,
p-ISSN: 2395-0072
7. Hagaragi PA, Shubhangi DC (2018) Brain tumour detection and ART classification technique
in MR brain images using RPCA QT decomposition. Int Res J Eng Technol (IRJET) 5(4):1717–
1725, e-ISSN: 2395-0056, p-ISSN: 2395-0072
8. Hagaragi PA, Shubhangi DC (2017) A decomposition technique for image fusion to detect
brain tumor using T1, T2 weighted MR images. Int J Curr Eng Sci Res (IJCESR) 4(12):62–73
(Part-VII) https://doi.org/10.21276/ijcesr, o-ISSN: 2394-0697, p-ISSN: 2393-8374
240 P. A. Hagargi

9. Hagaragi PA, Shubhangi DC (2017) A survey on fusion based methods using MR image
analysis for disease diagnosis. Int J Eng Dev Re (IJEDR) 5(4):328–333. ISSN-2321-9939
10. Hagaragi PA, Shubhangi DC (2016) Brain tumor MR image fusion using wavelet and
curvelet transforms. In: Proceedings of 2nd IEEE international conference on engineering
and technology ICETECH-2016. IEEE, 978-1-4673-9916-6/16
11. Hagaragi PA, Shubhangi DC (2016) Multi model MRI brain image fusion technologies and
graph cut segment. Int J Eng Future Technol 1(1):55–65. ISSN 2455-6432
12. Hagaragi PA, Shubhangi DC (2014) Comparisons of different methods of image edge detection
with morphologic method. J Innov Comput Sci Eng 4(1)
13. Hagaragi PA, Shubhangi DC(2012) Image edge detection using morphological operation. Int
J Res Comput Appl Manag. ISSN-2231-1009
14. Fakoor R, Ladhak F, Nazi A, Huber M (2013) Using deep learning to enhance cancer diagnosis
and classification. In: International conference on machine learning. Omnipress
15. Lee LK, Liew S-C (2015) A survey of medical image processing tools. https://doi.org/10.13140/
RG.2.1.3364.4241. In: 4th international conference on software engineering and computer
systems, Kuantan, Malaysia
16. He K, Zhang X, Ren S, Sun J (2016) Deep residual learning for image recognition. In: IEEE
conference on computer vision and pattern recognition. IEEE Comput Soc 770–778
Detection of Brain Tumor Using Image
Processing and Neural Networks

Vanshika Dhillo, Dipti Sakhare, and Shilpa Rudrawar

Abstract Artificial Intelligence (AI) is an umbrella consisting of many small blocks

like machine learning, evolution computation, robotics, vision, natural language
process and planning, speech processing etc. In the past years AI has developed
a lot and given its share to make human life better, easy and compact. The word
“technology” is a term which can be defined in many ways. The definition of the
word keeps evolving with the continuous development in various fields. Decades
ago, it used to take an entire room to accommodate a single computer but now we
use the same computer at the ease of our fingertips. With the rapid development
in technology mankind’s expectations and needs have also increased. Human race
demands accurate results in less time with easier methods. The current rapidly devel-
oping field is AI which has also shown some extra ordinary results like the robot
Sophie, Alexa, Siri etc. Mankind has very willingly adapted the use of AI in daily
life and wants to excel more in the field. Neural Networks which is a very small
section of AI but can be used vastly in various fields. The healthcare section has also
improved its facilities and increased life expectancy. The motive behind this project
is to use neural networks in the medical field for greater accuracy and instant results.

Keywords Image processing · Biomedical · Neural networks · Artificial

intelligence · Brain tumor

1 Introduction

In the past few decades medical studies have adapted a lot of changes in their approach
toward treating a patient which has improved the standards of living and provided
comfort. Biomedical tools have been developed which are used by the doctors to
analyze and study about a particular disease or disorder. Even after so much thrive
there are some life taking afflictions like cancer, leukemia or brain tumor which if
not detected on time can be proved as fatal.

V. Dhillo (B) · D. Sakhare · S. Rudrawar

Department of E&TC, MIT Academy of Engineering, Pune, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 241
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_25
242 V. Dhillo et al.

As per a survey taken by the National Brain Tumor Foundation (NBTF) the
death rate due to brain tumor has been increasing. The increase in deaths is caused
due to many causes like treatment not given on time, tumor not detected due to
poor image quality or contrast and many more. The idea of this project is to avoid
such situations and improve the quality of treatment. Machine learning has high
dimensional features which can definitely be utilized in such conditions. This will
not only help in increasing accuracy but also help the doctors to cure and recover the
patient within the required time. The target is to build a self-learning system which
scans the images and detects the tumor. Neural networks are nothing but a multilayer
architecture which is utilized for quantitative analysis. For example, the neurons in
our body they are a network of cells which perform functions as directed. The tasks
performed by neural networks is similar to the task of neurons.

2 Literature Survey

Image processing is being used in the medical field for a very long time and for many
tests. In order to detect the brain tumor doctors, suggest for tests like the positron
emission tomography (PET), compute tomography (CT) and excessively used is the
magnetic resonance image (MRI). MRI is preferred by the majority because it does
not emit harmful rays and has no side effects on the body on the other hand it gives
a precise view of the disorder. When it comes to image processing there are various
types of methods which are used to analyze or detect an image.
At the time of detecting any image many parameters are taken into consideration
like the color, contrast, pixels, intensity, boundaries, dimension etc. According to
these parameters any image can be categorized and the optimal required function
can be performed. In order to detect the brain tumor many processes were adapted
and with the development of technology many other processes and tests are being
developed [1]. Mentioned are some of the processes which have been used like
thresholding, region based, contour and shape, statistical based and machine learning
but these methods came with their own demerits. In the process of thresholding the
pixels separated and then grouped together on the basis of their intensity values [1].
For some cases, this method proved to be useful but, in some cases, where the tumor
is at its very first stage and the intensity or the color contrast of the lump formed
is very fade and the pixels could be grouped wrong showing no signs of tumor [2].
The second method in the list which was tested was the contour and shape method
in which boundaries and segments are detected in this the boundaries and segments
of the lump are detected but as the lumps are of irregular sizes and sometimes, they
are formed near the inner lining of the brain due to which boundaries could merge
and accuracy wasn’t gained [3]. For the third method which was the region-based
method groups of similar regions and made and then further detected. Last method in
the list was statistical based which was based on probable values, labels and optimal
Detection of Brain Tumor Using Image Processing … 243

distribution [4]. Machine learning when is further sub divided into categories in
which one of them is called as deep learning which will be used in this project. Used
methods, the approach, its merits and demerits will be discussed further.

3 Methodology

Deep Learning is a modern well-known technique which involves networks with

many layers to make predictions. These networks are in the headlines because due to
their exceptional results and accuracy. Before starting the implementation and seeing
the miracles of deep learning and neural networks it is very important to understand
how these networks work and how to make them.
(1) Neural Network and its formation
The hypothesis of this theory started in the ancient times, in the eighteenth century
by Alexander Bain (1873) and William James (1890). Through their work in due
course, they both proposed that thoughts and activities in the body are nothing but
interactions among the neuron and the brain. Similarly, the same theory can be used
in machines and a system can be made which is self-learning [5]. Neural network
is a circuit of multiple neurons and nodes in AI form which is used to perform any
desired task.
Neural networks are formed by multiple layers of information provided by the
user, tasks and functions but if one has to define a simple neural network it consists
of mainly three layers which is the input layer, hidden layers and the output layers
[6]. The first layer which is the input layer is where the input command is given
and the output layers is from which the required result is achieved. The second layer
which is the hidden layer is where all the tasks and functions take place. As shown
in Fig. 1 is the structure of a very basic neural network. The network trained can be

Fig. 1 Basic structural layout of a neural network

244 V. Dhillo et al.

Fig. 2 Image pre-processing

a universal set which includes everything or it can be a sub set which is dedicated
for a particular task [7]. To detect any image, it’s very important to train the network
and then assign proper functions for the required task. In order to train a network for
a set of images the pre-processing on the images necessary [8].
(2) Image Pre-processing
As mentioned earlier images can be classified into many sections and on the basis of
many parameters. In order to train a network all, the images which are being used or
which are being stored should be uniform [2]. Images can be divided on the basis of
RGB and Gray-scale. MRI scans are often in the gray scale format so all the images
stored in the folder should preferably be in the gray scale format. Dimensions of the
image should be uniform as well for which one should resize the image and change
its dimensions to ‘227 × 227’ which is the required dimension when a network is
being formed. Figure 2 shows the image pre-processing for RGB image hence it
shows ‘m × n × 3’ in which ‘m and n’ are the dimensions of the image [9].
3. Transfer Network
While we use deep learning, we often come around pre-trained or trained networks
which already have a defined data storage. For updating the network or add new data
a process is used which is called transfer network. In order to perform this process,
changes are made in the input layer and the layer just before the outcome. Another
important factor which should be taken into consideration is the memory manage-
ment [6]. The method of transfer network also helps in maintaining the storage. As
mentioned before neural networks are a self-learning system and it stores only the
required information in the dedicated folders and you don’t land up with double
images in the folder. When you compare Figs. 2 and 3 you can notice the changes in
networks. Figure 3 is nothing but the changes in the last layers of the same system
which is showed in Fig. 2. After you have completed the process of transferring the
network you have to again train the network with the new stored images.
(4) Network Training
The last task to perform is to train the network with the given data set and the
given commands. In order to detect the brain tumors, the network had to be trained
Detection of Brain Tumor Using Image Processing … 245

Fig. 3 Transfer of network

on numerous MRI scans and perform segmentation in order to detect the tumor
accurately. Tumors are nothing but lump like structures which are formed in the
brain. At times multiple tumors are developed within the brain. Human brain can
be divided into three main parts which is the cerebrum, the brain stem and the
cerebellum. After which it is further sub-divided into hemispheres which are the
frontal lobe, parietal lobe, temporal lobe and the occipital lobe. These lobes working
conjunction to maintain the functions of the body and each lobe carries out multiple
functions. Tumor can be formed in any of these lobes. The network can be trained
for detecting the tumor and also detect in which part of the lobe it is formed. The
input given has to be a data set of multiple scans and the region of the lobes has to
be defined.

4 Results

When compared to other methods listed above like manual inspection, digital photog-
raphy and automated analyzer the use of neural networks in combination with data
sciences has proved to provide more accuracy with a greater speed. The reason why
this method works exceptionally is because of its trained network. CNN (convo-
lutional neural network) is a mathematical process mainly used for image analysis
and has an adaptive behavior. The network gets trained over a thousand pictures and
improves its accuracy after each run. One can not only compare its speed or accu-
racy but also the how it maintains the data set and memory. The intension of this
project was to obtain results similar to the results given by a specialist. There are soft-
wares available to perform deep learning but the most feasible would be MATLAB
as there are inbuilt functions available like the image processing tool box and the
deep learning tool box. One can practice or test functions on MATLAB’s pre-trained
networks such as ‘alexnet’. The circled part in Figs. 4 and 5 is the tumor which was
246 V. Dhillo et al.

Fig. 4 Detection of the tumor

Fig. 5 Detection of the tumor

detected after segmentation. The results conclude that with the help of this system
single as well as multiple tumors can be detected with greater accuracy.

5 Conclusion

Image processing has found its way in the healthcare stream and will continue to grow.
There is a broad perspective of using image processing for many other tests as well
like detecting the hemoglobin, WBC and RBC in the blood. This method can also be
used for retinopathy which can detect diabetes and detect cataract as well. Adaptive
methods like this have proved to be useful and can be further developed as well.
With advancing technology and systems in industries the style of living is changing.
These new systems can be a little complicated to implement when compare to the
earlier methods but are more effective and life changing. Biomedical imaging hand
in hand with data science will not only make the process simpler but systematic as
Detection of Brain Tumor Using Image Processing … 247

well. Reduction of errors, less human efforts and advancement in the life expectancy
due to early detection are the aims to be achieved to further which we can add a more
digitalized system.

References

1. Abdalla HEM, Esmail MY (2018) Brain tumor detection by using artificial neural network.
In: International conference on computer, control, electrical, and electronics engineering
(ICCCEEE), Khartoum, pp 1–6. https://doi.org/10.1109/ICCCEEE.2018.8515763
2. Fathurachman M, Kalsum U, Safitri N, Utomo CP (2014) Heart disease diagnosis using extreme
learning based neural networks. In: International conference of advanced informatics: concept,
theory and application (ICAICTA), Bandung, pp 23−27. https://doi.org/10.1109/ICAICTA.
2014.7005909
3. Abdullah AA, Chize BS, Nishio Y (2012) Implementation of an improved cellular neural network
algorithm for brain tumor detection. In: International conference on biomedical engineering
(ICoBE), Penang, pp 611–615. https://doi.org/10.1109/ICoBE.2012.6178990
4. Jemimma TA, Vetharaj YJ (2018) Watershed algorithm based DAPP features for brain tumor
segmentation and classification. In: International conference on smart systems and inventive
technology (ICSSIT), Tirunelveli, India, pp 155–158. https://doi.org/10.1109/ICSSIT.2018.874
8436
5. Yazid MHA, Satria MH, Talib S, Azman N (2018) Artificial neural network parameter tuning
framework for heart disease classification. In: 5th International conference on electrical engi-
neering, computer science and informatics (EECSI), Malang, Indonesia, pp 674–679. https://
doi.org/10.1109/EECSI.2018.8752821
6. Ai H, Fang M, Lvov YM, Mills DK, Jones SA (2002) Applications of the electrostatic layer-
by-layer self-assembly technique in biomedical engineering. In: Proceedings of the second joint
24th annual conference and the annual fall meeting of the biomedical engineering society]
[engineering in medicine and biology, Houston, TX, USA, vol 1, pp 502–503. https://doi.org/
10.1109/IEMBS.2002.1136916
7. Joshi S, Borse M (2016) Detection and prediction of diabetes mellitus using back-propagation
neural network. In: International conference on micro-electronics and telecommunication
engineering (ICMETE), Ghaziabad, pp 110–113. https://doi.org/10.1109/ICMETE.2016.11
8. Lekha S, Suchetha M (2018) Real-time non-invasive detection and classification of diabetes
using modified convolution neural network. IEEE J Biomed Health Inform 22(5):1630–1636.
https://doi.org/10.1109/JBHI.2017.2757510
9. Banowati C, Novianty A, Setianingsih C (2019) Cholesterol level detection based on iris recogni-
tion using convolutional neural network method. In: IEEE Conference on sustainable utilization
and development in engineering and technologies (CSUDET), Penang, Malaysia, pp 116–121.
https://doi.org/10.1109/CSUDET47057.2019.9214690
Automatic Guided System for Image
Resolution Enhancement

Neeta P. Kulkarni, J. S. Kulkarni, and S. M. Karve

Abstract High-resolution images decide the performance of the system. Image reso-
lution enhancement is the process to extend no of pixels. During this, we increase
the quality of the image to form it suitable for all applications. Within the proposed
method, we’ve used a combination of DWT, SWT and Interpolation, i.e. combina-
tion of 3 traditional methods. DWT has edge loss problem, so alongside it, SWT
is employed to beat edge loss problem. Initially, DWT of a low-resolution image is
taken. The LL band of this is often interpolated by a factor of two and it’s combined
with SWT of the input low-resolution image. Again inverse DWT is taken of the
resultant image.

Keywords Image resolution enhancement · Wavelet transform · Interpolation ·

DWT · SWT · PSNR

1 Introduction

Image resolution improvement is one in all the foremost customary strategies of

low-level digital image process. Digital image process field defines the treatment of
digital pictures by suggests that of a pc. A method of low-level improvement has
each its inputs and outputs as pictures.
The image resolution is also a big side. The resolution suggests that the precise
look of the image to the user. Image process (Image Resolution Enhancement) has
several applications among the arena of Medical, Satellite Image process, Industrial
applications. Image resolution includes rising the standard of the image by increasing
the quantity of pixels in order that the image is a lot of appropriate for any applications.
The foremost crucial objective of improving quality is to increase the standard of the
image.

N. P. Kulkarni (B) · S. M. Karve

Department of ENTC, SVERI’s College of Engineering, Pandharpur, India
e-mail: [email protected]
J. S. Kulkarni
Department of ENTC, Vishwakarma Institute of Information Technology, Pune, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 249
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_26
250 N. P. Kulkarni et al.

Characteristics of the image and to restrain some useless attributes of the image.
The image, which has been improved, can satisfy some unique analysis better than
the primary one. It’s the method of manipulating a picture in order that resultant
image is more suitable than the first one for a specific application.

2 Motivation

In recent years there’s an increase in the demand for better quality images within
the various applications like medical, astronomy, visual perception. Image resolu-
tion improvement is to boot wide helpful for satellite image applications that embody
building construction, bridge recognition, in GPS technique. For image improvement
method, there a time-invariant transform. To restore the interpretation invariance is to
average some slightly different DWT, called decimated DWT, to define the stationary
wavelet transform (SWT). The Stationary wavelet transform (SWT) could also be a
wavelet transform algorithm designed to beat the shortage of translation-invariance
of the discrete wavelet transform (DWT). Translation invariance is achieved by
removing the down samplers and up samplers within the DWT and upsampling
the filter coefficients. The SWT is an inherently redundant scheme because the
output of every level of SWT contains an equivalent number of samples because
the input, so for the decomposition of N levels there’s a redundancy of N within the
wavelet coefficients. Hasan Demirel and Gholamreza Anbarjafari propose a picture
resolution enhancement technique [1] supported interpolation of the high-frequency
sub-band images obtained by discrete wavelet transform (DWT) and therefore, the
input image. The sides are enhanced by introducing an intermediate stage by using
stationary wavelet transform (SWT). During this correspondence, one level DWT
[2] (with Daubechies 9/7 as wavelet function) is employed to decompose an input
image into different sub-band images.
Another transform is the stationary wavelet transform (SWT) [3]. The Discrete
Wavelet Transform isn’t a time-invariant transform. To restore the interpretation
invariance is to average some slightly different DWT, called decimated DWT, to
define the stationary wavelet transform (SWT). In DWT [4] basic computational,
the step may be a convolution followed by decimation. The decimation retains even
indexed elements. But the decimation might be administered by choosing odd indexed
elements rather than even indexed elements. This selection issues each step of the
decomposition method, and thus at each level, we tend to selected oddly. It will thus
by suppressing the down-sampling step of the decimated algorithmic program and
instead up-sampling the filters by inserting zeroes between the filter coefficients.
Automatic Guided System for Image Resolution Enhancement 251

3 Method Adopted

The planned technique uses a combination of DWT, SWT and Interpolation to rein-
force low-resolution image to high-resolution image. The brief steps within the
method are as given below. The proposed method is as shown in Fig. 1.
Step 1: Take standard benchmark high-resolution image (say 512 × 512).
Step 2: Downsample the high-resolution image 2 times by using DWT. During
this project, the discrete wavelet transform algorithm is meant using Haar [5] wavelet
function for DWT decomposition and IDWT. An obtained low-resolution image is
of size 128 × 128.

High-Resolution

Low-
resolut
ion

SWT
DWT

LL LH HL HH
LL LH HL H

Corrected Cor Correcte

LH rect d HH

IDWT

High-
resolution

Fig. 1 General block diagram of the proposed method

252 N. P. Kulkarni et al.

Step 3: Apply 1 Level 2 Dimensional Discrete wavelet transform to the obtained

low-resolution image of size 128 × 128. DWT decomposes the low-resolution image
into 4 sub-bands (say a, h, v, d). These are representing LL, LH, HL and HH, respec-
tively. In DWT size of a, h, v, d is down-sampled by 2. i.e., its size is 64 × 64. Within
the proposed technique, 3 high-frequency sub-band images of DWT are interpolated
using bi-cubic interpolation by a factor of two. Then it’s size becomes 128 × 128
[6] (we get new h, new v, new d).
Step 4: Apply 1 level 2 Dimensional Stationary wavelet transform to the obtained
low-resolution image. SWT decomposes the low-resolution image into 4 sub-bands
which are LL, LH, HL and HH, respectively. It’s not down-sampled. 3 high-frequency
sub-bands of SWT are added with interpolated high-frequency sub-band images of
DWT. Add h1, v1, d1 with new h, new v, new d to urge corrected sub-bands.
Step 5: Obtained bands and input image are again interpolated by a factor of α/2.
Here α is the enlargement factor. Example 128 × 128 images are enhanced to 512 ×
512, here α = 4. Here input image is taken instead of considering LL band because
low-resolution images are obtained by down-sampling the high-resolution image.
Therefore the input image is often treated as low waveband of the high-resolution
output image.
Step 6: These interpolated bands and interpolated input image in step 5 are
combined through IDWT to get a high-resolution image at output. This obtained
image preserves the sides (high-frequency components). So as to point out the effec-
tiveness of the proposed method over the traditional and state-of-art image resolution
enhancement techniques, four well-known test images (Lena, Elaine, Baboon, and
Peppers) with different features are used for comparison. The quantitative (peak
signal-to-noise and root mean square error) and visual results show the prevalence
of the current technique over the conventional and state-of-art image resolution
enhancement techniques [7]. The high-resolution image is down-sampled to get
low-resolution image and the obtained low-resolution image is enhanced by using the
opted way of implementation. Peak signal to noise ratio is excellent over conventional
image resolution enhancement techniques.

4 Results and Discussions

The proposed method has been tested with a standard set of test grey images and
colour images. The image enhancement is carried out in the following steps.
1. As shown in Fig. 2 the high-resolution standard benchmark input images (size
512 × 512) are taken.
2. As shown in Fig. 3, these high-resolution input images are downsampled to
128 × 128 and low-resolution images are obtained.
3. As shown in Fig. 4, this low-resolution image is enhanced to 512 × 512 by the
proposed method. First, take level 2d DWT of low-resolution image.
4. Then take 1 level 2d SWT of the low-resolution image. It is as shown in Fig. 5.
Automatic Guided System for Image Resolution Enhancement 253

Fig. 2 High-resolution
image

Fig. 3 Downsampled image

254 N. P. Kulkarni et al.

Fig. 4 2D DWT of the

down-sampled
low-resolution image

Fig. 5 2D SWT of a
down-sampled
low-resolution image
Automatic Guided System for Image Resolution Enhancement 255

Fig. 6 Enhanced
high-resolution resolution

5. This low-resolution image is enhanced to 512 × 512 by the proposed method.

The resultant enhanced image is shown in Fig. 6.
6. The original high-resolution image is used as the ground truth image to calculate
Peak signal to noise ratio.

Proposed work gives the better result as compared to traditional methods. The
performance is calculated by calculating Peak Signal to Noise ratio (PSNR) as given
in Eq. (1) Signal to Noise Ratio (PSNR) is used to analyse the quality of image,
sound and video files in dB (decibels). The original image and resultant image are
having different PSNR values.

PSNR = 10 log 10(255 × 255/MSE)

MSE Mean Square Error (1)

Table 1 shows that PSNR value has been improved by using the proposed method
as compared to benchmark methods.

5 Conclusion

This method is giving superior results as compared to traditional methods. By

changing the equation and sort of filter used, the results can be further improved
by changing the sort of filter and its equation. The proposed method is are often also
256 N. P. Kulkarni et al.

Table 1 PSNR values (DB)

Method Lena Elaine
for images enhanced from
128 × 128 to 512 × 512 (α = Bilinear 26.34 25.38
4) Bi-cubic [1] 26.86 28.93
WZP (db9/7) 28.84 30.44
DWT 34.79 32.73
CWT [2] 33.74 33.05
SWT and DWT (db9/7) 34.82 35.01
Proposed method 35.53 36.65

tested for colour images and give higher results. This is a good method for image
resolution enhancement.

References

1. Demirel H, Anbarjafari G (2010) Satellite image resolution enhancement using complex wavelet
transform. IEEE Geosci Remote Sens Lett 7(1):123–126
2. Zhao S, Han H, Peng S (2003) Wavelet domain HMT-based super image resolution. In:
Proceedings IEEE international conference image processing, vol 2, pp 933–936
3. Demirel H, Anbarjafari G (2011) Image resolution enhancement by using discrete and stationary
wavelet decomposition. IEEE Trans Image Process 20(5):1458–1460
4. Demirel H, Anbarjafari G, Izadpanahi S (2009) Improved motion-based localized super-
resolution technique using discrete wavelet transform for low-resolution video enhancement. In:
Proceedings 17th European signal processing conference, Glasgow Scotland, pp 1097−1101
5. Temizel A, Vlachos T (2005) Wavelet domain image resolution enhancement using cycle-
spinning. Electron Lett 41(3):119–121
6. Phanindra P, Chinna Babu J, Usha Shree V (2013) VLSI implementation of medical image
fusion using Haar transform. Int J Sci Eng Res 4(9). ISSN 2229-5518
7. Anbarjafari G, Demirel H (2010) Image super resolution based on interpolation of wavelet
domain high frequency sub-bands and the spatial domain input image. ETRI J 32(3):390–394
Residual Network for Face Progression
and Regression

Dipali Vasant Atkale, Meenakshi Mukund Pawar,

Shabdali Charudtta Deshpande, and Dhanashree Madhukar Yadav

Abstract In computer vision application, the style transfer is a most active area,
where deep generative networks have been used to achieve desired results. The
development of adversarial networks training produces a high-quality image result in
terms of face age progression and regression that is face aging and de-aging. Inspired
by Ian Goodfellow, in this paper, we have designed the combinational network using
the residual block, convolution and transpose convolutional in CycleGAN for face
age progression and regression. Face aging is an image to image translation concept
which is used in many applications such as cross-age verification and recognition,
entertainment, in smart devices like biometric system for verification purpose etc.
The proposed architecture preserves the original identity as it is and converts young
people to old and vice versa. The network consists of residual blocks to extract
deep features. The UTKFace unpaired image dataset is used to do experiments. The
qualitative analysis of proposed methods in terms of performance metrics which
gives better results. The performance metrics calculated such as Mean Squared Error
(MSE), Root Mean Squared Error (RMSE) and Structured Similarity Index (SSIM)
to the quality of image.

Keywords Convolution · Deep learning · Discriminator · Generator · Residual

block · Style transfer · Unpaired image dataset

D. V. Atkale (B) · M. M. Pawar · S. C. Deshpande · D. M. Yadav

SVERI’s COE, Pandharpur, India
M. M. Pawar
e-mail: [email protected]
S. C. Deshpande
e-mail: [email protected]
D. M. Yadav
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 257
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_27
258 D. V. Atkale et al.

1 Introduction

The sketch of ‘Afghan Girl’ earned global recognition when it was highlighted on the
cover page of National Geographic Magazine in 1985, considering that the person in
imaginary which remains unidentified for next few years until she was found in 2002
[1]. In such thousands of similar examples of finding the lost person, the key to find
the missing person is an old photo of that person. Recently, deep learning models
are widely used. The reason behind this is that deep learning models give faster and
more precise results as well as computational complexity is less compared to existing
machine learning techniques. To work on face age which is helpful in the industry
as well as academics. Consider an example of movies, for the post-production event
many of the actors were reworked either for the beautification of face or texture is
prepared as per the expectations of the producer. More precisely, face aging and
de-aging is nothing but effects generated by applying different cosmetics on the face
that is makeup or some lighting effects.
Face aging is the process which changes according to the hormonal (gens) and
environment throughout our life. The main goal of face aging is to transform age of
the photo based on the ground truth image or conditions given by users. Face Aging
is used for many applications such as entertainment [2], finding missing persons
[3], cross-age face recognition [4], etc. Recently many researchers have worked
on face aging problems and designed the number of their models. Still, the face
aging is a challenging problem in terms of accuracy, performance measures, cross-
age verification and recognition, identity missing, etc. Due to the limitations of the
feature representation, classification and illumination variation, we are going to use
the deep generative network. In this paper, we designed a deep generative adversarial
network for face aging. The proposed network consists of an encoder and decoder
concept using residual blocks. Next section is introduced to get a brief idea about
the existing approaches for face age progression and regression.

2 Literature Survey

Traditional face aging has two approaches prototype model and physical model:
prototype model determines the average of a face age group and to construct the aged
face the difference between the average faces taken to generate an age a pattern. This
model generates the unrealistic face images the information would be lost [3–5]. The
physical model deals with facial structure and shape. Aging is done by changing the
muscles, wrinkles, size of chin expanded, hair color etc. The restrictions of this model
require a large number of paired datasets during training, and the computational rate
is too high [6, 7].
Later, for face age progression, deep neural networks were used for feature extrac-
tion. As compared to previous methods, automatically the features were extracted
which gives better aging effects using deep neural network algorithms such as Deep
Residual Network for Face Progression and Regression 259

Restricted Boltzmann Machine [8, 9], Recurrent Neural Network [2], etc. Still the
originality, age information missing and at the time of training the same person many
face images with different ages.
Afterwards, Generative Adversarial Networks (GANs) [10], works successively in
different tasks such as image synthesis [11–14], image super resolution [15], image-
image translation [16–18], image inpainting [19]. The GAN models give realistic
images with aging effects. However, GANs still have instability during training. In
conditional generative adversarial networks (cGANs) [11] the generator learns based
on conditions such as information related to the labels [20] used for the multi-domain
image to image translation and language descriptions [21] in case of text to image
synthesis, etc. But the discriminator was unable to distinguish between a real face
and generated face.
Later to overcomes on this issue the identity-preserved conditional generative
adversarial network (IPCGANs) [22] comes in existence, they were added the percep-
tual loss [23] to keep original information, this model is useful in case of the paired
image dataset. The conditional adversarial auto-encoder (CAAE) [24] first time
successfully generates face age progression and regression using unpaired image
dataset.
It has the least learning ability of generators as well discriminator due to that only
rough wrinkles were added on the face. Recently, the style transfer [25] concept is
a famous research area where the image is translated from one domain to another
domain. The cycle consistent adversarial network (Cycle GAN) [26] which consists
of pairs of generator and discriminator network to generate targeted age face image.
This CycleGAN has achieved better performance in image generation tasks. We
inspired from this network, and we have designed our model, the details of the
proposed method are given below.
The drawback of only residual blocks, it creates the artifacts, so we have designed
a combinational network.
The combinational network consists of three blocks residual, convolution and
Transpose convolution to improve the quality of the generated image; the quality is
checked by using performance metrics.

3 Proposed Approach

In this proposed approach, we have developed the combinational network in

CycleGAN [27] for face age progression and regression. The combinational network
in the CycleGAN consists of two deep generative networks, generator and discrim-
inator. The goal of the generator model is to generate fake images to confuse the
discriminator and discriminator network distinguish between generated images and
real images that is ground truth. To generate high-quality images, the PatchGAN is
used as a discriminator model. The PatchGAN helps discriminator networks to make
decisions as real or fake in terms of patches instead of the whole image.
260 D. V. Atkale et al.

Fig. 1 The overall block

diagram of the proposed
method

3.1 Block Diagram

The overall block diagram of the proposed approach shown in Fig. 1. It consists of
several blocks such as image dataset to call the database images for further processing,
the task of preprocessing unit is to remove the unwanted noise from the image, or it
enhances some features of image which is required for further image analysis and
processing. The pre-processing based on the requirements such as rescaling, reshape
etc., network has mainly two sub-parts generator and discriminator. The generator
network generates the images based on the loss functions definition, whereas the
discriminator checks whether the generated image is real or fake. Training means
there is learning relation data and attributes out the whole training dataset some
images were considered. The training block is used to train the system neurons
based on the user’s conditions, at last, a testing block which helps to check trained
module performance by applying the trained module weights to any external image
or any image from the same dataset.

3.2 Network Architecture

3.2.1 Generator Network

The generator network architecture is as shown in Fig. 2. It consists of basically

three blocks: convolutional, residual and deconvolution. The drawback of ideally
using the connection of residual block it creates an artifact [27], to avoid this issue,
we have designed the combinational network. The connection consists of an alter-
nate convolutional and residual block with stride 2 which means down-sampling
operation. Later upsampling using the transpose convolutional (deconvolution) and
residual block with stride 2. The activation function used in this network is Rectified
Residual Network for Face Progression and Regression 261

Fig. 2 The generator network architecture which consists of three blocks such as Conv2D, Residual
and Transpose Conv2D

Linear Unit (Relu). The convolutional layer extracts the features; residual block is
used for skip connection to reduce the complexity of the network. The activation
function used in the generator network is Rectified Linear Unit (Relu), which helps
to improve our results. The residual block has stride 2 for down-sampling, the stride
is nothing but the number of pixels by which the window moves after each operation
at the time of training the data.
262 D. V. Atkale et al.

Table 1 The discriminator

Layer Input Output shape
network structure
Conv2D (256, 256, 3) (128, 128, 64)
Conv2D (128, 128, 64) (64, 64, 128)
Conv2D (64, 64, 128) (32, 32, 256)
Conv2D (32, 32, 256) (16, 16, 512)
Conv2D (16, 16, 512) (16, 16, 1)

3.2.2 Discriminator Network

The discriminator network architecture is represented in Table 1. The discriminator

network consists of 4 convolution layers with stride 2. The PatchGAN concept is
used to improve the quality of image. The discriminator network gives the decision as
real or fake. The Conv2D consists of Convolution, Instance Normalisation and Leaky
Relu (slope 0.2). The Leaky Relu is the advanced activation function of Relu, the
Relu allows only the positive values, and negative values suppresses. In some cases,
information may get lost, but in Leaky-Relu has both positive as well as negative
values. The activation function applied based on the user’s requirements. The Adam
optimizer algorithm is used to updates the weights of the network during training in
an iterative manner.

3.3 Network Objectives

This section consists of summaries of the network object functions. Basically, the
proposed network is based on the generative adversarial networks (GANs), which
consists of a generator (G) and discriminator (D). This basic network is trained using
the min–max game equation as,
   
Ex∼pdata log logD(x) + Ez∼p(z) log log(1 − D(G(Z))) (1)

where z represents the noise vector, the first part of the equation gives information
related to the original dataset images are identical or not whereas the second part
tells about the generator images are real or fake, this all decisions were given by
discriminator network.
The loss functions used in the proposed approach as Adversarial loss function,
     
LGAN G, Dy, X , Y = 0.5 ∗ Ex∼Pdata(y) DY (y) + Ex∼Pdata(x) [1 − DY (G(x))] (2)

The X is related to young people, Y for old people, G generator and DY differ-
entiates between the original image and generated image based on age domain that
is young or old. Cyclic Loss function, the cyclic Loss is used to convert one domain
Residual Network for Face Progression and Regression 263

image to another domain that is in face aging. The young age group is converted to
the older age category and vice versa.
The Instance Normalization and layer normalization operations are similar to each
other. The layer normalization is used to normalize the input features, whereas the
Instance Normalization is used for normalization through each channel in training
example. Mostly, the normalization operation uses the mean and standard deviation
formulas. The Instance Normalization is expressed as,

1  1 
W H W H
xtijk − μti
ytijk =  , μti = xtilm , σti2 = (xtilm − muti )2 (3)
σti + ε
2 HW HW
l=1 m=1 l=1 m=1

4 Results and Discussion

The overall experiments have been conducted on UTKFace dataset. All the exper-
iments performed on Google Colab platform with Tesla P100-PCIE-16 GB GPU
machine. The results of the proposed approach are as shown in Fig. 3. It has divided
into two age groups as (20–30) as young people and (50–60) as old people age group
and conversion of them shown in results. The results are shown in two categories first
where we are transforming (20–30) age group to (50–60) and again reconstructed it
back to its original age group that is (20–30) this our young people face aging and
de-aging. Secondly, the (50–60) age category is converted into (20–30) and trans-
forming it back to its original age group that is (50–60) which an old to young face
image generation. In both the cases face aging and de-aging operation takes place.

4.1 Dataset

The UTKFace dataset is used, which has around 20 k face images. The age ranging
from 0 to 116 years of different races and genders. The dataset image resolution is
256 × 256 pixel. In the proposed approach, we are only concerned about the two age
categories as the young age group is (20–30) and older (50–60) remaining images
are filtered out.

4.2 Quantitative Analysis

Mean squared Error (MSE) is calculated by performing the arithmetic operation that
is the difference between the real image pixel and predicted image pixel. Lower the
value of MSE denotes higher the quality of image. The value of MSE is noted in
264 D. V. Atkale et al.

Fig. 3 The results of our combinational network, which consists of two age group (20–30) and
(50–60). The first row of the table indicates the face aging of young to old that is (20–30) age
category face image translated to (50–60) and again reconstructed it back to its original age group.
Similarly, the second row, old to young where (50–60) age group converted to (20–30) age category
and again reconstructed it back to its original form that is (50–60) age group

Table 2 MSE is calculated by using the observed value and predicted for an N number
of data points as,

1 
N
2
MSE = Yi − Ỹi (4)
N i=1
Residual Network for Face Progression and Regression 265

Table 2 The quantitative analysis of the proposed method by comparing the original image and
reconstructed image

Root Mean squared Error (RMSE) which is similar to the MSE only the root has
been taken to get the accurate results. The values of RMSE are given in Table 2. The
formula is as given below,
(1/2)
1  
N
RMSE = MSE = Yi −Yi )2 (5)
N i=1

Structured Similarity Index (SSIM) is the similarity distribution between the real
image and reconstructed image. The SSIM range is from 0 to 1. SSIM is one of
the best ways to check the quality of an image. The SSIM is represented in Table
2. The formula for calculating the SSIM between the two images that is original is
considered as A and reconstructed image as B given below,

(2μA μB + C1 )(2σAB + C2 )
SSIM (A, B) =  2   (6)
μ A + μ2 B + C1 σ 2 A + σ 2 B + C2
266 D. V. Atkale et al.

5 Conclusions

Considering the huge achievement of the generative adversarial network (GAN) in

image generation, translation, we proposed the combinational network in CycleGAN
which consists of three basic blocks that is convolutional, residual and deconvolution
blocks for face aging. The generator has residual block architecture. The performance
of both generator and discriminator networks is improved using different loss func-
tions. To train the proposed approach Google Colab platform is used. To check
the quality of original dataset image and reconstructed image by network, different
performance metrics have been calculated that is Mean squared Error (MSE), Root
Mean squared Error (RMSE), Structured Similarity Index (SSIM). For experimenta-
tion, UTKFace large image dataset is used. Our future work mainly focuses on further
enhancement in the model, so that it will be applicable to all the age group face aging
and de-aging effectively as that of two age category conversation by preserving the
original identity, and important features of the Image. As well as we will test the
same network for multiple face image dataset.

Acknowledgements The authors wish to thank the department of Electronics and Telecommu-
nication Engineering, SVERIs College of Engineering Pandharpur for the support during this
research.

References

1. The Afghan Girl: a life revealed. https://www.nationalgeographic.com/magazine/2002/04/afg

han-girl-revealed/
2. Wang W et al (2016) Recurrent face aging. Proc IEEE Conf Comput Vis Pattern Recogn
3. Kemelmacher-Shlizerman I, Suwajanakorn S, Seitz SM (2014) Illumination-aware age
progression. Proc IEEE Conf Comput Vision Pattern Recogn
4. Park U, Tong Y, Jain AK (2010) Age-invariant face recognition. IEEE Trans Pattern Anal Mach
Intell 947−954
5. Tang J et al (2017) Personalized age progression with bi-level aging dictionary learning. IEEE
Trans Pattern Anal Mach Intell 905–917
6. Lanitis A, Taylor CJ (2000) Towards automatic face identification robust to ageing variation. In:
Proceedings fourth IEEE international conference on automatic face and gesture recognition
(Cat. No. PR00580). IEEE
7. Tazoe Y et al (2012) Facial aging simulator considering geometry and patch-tiled texture. In:
ACM SIGGRAPH, Posters, 1-1
8. Nhan Duong C et al (2016) Longitudinal face modeling via temporal deep restricted Boltzmann
machines. Proc IEEE Conf Comput Vision Pattern Recogn
9. Nhan Duong C et al (2017) Temporal non-volume preserving approach to facial age-progression
and age-invariant face recognition. Proc IEEE Int Conf Comput Vision
10. Goodfellow I et al (2014) Generative adversarial nets. Adv Neural Inf Process Syst
11. Isola P et al (2017) Image-to-image translation with conditional adversarial networks. Proc
IEEE Conf Comput Vision Pattern Recogn
12. Liu Y et al (2017) Auto-painter: cartoon image generation from sketch by using conditional
generative adversarial networks. arXiv preprint arXiv:1705.01908
Residual Network for Face Progression and Regression 267

13. Makhzani A, Shlens J, Jaitly N, Goodfellow I, Frey B (2015) Adversarial autoencoders. arXiv
preprint arXiv:1511.05644
14. Mirza M, Osindero S (2014) Conditional generative adversarial nets, arXiv preprint arXiv:
1411.1784
15. Ledig C et al (2017) Photo-realistic single image super-resolution using a generative adversarial
network. Proc IEEE Conf Comput Vision Pattern Recogn
16. Huang X et al (2018) Multimodal unsupervised image-to-image translation. Proc Eur Conf
Comput Vision (ECCV)
17. Kim T et al (2017) Learning to discover cross-domain relations with generative adversarial
networks. Proc 34th Int Conf Mach Learn 70. JMLR. Org
18. Liu M-Y, Breuel T, Kautz J (2017) Unsupervised image-to-image translation networks. Adv
Neural Inf Process Syst
19. Yeh R et al (2017) Semantic image in painting with perceptual and contextual losses. arXiv
preprint arXiv:1607.07539
20. Choi Y et al (2018) Stargan: unified generative adversarial networks for multi-domain image-
to-image translation. Proc IEEE Conf Comput Vision Pattern Recogn
21. Zhang H et al (2017) Stackgan: text to photo-realistic image synthesis with stacked generative
adversarial networks. Proc IEEE Int Conf Comput Vision
22. Wang Z et al (2018) Face aging with identity-preserved conditional generative adversarial
networks. Proc IEEE Conf Comput Vision Pattern Recogn
23. Dosovitskiy A, Brox T (2016) Generating images with perceptual similarity metrics based on
deep networks. Adv Neural Inf Process Syst
24. Zhang Z, Song Y, Qi H (2017) Age progression/regression by conditional adversarial
autoencoder. Proc IEEE Conf Comput Vision Pattern Recogn
25. Zhang S et al (2019) Stylistic scene enhancement GAN: mixed stylistic enhancement generation
for 3D indoor scenes. Vis Comput 35(6–8):1157–1169
26. Zhu J-Y et al (2017) Unpaired image-to-image translation using cycle-consistent adversarial
networks. Proc IEEE Int Conf Comput Vision
27. Thengane VG et al (2018) Cycle face aging generative adversarial networks. In: 2018 IEEE
13th international conference on industrial and information systems (ICIIS). IEEE.
Design and Simulation of 2-Element,
Circular Shaped MIMO Antenna
for C-Band Frequencies

Ashish Jadhav, Nagashettappa Biradar, Husain Bhaldar, Mahesh Mathpati,

Manoj Deshmukh, and Renuka Wadekar

Abstract This paper summarizes the novel approach of designing MIMO antenna
working at dual-band frequencies with wide bandwidth. The 2-Element, circular-
shaped MIMO antenna proposed here is operating at C-band frequencies which
can be used for satellite communication, cellular communication such as 4G and
5G, Wi-MAX, and WBAN wearable applications. Two circular shaped microstrip
patches fed with the tapered feed line, are placed diagonally opposite of each other.
The designed MIMO antenna resonates at 3.1 and 6.9 GHz frequencies. The first
operating band is from 2.83 to 5.58 GHz with a maximum bandwidth of 2.75 GHz
(88.6% Bandwidth) and the second operating band is from 6.19 to 7.95 GHz with
a maximum bandwidth of 1.76 GHz (29% Bandwidth). Reflection Coefficients at
those frequencies are less than −10 dB and VSWR value is between 1 and 2 which is
necessary and sufficient condition for any antenna to radiate efficiently. The radiation
efficiency of this antenna is 58% and observed directivity is 3.4 dBi.

Keywords Antenna array · Microstrip antennas · MIMO systems · Multiple band

antennas · Mutual coupling · Multi-frequency antennas · Reflection coefficient

A. Jadhav (B) · H. Bhaldar

Department of Electronics and Communication Engineering, Bheemanna Khandre Institute of
Technology, Bhalki, affiliated to Visvesvaraya Technological University Jnana Sangama,
Belagavi, Karnataka, India
A. Jadhav · H. Bhaldar · M. Mathpati · M. Deshmukh · R. Wadekar
Department of Electronics and Telecommunication Engineering, Shri Vithal Education and
Research Institute’s College of Engineering, Pandharpur affiliated to PAH Solapur University,
Solapur, India
N. Biradar
Department of Electronics and Communication Engineering, Bheemanna Khandre Institute of
Technology, Bhalki, India
M. Mathpati
GNDCE Bidar, affiliated to Visvesvaraya Technological University Jnana Sangama, Belagavi,
Karnataka, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 269
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_28
270 A. Jadhav et al.

1 Introduction

In modern wireless communication systems, antenna plays a vital role in sending

and receiving the information. The antenna is a transducer that transmits and receives
the signals. At the transmitter, the antenna converts electrical signals into electro-
magnetic waves and sends it in the air. At the receiving end, the antenna receives the
electromagnetic signals and converts it into an electrical form which can be further
modulated and decoded. Here the question arises, at what frequency the commu-
nication is carried out? Different bands of frequencies are allocated for different
communication systems. The antenna used with that communication system must
be able to radiate (transmit/receive the signals) at the allocated frequency. Most of
the time a single antenna is used by multiple communication systems, such as in
mobile phones which are featured with Wi-Fi, Bluetooth along with voice and data
communication. In that case, the antenna used with mobile phones must be able
to resonate at different frequencies are located for Wi-Fi, Bluetooth voice, and data.
That means the antenna should exhibit multiband operating property. Also, the band-
width, which is nothing but the difference between an upper frequency and lower
frequency of communication should be large enough to encompass all the ranges
supported by that communication technique.

2 Literature Review

Details of work carried out by other researchers along with methodology and
outcomes are compared with the work proposed here; given in Table 1.
In [1], the design of Wideband GNSS Antennas has been reported for the Global
Positioning System (GPS) antennas. RT6010 substrate with εr = 10.2 and loss tan δ
= 0.0023 is used as substrate material. Bandwidth enhancement is done up to 38%
(1.16–1.61 GHz).
In [2], Integrated Duplexing and Multi-Band Filtering are implemented for
triple-band frequency (2.46–2.56, 3.45–3.56, and 5.01–5.26 GHz). The reflection
coefficients for all those three bands of frequencies are between −20 to −30 dB.
Design of Vertical Metallic Strip Pairs Feeding Structure is described in [3], in
which Rogers RT6010 and Rogers RO4003 are used as substrate materials and it is
designed to radiate at Five frequencies (2.4, 3.5, 4.1, 4.8, 5.2 GHz).
The novel triple-band Rectifier antenna reported in [4], in which FR-4 substrate
with εr = 4.4, δ = 0.02, and a thickness of 1.6 mm is used. It is designed for
Energy Harvesting Applications in GSM-900, GSM-1800, and Wi-Fi-2450 having
dimension 54 mm × 42 mm.
To improve the bandwidth of the antenna, Defected Ground Structure is used in
[5], due to which bandwidth enhancement is achieved up to 27.2, 21.2, and 17.5%.
In [6], Triple-Band Monopolar Antenna is designed using Rogers Duroid 5880LZ
εr = 2.0 for Triple Band: 2.51, 3.56, and 4.62 GHz. A circular patch with a diameter
Table 1 Comparison of this proposed work with existing results
References Technique Substrate Operating Overall S-parameters Bandwidth/% Reference VSWR Applications
and εr frequency band(s) dimensions Bandwidth impedance mentioned
(mm)
[1] Wideband 1.27 mm 1.16–1.61 GHz 112 × 112 Return loss 1.16–1.61 GHz 25–50  NA Global
GNSS RT6010 × 30 mm better than (38%) Navigation
antenna substrate 10 dB Satellite
with εr = Systems
10.2 (GNSS)
antennas
[2] Integrated εr = 2.55 2.46–2.56, 45 × 50 × −20 to − 5.7 and 10.4% NA NA NA
duplexing 3.45–3.56, and 0.8 mm 30 dB
filtering 5.01–5.26 GHz
[3] Metallic Rogers Five frequencies 30 mm × Up to −20 dB 150, 360, 150, 50, NA NA WLAN,
strip pairs RT6010 and (2.4, 3.5, 4.1, 4.8, 30 mm × 55 MHz WiMAX, and
feeding Rogers 5.2 GHz) 9.813 mm 5G
structure RO4003
[4] Triple-band FR-4 with εr 0.9, 1.8, and 54 mm × −20 to − NA 50  NA Energy
rectifier = 4.4, δ = 2.45 GHz 42 mm 30 dB Harvesting
0.02 and Applications
thickness of GSM-900,
1.6 mm GSM-1800,
and
Design and Simulation of 2-Element, Circular Shaped MIMO Antenna …

Wi-Fi-2450
[5] Defected FR4 with 7.317–9.58, 144 mm × Below − 27.2, 21.2, 17.5% NA NA X, Ku, and
ground thickness 25.45–31.4, and 144 mm 10 dB Ka bands
structure 0.3 mm 35.37–42.167 GHz
(continued)
271
Table 1 (continued)
272

References Technique Substrate Operating Overall S-parameters Bandwidth/% Reference VSWR Applications
and εr frequency band(s) dimensions Bandwidth impedance mentioned
(mm)
[6] Triple-band Rogers Triple Band: 2.51, Circular Below − 22.80% 50  1–2 Not
monopolar Duroid 3.56, and 4.62 GHz patch 10 dB mentioned
antenna 5880LZ εr diameter of
= 2.0 150 mm
[7] Parasitic 1.0 mm 960,1860, and 118 mm × Return loss 25.6% for NA 3:1 Wi-Fi, RFID,
shorted strip thick FR-4 2590 MHz 42 mm less than − 850–1100 MHz, and LTE
substrate εr 6 dB 13.3% for
= 4.4 1750–2000 MHz
and 22.7% for
2190–2750 MHz
This Work 2 elements, FR-4 3.1 and 6.9 GHz 55 mm × −15 dB 88.6% 56  1.03 4G and 5G,
circular substrate 80 mm Bandwidth (2.83 Wi-MAX
shaped εr = 4.4 to 5.58 GHz with and WBAN
MIMO a maximum wearable
antenna bandwidth of applications
2.75 GHz), and
29% Bandwidth
(for 6.19 to
7.95 GHz with a
maximum
bandwidth of
1.76 GHz)
A. Jadhav et al.
Design and Simulation of 2-Element, Circular Shaped MIMO Antenna … 273

of 150 mm is used to get the reflection coefficient Below −10 dB and bandwidth of
22.8%.
In [7], Parasitic Shorted Strip is designed for 960, 1860, and 2590 MHz using
1.0 mm thick FR-4 substrate εr = 4.4 with dimensions of 118 mm × 42 mm. This
antenna has a return loss of less than −6 dB and bandwidth 25.6% for 850–1100 MHz,
13.3% for 1750–2000 MHz, and 22.7% for 2190–2750 MHz which can be used for
Wi-Fi, RFID, and LTE.

3 Methodology

The methodology adopted to design such a system is a microstrip patch antenna.

There are plenty of shapes in which microstrip patch antenna can be designed. One
of which is a circular shaped microstrip patch antenna as shown in Fig. 1a.
In a circular microstrip patch antenna, the ground plane and circular patch are
separated by dielectric substrate material. In this proposed antenna, the dielectric
substrate material used is flame retardant-4 (FR4) having a dielectric constant of 4.4
and thickness 1.6 mm. As shown in Fig. 1a, the design of a circular-shaped patch
antenna is simulated with an EM solver. Dimensions of the fabricated antenna are
mentioned in Fig. 1b. The feedline of the antenna is made tapered to widen the
bandwidth of the MIMO Antenna.
Defective ground structure (DGS) is the method in which creating defects in the
ground plane of the antenna. Sometimes the ground plane is partially removed to
improve the performance of the antenna. The performance improvement in terms
of improving various parameters such as bandwidth polarization gain and so forth.
Which method is adopted in the work proposed here and 40% partial ground concept
is used to enhance the bandwidth of antenna as shown in Fig. 2.

Fig. 1 Design of 2-element, circular shaped MIMO antenna. a Design of 2-elements, b parameters
274 A. Jadhav et al.

Fig. 2 40% Partial ground plane of MIMO antenna

4 Simulation Results and Discussions

The proposed antenna is designed with EM solver software and the results obtained
for various parameters are plotted. Many antenna parameters can be identified to
improve. But here the focus is mainly on scattering parameter, bandwidth, VSWR,
and reference impedance. Scattering parameters are the performance metric of any
antenna. For the MIMO antenna using 2 Patches, scattering parameters will be S11,
S12, S21, and S22.
S11 is an indication of the reflection coefficient of patch-1. At 3.1 GHz frequency,
the value of the observed S11 parameter is −11.04 dB and At 6.9 GHz frequency,
the value of the observed S11 parameter is −13.36 dB as shown in Fig. 3 and Table
2. S22 is an indication of the reflection coefficient of patch-2. At 3.1 GHz frequency,
the value of the observed S22 parameter is −11.03 dB and At 6.9 GHz frequency,
the value of the observed S22 parameter is −13.62 dB as shown in Fig. 4 and Table
2.
Bandwidth is the range of frequencies at which the communication can be carried
out. As far as antenna design is concerned, is the necessary and sufficient condition
for the antenna to radiate effectively is its reflection Coefficient should be less than
−10 dB and its VSWR value should be between 1 and 2. The antenna bandwidth is
calculated below −10 dB line. The difference between higher frequency (FH) and the
lower frequency (FL) is referred to as bandwidth. In this, the bandwidth is expressed
Design and Simulation of 2-Element, Circular Shaped MIMO Antenna … 275

Fig. 3 S11 Parameter of patch 1

Table 2 S11 and S22

MIMO Freq. (GHz) S-Parameter (dB)
parameter of patch 1 and 2 at
Simulation result
3.1 and 6.9 GHz
Patch-1 3.1 −11.04
6.9 −13.36
Patch-2 3.1 −11.03
6.9 −13.62

in percentage and it can be obtained by dividing this frequency difference by Centre

frequency (Fc).
For 1st band of patch 1, bandwidth is measured from 2.83 to 5.58 GHz. So, the
frequency difference obtained here is 2.75 GHz and percentage band with approaches
to 88.6% with Fc = 3.1 GHz. For 2nd band of patch 1, bandwidth is measured
from 6.19 to 7.95 GHz. So, the frequency difference obtained here is 1.76 GHz and
percentage band with approaches to 29% with Fc = 6.9 GHz action in Fig. 5 and
Table 3.
For 1st band of patch 2, bandwidth is measured from 2.69 to 5.64 GHz. So, the
frequency difference obtained here is 2.95 GHz and percentage band with approaches
276 A. Jadhav et al.

Fig. 4 S22 Parameter of patch 2

Fig. 5 Bandwidth of patch 1

Design and Simulation of 2-Element, Circular Shaped MIMO Antenna … 277

Table 3 Bandwidth of patch 1 and 2 at 3.1 and 6.9 GHz

MIMO Freq. (GHz) Bandwidth
Patch-1 3.1 2.75 GHz from 2.83 to 5.58 GHz
(% Bandwidth = 88.6% with Fc = 3.1 GHz)
6.9 1.76 GHz from 6.19 to 7.95 GHz
(% Bandwidth = 29% with Fc = 6.9 GHz)
Patch-2 3.1 2.95 GHz from 2.69 to 5.64 GHz
(% Bandwidth = 95% with Fc = 3.1 GHz)
6.9 1.75 GHz from 6.24 to 7.99 GHz
(% Bandwidth = 25.4% with Fc = 6.9 GHz)

to 95% with Fc = 3.1 GHz. For 2nd band of patch 2, bandwidth is measured from 6.24
to 7.99 GHz. So, the frequency difference obtained here is 1.75 GHz and percentage
band with approaches to 25.4% with Fc = 6.9 GHz action in Fig. 6 and Table 3.
Voltage standing wave ratio (VSWR) is a term that is used to represent the reflec-
tion of the signal into the system. It also indicates the amount of impedance mismatch.
If the impedance of the transmitter circuit does not match with the impedance of the
antenna then most of the transmitted waves are reflected. And those reflected waves
will set up standing waves. VSWR is also called a standing wave ratio. Ideally, the
value of voltage standing wave ratio is 1 and practically it may vary from 1 to 2.
For the proposed antenna patch 1, the VSWR value for 3.1 GHz is 1.78 and for
6.1 GHz it is 1.54. For Patch 2, he VSWR value for 3.1 GHz is 1.77 and for 6.1 GHz
it is 1.52 as shown in Figs. 7 and 8; Table 4.

Fig. 6 Bandwidth of patch 2

278 A. Jadhav et al.

Fig. 7 VSWR of patch 1

Fig. 8 VSWR of patch 2

Design and Simulation of 2-Element, Circular Shaped MIMO Antenna … 279

Table 4 VSWR of patch 1

MIMO Freq. (GHz) VSWR
and 2 at 3.1 and 6.9 GHz
Patch-1 3.1 1.78
VSWR1 6.9 1.54
Patch-2 3.1 1.77
VSWR2 6.9 1.52

5 Conclusion

This article presents a Simulation of 2 elements, circular-shaped MIMO antenna.

This antenna is operating at C band frequency such as 2.69 to 5.64 GHz and 6.19
to 7.95 GHz. The defected ground structure (DGS) technique is used to widen the
bandwidth and enhance the polarization. Tapered of the feedline is done to obtain dual
resonating bands at 3.1 and 6.9 GHz. By using this method the reflection coefficient
can be made as low as −10 dB and the voltage standing wave ratio can be maintained
between 1 and 2. This novel approach of antenna design can be used in wideband
communication systems where applications need to be operated on a wide range of
frequencies such as satellite communication, cellular communication such as 4G and
5G, Wi-MAX and WBAN.

References

1. Tamjid F, Foroughian F, Thomas CM, Ghahreamani A, Kazemi R, Fathy AE (2020) Toward

high-performance wideband GNSS antennas-design tradeoffs and development of wideband
feed network structure. IEEE Trans Antennas Propag 68(8):5796–5806
2. Xie Y, Chen F-C, Qian J-F (2020) Design of integrated duplexing and multi-band filtering slot
antennas. IEEE Access 8:126119–126126. https://doi.org/10.1109/ACCESS.2020.3006831
3. Afifi AI, Abdel-Rahman AB, Abd El-Hameed AS, Allam A, Ahmed SM (2020) Small frequency
ratio multi-band dielectric resonator antenna utilizing vertical metallic strip pairs feeding
structure. IEEE Access 8:112840–112845. https://doi.org/10.1109/ACCESS.2020.3002789
4. Tafekirt H, Pelegri-Sebastia J, Bouajaj A, Brite MR (2020) A sensitive triple-band rectifier
for energy harvesting applications. IEEE Access 8:73659–73664. https://doi.org/10.1109/ACC
ESS.2020.2986797
5. Fahad AK, Ruan C, Ali SAKM, Nazir R, Haq TU, Ullah S, He W (2020) Triband ultrathin polar-
ization converter for X/Ku/Ka-band microwave transmission. IEEE Microwave Wirel Compon
Lett 30(4):351–354
6. Paramayudha K, Chen SJ, Kaufmann T, Withayachumnankul W, Fumeaux C (2020) Triple-
band reconfigurable low-profile monopolar antenna with independent tunability. IEEE Open J
Antennas Propag 1:47–56. https://doi.org/10.1109/OJAP.2020.2977662
7. Chen A, Sun M, Zhang Z, Fu X (2020) Planar monopole antenna with a parasitic shorted strip for
multistandard handheld terminals. IEEE Access 8:51647–51652. https://doi.org/10.1109/ACC
ESS.2020.2979483
Attribute Inspection of Product Using
Image Processing

Anup S. Vibhute, Reshma R. Deshmukh, P. S. Valte, B. D. Gaikwad,

and Shrikant Pawar

Abstract Automation is a crucial thing about an industry which manufactures

product within the mass quantity. After manufacturing product; to form the deci-
sion of rejecting or accepting is taken by measuring quality parameters. To test
quality parameters like dimensions and features of manufactured product inspection
is mostly done manually in manufacturing industries. Manual assessment is time-
consuming, costly, sometimes inaccurate and manual assessment for elegant shapes
is incredibly difficult. To resolve these problems, control and quality management
of the commercial product is feasible by the use of image processing techniques.

Keywords DIP · Feature extraction · Gray-scale · Image analysis · Image

classification · Image quality · Image capture · Image denoising · Image
enhancement · Image edge detection · Image filtering · Image processing · Image
recognition · Image segmentation

A. S. Vibhute · R. R. Deshmukh (B) · P. S. Valte

Department of Electronics and Telecommunication, SVERI’s College of Engineering, PAH
Solapur University, Pandharpur, Solapur, India
A. S. Vibhute
e-mail: [email protected]
P. S. Valte
e-mail: [email protected]
B. D. Gaikwad
Department of Mechanical Engineering, SVERI’s College of Engineering, PAH Solapur
University, Pandharpur, Solapur, India
e-mail: [email protected]
S. Pawar
Top Gear Transmission, Satara, (M.S.), India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 281
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_29
282 A. S. Vibhute et al.

1 Introduction

There are two forms of visual-based defect detection; the manual inspection and
also the automated one [1]. In many industries, the quality of products is tested by
manual review with the assistance of gauges, and if it doesn’t fit on gauge properly,
then the job is taken into account to be faulty. Also in some industries, the standard
testing is completed by a human eye wherein human observes the form and size of
the output products are starting of the product line. But in practice, just some pieces
are taken and verified for dimension and orientation of the shapes. This manner of
manual testing could be a tedious process and at risk of human errors which reduce
the standard of the products.
This approach has been utilized in the industry before the existence of automated
visual inspection. The activities during this manual inspection are to search the
defect, recognize the fault and make the sole decision. Thus, the training for quality
inspection is crucial to boost the skill of examination and to attenuate errors during
the manufacturing process. They have to look out a way to identify the task with
defects than to form clear decision either to accept or to reject or to remodel the
defected part. Moreover, human inspectors are slow and have become ineffective
after completing the task that required an extended time. They’re either full of fatigue
or sickness or human weaknesses. Hence, they have frequent rest to want care of a
high-performance level. So we want to travel for an automatic visual inspection [2].
Today there is no field of technical endeavour neglected without the impact of
Digital Image Processing. Digital Image Processing may well be a way for an auto-
matic visual inspection [3]. We proposed here the attribute inspection technique that
uses a camera which captures a digital image of every job. The captured image is
preprocessed, filtered. Then the size or attributes are extracted, measured using edge
detection [2, 4] and segmentation techniques [5, 6]. Finally, the output is compared
with a reference or actual dimension in drawing using feature matching. The image
processing task is to look out the faulty piece and to form a decision whether to accept
or reject manufactured product using classification [7]. This increases the speed and
accuracy and avoids human errors which are common in quality testing and also
increases productivity. Quality testing using DIP performs acceptable range.
Some researchers illustrated previous research works which are studied to beat
the restrictions of subjective Evaluation in the visual quality inspection by a human
inspector. They developed an automatic procedure replacement by using computer
vision and image processing technologies to automate the method. These attempts
are to figure out the defects in manufacturing by using digital image processing [8].
Here, we proposed a system which helps to keep up the count of excellent products
furthermore as faulty products produced within the entire day [9] because it identifies
the defect and makes a choice to accept or reject the manufactured product on the
premise of attribute inspection [10].
Attribute Inspection of Product Using Image Processing 283

2 Methodology

See Fig. 1.

2.1 Image Acquisition

Images of the desired product are acquired through digital cameras. Photos are usually
obtained by one or more cameras placed at the position under inspection. The func-
tions of the cameras are typically fixed. In most cases, industrial automation systems
are designed to inspect only known objects at fixed positions [3]. The picture needs
to be adequately illuminated and arranged to facilitate good image acquisition. The
data flow starts from the image acquisition module by connecting the digital camera
with laptop pc (Fig. 2). The image acquisition module captures and transfers Image
into the computer for processing [8], in our paper input image shown in Fig. 3.

2.2 Image Preprocessing

Once images are acquired, they’re filtered to get rid of background or unwanted
reflections from the illumination system. Image restoration can also be applied to
boost image quality by correcting geometric distortions introduced by the acquisition
system i.e. camera. The acquired input images are preprocessed by using multiple

Fig. 1 Block diagram of attribute inspection using image processing

284 A. S. Vibhute et al.

Fig. 2 Camera setup with an illumination source

Fig. 3 Acquired input image

operations like grayscale conversion, threshold effect and noisy objects elimination
which are present in the pictures, as shown in Fig. 4.
Attribute Inspection of Product Using Image Processing 285

Fig. 4 Pre-processing output

2.2.1 Gray Scale Conversion

Sequence of input images is acquired from a digital camera, and then it is converted
into a grayscale image.

2.2.2 Threshold Effect

After Grayscale conversion threshold effect is applied with the particular value. As
a result, we get Binary Image from the Gray Scale Image as shown in figure.

2.2.3 Denoised Image Using Image Filtering

Image filtering is used to:

a. Remove noise
b. Sharpen contrast
c. Highlight contours
d. Detect edges.
The Median Filter could also be a non-linear digital filtering technique accustomed
remove noise from an image which we are going to get Denoised Image which
enhances the output of next step edge extraction [11].
286 A. S. Vibhute et al.

2.3 Edge Detection

An Edge in an exceeding picture could be a significant change within the brightness;

it’s a discontinuity in image intensity. Edge detection is one reasonably feature extrac-
tion. Edge detection identifies the points in an exceedingly digital image at which the
image brightness changes sharply, have discontinuities. The points at which image
brightness changes smartly are organized into a gaggle—this group of curved line
segments called as edges. Image segmentation is completed using various edge detec-
tion techniques like Sobel, Prewitt, Roberts, Canny, Log [6, 12]. Here we are using
the Sobel Edge Detection technique for edge extraction [13].
Sobel Edge Detection: It works by calculating the gradient of image intensity at each
pixel within the Image and then emphasizes regions of high spatial frequency that
correspond to edges. The convolution masks [14] of Sobel operator is as shown in
Fig. 3, which are accustomed, obtain the gradient magnitude of the Image from the
initial. The output of a Sobel edge detector is shown in Fig. 7.


The gradient magnitude is given by |G| = Gx 2 + Gy 2 .
Typically, [15] an approximate magnitude is computed using:

|G| = |Gx| + |Gy|

2.4 Segmentation

Image segmentation may be a necessary technique used for image analysis. It’s the
tactic of partitioning a digital image into multiple segments (sets of pixels, also
called image objects). This step tries to partition the image into regions of interest
that correspond to part or whole objects inside the scene [12]. The varied segmen-
tation techniques used are EM algorithm, OSTU algorithm and Genetic Algorithm
[6]. Threshold selection is employed in OTSU algorithm. Compared with all other
segmentation methods, the Otsu method is one of the only successful ways for image
Thresholding because of its simple calculation. Thresholding creates binary images
from grey-level ones by turning all pixels below some threshold to zero, and each
one pixel this threshold to a minimum of one [16]. If g(x, y) might be a threshold
Attribute Inspection of Product Using Image Processing 287

version of f(x, y) at some global threshold T, it’s often defined as [17],

(x, y) = 1; (x, y) ≥ T
(x, y) = 0; other wise

The segmentation output is shown in Fig. 7.

2.5 Feature Measurement

We have used Euclidean distance for feature measurement like inner diameter, outer
diameter and have matched. The Euclidean distance gives straight line distance
between two points. It’s appropriate once we have continuous numerical variables
and need to reflect absolute distances. This distance takes into consideration every
variable and doesn’t remove redundancies. Moreover, this distance doesn’t scale-
invariant, so generally must scale previously to use the gap. The Euclidean distance
[10] between two weight vectors provides a measure of similarity between the corre-
sponding images Imgref and Imgtest, as shown in Fig. 10. The measurement output
is shown in Fig. 9. The formula for which is,
 2
Euclidean Distance (Imgref, Imgtest) = (Imgref, Imgtest)
   
(2|A ∩ B|) (2|A ∩ B|)
Matching = Mismatching = 1 −
|A| + |B| |A| + |B|

2.6 Decision Making

In case of visual inspection, the system has to decide if the result of manufacturing
meets the quality standards, by matching [18] the computed features with a known
model. If the product satisfies the matching criterion, it is considered to be accepted
or else rejected.

2.7 Graphical User Interface

A graphical program (GUI) may be a visual interface to a plan. An honest GUI can
make schedules more comfortable to use by providing them with a uniform appear-
ance and with intuitive controls like push buttons, list boxes, sliders, menus, then
forth. A graphical-based language allows the user to figure directly with graphics.
288 A. S. Vibhute et al.

The developed GUI window gives a matching result for the acceptance and rejection
process [19]. From the all MATLAB functions that are presented and implemented,
all of them combine in the GUI screen to be the last final model for our proposed
system [20] as shown in Fig. 11.

3 Results and Discussion

Initially, we have captured 100 images of the required job (Ring Gear). The photos
are taken through all angles in such a way that all dimensions are covered. While
acquiring images from different angles, care is being carried by maintaining the same
distance from every angle (top view, side view and 45° angle). We formed three
databases for three different perspectives. Then we have taken new input images
one by one (around 100 test images) and compared with the reference database for
getting a final result of acceptance or rejection of input test job image.
After preprocessing dataset with the help of image processing techniques such
as Thresholding, denoising using the median filter, we need to do edge detection.
Initially, used canny method and results are as shown in Fig. 5. So, to get the more
precise output, we have gone for Sobel edge detection whose products are better as
shown in Fig. 6, than the Canny edge detection.

Fig. 5 Canny edge detection output

Attribute Inspection of Product Using Image Processing 289

Fig. 6 Sobel edge detection output

Because of simplicity, we have used Otsu segmentation. The segmentation result

of Otsu segmentation algorithm is stable or profitable as shown in Figs. 7, 8 and 9.
Figure 10 has shown either the test image matched with the reference database or
not (Fig. 11).
GUI used to give a straightforward approach for checking image processing
results. We have offered training database, an input image, preprocessing, edge detec-
tion, selecting parameter to be measured, segmentation and matching result buttons
in GUI (Fig. 12).
Previous researchers have given defect detection only. But our proposed system
adds to earlier research works, defect detection with decision making. The decision
is taken in terms of acceptance or rejection of the product. Out of 100 test images
of the product, almost 92 are accepted, and 8 are rejected due to excess positive
tolerance. As our product is industrial Ring gear mostly circles, we found. So, if
diameter exceeds than the considered positive tolerance, product becomes faulty.
Thus, our proposed system gives 92% accurate results (Table 1).

4 Conclusion

In this study, the image processing technique is used for automatic inspection and
internal quality control. Although lot of research carried by different researcher
doing research in images processing, there’s scope to use image processing tech-
niques for internal control of commercial product. The image processing techniques
290 A. S. Vibhute et al.

Fig. 7 Segmentation output

Fig. 8 Small circle detection

Attribute Inspection of Product Using Image Processing 291

Fig. 9 Attribute measurement

Fig. 10 Final comparison output

makes inspection automatic and fast. Also improves quality and production rate of
an industry. Algorithm is proposed for real time quality monitoring of manufactured
product. This proposed system can replace manual inspection of commercial product.
Result will indicate product is appropriate or not. Using this automatic inspection
system cost of inspection is reduced as we require only 1 time installation cost. Also
accuracy of inspection will increase.
292 A. S. Vibhute et al.

Fig. 11 GUI results for various parameters

Fig. 12 Final result in excel format

Table 1 Results of acceptance or rejection

No. of test images Accepted with (+3 Accepted with Exact images Rejected products
taken pixel) positive (−3 pixel)
tolerance negative tolerance
100 45 42 5 8
Attribute Inspection of Product Using Image Processing 293

References

1. Rahman NNSA, Saad NM, Abdullah AR, Ahmat N (2019) A review of vision based defect
detection using image processing techniques for beverage manufacturing industry. J Teknol
81(3):33–47
2. Canny J (1986) A computational approach to edge detection. IEEE Trans Pattern Anal Mach
Intell 6:679–698
3. Malamas EN, Petrakis EG, Zervakis M, Petit L, Legat J-D (2003) A survey on industrial vision
systems, applications and tools. Image Vis Comput 21(2):171–188
4. Threshold HNN (2016) A survey on relevant edge detection technique for medical image
processing. Eur J Biomed 3(7):317–327
5. Fouzia MT, Nirmala K (2010) A literature survey on various methods used for metal defects
detection using image segmentation. Evaluation 8
6. Ramadevi Y, Sridevi T, Poornima B, Kalyani B (2010) Segmentation and object recognition
using edge detection techniques. Int J Comput Sci Inf Technol (IJCSIT) 2(6):153–161
7. Lu D, Weng Q (2007) A survey of image classification methods and techniques for improving
classification performance. Int J Remote Sens 28(5):823–870
8. Peansupap V, Laofor C (2010) Digital image processing for evaluating defect level in visual
quality inspection. In: W078-special track 18th CIB world building congress, Salford, May
2010, p 110
9. Watpade AB, Amrutkar MS, Bagrecha NY, Vaidya A (2014) Qcuip: quality control using image
processing. Int J Eng Res Appl 4:15–18
10. Chavan HL, Shinde SA (2016) Defective product detection using image processing. Int J Sci
Res (IJSR) 5(6)
11. Shinde MD (2016) To detect and identify the defects of industrial pipe. Int J Sci Res (IJSR)
5(7):1561–1564
12. Swarnalakshmi R (2014) A survey on edge detection techniques using different types of digital
images. Int J Comput Sci Mob Comput 3(7):694–699
13. Ramesh TP, Bisht Y (2014) Detection and classification of metal defects using digital image
processing. Int J Adv Res Sci Eng 3
14. Senthilkumaran N, Rajesh R (2009) Image segmentation—a survey of soft computing
approaches. In: 2009 international conference on advances in recent technologies in commu-
nication and computing. IEEE, pp 844–846
15. Image filtering median filtering. https://www.cs.auckland.ac.nz/courses/compsci373s1c/Patric
esLectures/Image%20Filtering_2up.pdf
16. Cai H, Yang Z, Cao X, Xia W, Xu X (2014) A new iterative triclass thresholding technique in
image segmentation. IEEE Trans Image Process 23(3):1038–1046
17. Vala HJ, Baxi A (2013) A review on Otsu image segmentation algorithm. Int J Adv Res Comput
Eng Technol (IJARCET) 2(2):387–389
18. Karami E, Prasad S, Shehata M (2017) Image matching using SIFT, SURF, BRIEF and ORB:
performance comparison for distorted images. arXiv preprint arXiv:1710.02726
19. Kiran R, Amarendra H, Lingappa S (2018) Vision system in quality control automation. In:
MATEC web of conferences, vol 144. EDP Sciences, p 03008
20. Humod AH, Leman Z, Samin IR, Ashrafi N (2015) Improving productivity and quality in
manufacturing by applying computer vision systems. Image Process 10(5):1–9
A Proposed Method for Audio
Steganography Using Digital Information
Security

Pratik Kurzekar and Shrinivas Darshane

Abstract In the current era of digital technology, the information security is the
challenging task. For the secrete communication information hiding is an essential
element. The current information steganography system uses objects like audio,
image and video. The audio steganography is the technique that convey hidden
message by modifying an audio signal in an unnoticeable manner. It is a technique for
the hiding secret message in the host audio signal. The original audio message before
steganography and after encoding message is having uniform characteristics. The
embedding secrete audio message in the original audio file is a more challenging and
difficult task. This paper presents the comprehensive survey of audio steganography
techniques for information security. The experiment was tested using proposed LSB
technique for audio steganography. This paper extended towards quality measure of
steganography message. The quality of audio steganography measures using energy
score, Mean square error, Peak signal to noise ratio. From this experiment the quality
of audio steganography is observed as 92.759% for MSE and 94.971% for PSNR
technique. Audio information hiding is the one of the robust and dynamic ways of
protecting the privacy and secretes communication.

Keywords Information · Steganography · Quality · LSB · PSNR · MSE ·

Quality · Energy · HAS

1 Introduction

In this era of rising technologies, digital communication has become an integral and
significant part of everyone’s life. In the rapid development of digital communica-
tion the information security becomes an important concern. The methods and algo-
rithms available for digital data security uses a cryptographic primitive for secure
data transmission and secrete communication. With the advent of the technology,
people started to use private communication for sharing and transmission using the

P. Kurzekar (B) · S. Darshane

SVERI’s College of Engineering, Pandharpur, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 295
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_30
296 P. Kurzekar and S. Darshane

technological approach. As a result, securing these secret messages became a critical

issue for everyone. Cryptography and Steganography are two security methods for
secure data communication and data confidentiality [1].
In the cryptography method it ciphers the secret message, so that it cannot be read-
able by everyone without part of that communication, while steganography hides the
secret message into an original file so that it cannot be seen by eavesdroppers [2]. For
the third party user an encrypted message would right away imply a secret commu-
nication. The hidden message is not able to draw any attention and therefore would
not raise suspicions that a secret communication. Due to this reason steganography is
often regarded as a surreptitious method for transmitting receptive information into
total secrecy across public channels [3].
The audio steganography is a type of digital steganography, which hides digital
information into the digital audio media [4]. Human Auditory System (HAS) cannot
observe slight variation of high frequency based audible message so; audio steganog-
raphy has a great choice for secret communication [5]. The hiding speech in the audio
file algorithms could be embedded with the bit rate that is a considerable portion of
the host audio bit rate, up to 150 kbps [6]. The robustness of the audio steganography
method is referred to as the capability of the data detector to extract the embedded
message after common signal processing manipulations [7].
This paper presents a comprehensive survey of audio steganography techniques
for information hiding. The LSB technique was used for the implementation of audio
steganography. The experiment tested in time domain and frequency domain. The
rest of the paper is structured in seven sections. Audio steganography is described
in Sect. 2. Related work is explained in Sect. 3. The techniques of audio steganog-
raphy are illustrated in Sect. 4. Section 5 deals with proposed techniques for audio
steganography. Experimental analysis is described in Sect. 6. Section 7 deals with
conclusion followed by references.

2 Audio Steganography

Audio steganography is the ability and science of thrashing digital information such
as text messages, documents, and binary files into audio files. The primary message
is known as the carrier signal or message and the secondary message is known as to
the payload signal or message. Characteristically normal audio file and carrier file
are same and not recognized in tapping communication technology [8]. The general
steganography technique is shown in Fig. 1.
Steganography is achieved by means of three varieties of techniques: injection,
substitution, and generation. The insertion technique is embedded in the data to
cover in the insignificant part of the carrier file, which is normally unseen by oper-
ating systems and application software. The substitution technique substitutes the
insignificant bits in the original carrier message with the bits of the data to secrete.
Insignificant bits are those bits that can be modified without destroying the eminence
or destroying the reliability of the carrier message. This technique takes advantage of
A Proposed Method for Audio Steganography Using Digital … 297

Fig. 1 General steganography system

the limited abilities of the human auditory system; which cannot identify two sounds
that are slightly not alike. The generation technique examines the data to cover and
produces out of them a new set of data. It is a dynamic method of creating a carrier
message based on the information enclosed in the data to cover [9].

3 Review of Literature

Researchers proposed a three layered architectural model for audio steganography

which defines the replacement of Least Significant Bit. Before storing the cover
message into the last layer, the private message to be transmitted and it is passed
through the two layers. The stego message is transmitted over the network towards
the receiver side and secret message is obtained by performing the operations in
reverse order. The main objective of the paper is to keep the security and robustness
of the carrier message. They define the different parameters such as capacity, trans-
parency and robustness for the implementation of three layered architecture. This
experimental analysis proposed by Muhammad Asad et. al. gave the signal to noise
ratio of 54.78 dB compared with conventional LSB method having 51.12 dB SNR
[10].
Audio steganography system implemented by Lovey Rane gives improved secu-
rity. For this, they used a dual layer randomization approach. In this system, the
first layer is obtained by choosing randomly the byte number or samples. Here, an
additional layer of security is provided by selecting the bit position randomly at
which embedding is done in selected samples. By using this system, transparency
and robustness of the technique is improved [11].
Researchers observe a new method which is similar to the well known LSB
method. Due to less robustness and more susceptibility, LSB method is not desired.
In this method, two bits are used for protecting the message by increasing data hiding
298 P. Kurzekar and S. Darshane

capacity. A filter is added to restrict the changes in the stego file. Obtained stego files
are used to generate unique keys. The filtered file and the generated key sent to the
receiver. The key is used to extract the correct message at the receiver side [12].
Sridevi proposed a useful method of audio steganography by customizing the
LSB algorithm and a strong encryption key with enhanced security is suggested.
Enhanced Audio Steganography (EAS) is a combination of audio steganography
and cryptography. EAS works in two steps: it uses the most effective encryption
algorithm in the first level and in the second level it uses a modified LSB (Least
Significant Bit) algorithm to enclose the message into audio [13].
Submission technique is also a good choice for audio steganography. Message bits
are placed into multiple and higher LSB layer values using genetic algorithms ensuing
in enhancement of robustness. The robustness of the system should be increased
against intruders which try to exhibit the secret message and also some involuntary
attacks like noise etc. [14].
Mane suggested a method known as Least Significant Bit (LSB) method. In this
method, consecutive least significant bits are replaced with private message bits from
each sample of cover audio. LSB method is very simple but less powerful. This paper
differentiated the spectrum of original audio before embedding and audio signal after
embedding [15].
A new 4th bit rate LSB audio steganography method is a proposed new approach in
the current era. This method minimizes the embedding distortion of the host audio. In
this method, message bits are fixed into the 4th LSB layer. It leads in high robustness
against noise addition. As compared to standard LSB method, the perceptual quality
of audio is more in proposed method [16].
The substitution method is having some limitations for audio steganography. The
main problem of this technique is that it is less significant against attack. There are
two types of attack: One, it tries to extract the private message and other tries to
destroy it. As in the standard LSB method, the secret message is stored into the least
significant bit, so this method is more susceptible to attack. Therefore for security
purposes, the message is stored into a bit other than LSB. If a message is stored into
deeper bits, the system will become more powerful. But the main disadvantage is
that when the message is stored into MSB, the host audio signal gets altered. So,
by using an intelligent algorithm this problem is solved where the message bits are
embedded into MSB and other bits are altered to decrease an error. The message is
stored into multiple MSB to make a system more robust and high capacity [17].
Divya proposed a method where multiple LSB bits are used for hiding a text in
an audio signal using steganography and cryptography is used for security purposes.
For LSB audio steganography, the maximum number of bits is altered from 16 bit
audio samples. They use two novel approaches for substitution technique of audio
steganography improving capacity of cover audio for storing additional data. In
this technique, the message bits are stored into 35–70% compared to standard LSB
technique which uses 4 LSBs for data storing [18].
The researcher proposed a Genetic algorithm is also a good choice for audio
steganography. They studied the various audio steganography techniques using
A Proposed Method for Audio Steganography Using Digital … 299

genetic algorithms and LSB approaches. They tried some approaches which help
in audio steganography. In these techniques, hidden messages are written in such a
way that only the sender and corresponding receiver are able to see the message [19].

4 Techniques of Audio Steganography

Researchers currently turn towards hiding the high quality secret message in audio
files. The abundance of audio messages makes them eligible to convey secret informa-
tion. Many researchers started to explore how the audio signals and audio properties
can be exploited in the era of information security [20]. Several approaches were
considered, the most robust and significant ones are Least Significant Bit [21], Echo
hiding [22], Hiding in Silence Interval [23], Phase Coding [24], Amplitude Coding
[25], Spread Spectrum [26], and Discrete Wave Transform [27].

4.1 LSB Technique

Fundamentally, the LSB technique depends on implanting every bit from the data
to hide into the rightmost bits of every audio sample of the carrier audio message.
The LSB technique proved as advancement towards a HAS unable to understand
the slight variation of audio sampling frequency towards the high frequency region
of the audible spectrum. The LSB technique allows high embedding rate without
impairing the quality of the audio file. This technique is robust and dynamic for
audio steganography.
The technique uses the very fact that the majority of the knowledge in a very
sample in any audio file is contained within the MSBs instead of LSBs. In the LSB
coding approach the slandered data broadcasting rate is 1 kbps per 1 kHz. In some
implementations of LSB coding, the two smallest significant bits of a sample are
substituted with two message bits. This implementation increases the amount of data
that will be determined but also enlarges the amount of resultant noise in the audio
file as well. The representation of LSB coding technique is shown in Fig. 2.

4.2 Echo Hiding Technique

In this echo hiding technique, the secret data are embedded into the audio signals as a
short acoustic echo. In fact, an echo is a reproduction of sound, however, received by
the listener some time after the original sound. The echo is perceptible; its amplitude
must be reduced and undetectable. In order to hide data, bits whose values are 0, it
is characterized by an echo overdue 1 ms; bits whose values are 0 are represented by
an echo delayed 2 ms.
300 P. Kurzekar and S. Darshane

Fig. 2 Graphical representation of LSB coding technique

Fig. 3 Graphical representation of echo hiding technique

In this technique, the original signal is divided into chunks before the encoding
process. Once the encoding process is finished, the blocks are concatenated back
together to create the final signal [28]. The complete block diagram of echo hiding
technique is shown in Fig. 3.

4.3 Amplitude Coding

Amplitude coding technique conceals secret data in the magnitude speech spectrum
while not distorting the carrier audio signal. It is based on finding a safe spectral area
in the signal whose magnitude speech spectrum is below a certain value. In addition,
the carrier locations are preferred based on how much they can badly affect the audio
signal [25].
A Proposed Method for Audio Steganography Using Digital … 301

Fig. 4 Synthesized spread

spectrum information
encoded in original message

4.4 Spread Spectrum

Spread Spectrum technique scatters the secret data over the frequency spectrum of
the audio file using a precise code independent of the original signal. Fundamentally,
secret data are multiplied by a code known to the corresponding level only, and then
implanted in the carrier audio message. In this technique the data is generated by m-
sequences code known as sender and receiver for the secret communication [29]. To
control stego speech distortion, [30, 31] have proposed an embedding method where
splitted data is secreted under a frequency cover. The spread spectrum is combined to
phase changing to increase the strength of the transmitted data against additive noise
and allows easy detection of the fixed data. In this method, a reliable hiding capacity
of 3 bps was attained. The graphical representation of spread spectrum information
encoded in the original message is described in Fig. 4.

4.5 Discrete Wave Transform

Discrete Wave Transform technique private messages are encoded in the smallest
significant bits of the wavelet coefficients of the audio signals. Often, private data
is chosen to be secreted in the wavelet coefficients and not in silent sections of the
audio signal so as to promote the imperceptibility of the audio file [27].

5 Proposed Approach

From the enriched literature observed the limitation of available techniques and
their procedure. The human ear is highly sensitive and can regularly notice even the
slightest bit of noise introduced into a sound file. The main limitation associated with
parity coding is not much closure making introduced noise inaudible. The limitation
of phase coding towards data transmission rate because of this technique message is
encoded in the first segment of signal only. Phase coding technique recommended
only when the small amount of data considered for a steganography approach. Least
significant bit (LSB) coding is the robust way to encode messages in a digital audio
302 P. Kurzekar and S. Darshane

file. Substituting the least significant bit to each frequency point with a binary message
allows for a large amount of data to be encoded [21]. In the available data hiding
methods in enriched literature the proposed method for embedding secret messages
within audio file, LSB is the simple and robust method for inserting messages in audio
signal towards noise free environment. It embeds secret message-bits in a subset of
the LSB levels of the audio message.
The following steps are:
a. Receive Audio file
b. Convert the file into bit Pattern
c. Each character of secrete message convert into the bit pattern
d. Replaces the LSB bit from audio with LSB bit from character in the message.
The proposed system is to provide a good, robust and dynamic method for infor-
mation hiding in audio from hackers and sent to the destination for secret commu-
nication. The advantages of the proposed system does not change the size of the
original audio file even after encoding and also suitable for all audio formats.

6 Experimental Analysis

For these experiments a real example was tested to illustrate how the proposed
steganography technique works. It involved using the various steps of the algorithm
to envelop a secret text message into a 16-bit WAV carrier audio file. For this exper-
iment we selected the different 10 secret messages. The secret data to hide is a text
message shown in the experiment that says “we will kill you”. The private message
is preprocessed and converted into a binary form. Ten audio samples are randomly
selected for this experiment. The chunks obtained in implementation step one are
stored into the three LSBs of the audio samples selected in step two. The final output
is the carrier WAV audio file now carrying the secret data. The graphical represen-
tation of original audio message “you are very nice” is shown in Fig. 5. Figure 6
represents the audio file with encoded secrete message. The spectrogram of audio
file with encoded message is shown in Fig. 7. Figure 8 represents the spectrogram
of original audio message after secrete message extracted.

Fig. 5 Original audio signal message for the “you are very nice”
A Proposed Method for Audio Steganography Using Digital … 303

Fig. 6 The audio file with encoded message “we will kill you”

Fig. 7 The spectrogram of combine original signal and secrete message

6.1 Stegno Audio Quality Measurement

Steganography audio quality measurement replaces the listener panel with a

computational algorithm, thus facilitating automated real-time quality measurement.
Indeed, for the purpose of real-time quality examining and organization on a network-
wide scale, objective speech quality measurement is the only viable option. Objective
measurement methods aim to convey quality examination that is highly correlated
with those obtained from subjective listening experiments. In the objective quality
measure mean square error (MSE) and peak signal-to-noise ratio (PSNR) technique
was used.
304 P. Kurzekar and S. Darshane

Fig. 8 Original signal after extracted secrete message

6.2 Mean Square Error (MSE)

The mean square error (MSE) of associate reckoner calculates the common of the
squares of the errors, that is, the variation between the estimator and what is estimated.
MSE is a risk function, equivalent to the predictable value of the squared error loss
or quadratic loss. The distinction happens attributable to uncertainty or as a result
of the reckoner does not account for data that might manufacture a lot of correct
estimation of speech synthesis [32].

6.3 Peak Signal to Noise Ratio (PSNR)

Peak Signal to Noise Ratio, often abbreviated PSNR, is the ratio between the most
probable influence of a signal and the power of corrupting noise that affects the
quality of its illustration. PSNR is usually articulated in terms of the logarithmic
decibel scale. PSNR is most ordinarily accustomed live the standard of reconstruction
of signal and image. The signal in this case is the original data, and the noise is
the error introduced by synthesis [33]. The PSNR and MSE method was used for
quality measure of steganography audio based on proposed LSB technique. Table 1
represents the MSE and PSNR values for LSB technique.
A Proposed Method for Audio Steganography Using Digital … 305

Table 1 MSE and PSNR results for audio steganography using proposed LSB technique
Sr. No. Original audio signal Stegno audio signal MSE PSNR
1 S001 N001 5.23 1.26
2 S002 N002 8.9 7.56
3 S003 N003 7.61 1.29
4 S004 N004 5.32 3.24
5 S005 N005 7.61 7.32
6 S006 N006 9.1 9.78
7 S007 N007 8.06 11.23
8 S008 N008 5.32 1.29
9 S009 N009 8.06 3.24
10 S0010 N0010 7.2 4.08
Average 7.241 5.029
Quality (100 − average) 92.759 94.971

7 Conclusion

In this experiment, we have proposed a LSB robust technique for imperceptible

audio steganography. This system is useful to provide a good, efficient method for
data secrecy from hackers and send it to the destination safely. This proposed tech-
nique will not change the size of the file even after encoding and also suitable for
any type of audio file format. Thus we conclude that audio steganography techniques
can be used for a number of purposes other than covert communication and informa-
tion tracking. The superiority of steganography messages plays an important role in
secret communication. The experiment extended towards quality measures of audio
steganography. The quality measure experiment tested using MSE and PSNR tech-
niques. The observed quality is extracted as 92.759% from MSE and 94.971% from
PSNR. The authors recommended the PSNR technique is robust and dynamic for
steganography audio quality measure.

References

1. Wayner P (2009) Disappearing cryptography: information hiding: steganography & water-

marking, 3rd edn. Morgan Kaufmann Publishers
2. Petitcolas FAP, Anderson RJ, Kuhn MG (1999) Information hiding—a survey. Proc IEEE Spec
Issue Prot Multimed Content 87(7):1062–1078
3. Bender W, Gruhl D, Morimoto N, Lu A (1996) Techniques for data hiding. IBM Syst J 35(3–
4):313–336
4. Dong X, Bocko M, Ignjatovic Z (2004) Data hiding via phase manipulation of audio signals.
In: IEEE international conference on acoustics, speech, and signal processing (ICASSP), vol
5, pp 377–380
5. Kandel ER, Schwartz JH, Jessell TM (2000) Principles of neural science, 4th edn. McGraw-Hill
306 P. Kurzekar and S. Darshane

6. Lee YK, Chen LH (2000) High capacity image steganographic model. IEEE Proc Vis Image
Signal Process 288–294
7. Alvaro M, Guillermo S, Gadiel S (2005) Is image steganography natural? IEEE Trans Image
Process 14
8. Cvejic N, Seppben T (2002) Increasing the capacity of LSB-based audio steganography. Finland
9. Johnson NF, Jajodia S (1998) Exploring steganography: seeing the unseen. Comput J 31(2):26–
34
10. Asad M, Gilani J, Khalid A (2012) Three layered model for audio steganography. In:
International conference on emerging technologies (ICET)
11. Rana L, Banerjee S (2013) Dual layer randomization in audio steganography using random
byte position encoding. Int J Eng Innov Technol 2(8)
12. Gandhi K, Garg G (2012) Modified LSB audio steganography approach. Int J Emerg Technol
Adv Eng 3(6):158–161
13. Sridevi R, Damodaram A, Narasimham S (2009) Efficient method of audio steganography by
modified LSB algorithm and strong encryption key with enhanced security. J Theor Appl Inf
Technol 771–778
14. Bankar Priyanka R., Katariya Vrushabh R, Patil Komal K (2012) Audio steganography using
LSB. Int J Electron Commun Soft Comput Sci Eng 90–92
15. Mane A, Galshetwar G, Jeyakumar A (2012) Data hiding technique: audio steganography using
LSB technique. Int J Eng Res Appl 2(4):1123–1125
16. Gadicha AB (2011) Audio wave steganography. Int J Soft Comput Eng (IJSCE) 1:174–177
17. Zamani M et al (2009) A secure audio steganography approach. In: International conference
for internet technology and secured transactions
18. Divya SS, Ram Mohan Reddy M (2012) Hiding text in audio using multiple LSB steganography
and provide security using cryptography. Int J Sci Technol Res 1:68–70
19. Nehru G, Dhar P (2012) A detailed look of audio steganography techniques using LSB and
genetic algorithm approach. Int J Comput Sci (IJCSI) 9:402–406
20. Djebbar F, Ayad B, Abed-Meraim K, Hamam H (2011) A view on latest audio steganography.
In: 7th IEEE international conference on innovations in information technology
21. Gopalan K (2003) Audio steganography using bit modification. Proc Int Conf Multimed 1:629–
632
22. Gruhl D, Bender W (1996) Echo hiding. In: Proceeding of information hiding workshop, pp
295–315
23. Shirali-Shahreza S, Shirali-Shahreza M (2008) Steganography in silence intervals of speech.
In: Proceedings of the fourth IEEE international conference on intelligent information hiding
and multimedia signal, pp 605–607
24. Qi Y-C, Ye L, Liu C (2009) Wavelet domain audio steganalysis for multiplicative embedding
model. In: Proceedings of the 2009 international conference on wavelet analysis and pattern
recognition
25. Djebbar F, Ayad B, Abed-Meraim K, Habib H (2012) Unified phase and magnitude speech
spectra data hiding algorithm. J Secur Commun Netw
26. Khan K (1984) Cryptology and the origins of spread spectrum. IEEE Spectr 21:70–80
27. Cvejic N, Seppanen T (2002) A wavelet domain LSB insertion algorithm for high capacity
audio steganography. In: Proceedings 10th IEEE digital signal processing workshop and 2nd
signal processing education workshop, p 5355
28. Gopalan K, Wenndt S (2004) Audio steganography for covert data transmission by impercep-
tible tone insertion. In: WOC 2004, Banff, 8–10 July 2004
29. Czerwinski S, Fromm R, Hodes T. Digital music distribution and audio watermarking. https://
reference.kfupm.edu.sa/content/d/i/digital_music_distribution_and_audio_wat_1045219.pdf
30. Queirolo F. Steganography in images. Final communications report. https://eric.purpletree.org/
file/Steganography%20In%20Images.pdf
31. Cox IJ, Kalker T, Pakura G, Scheel M (2005) Information transmission and steganography, vol
3710. Springer, pp 15–29
A Proposed Method for Audio Steganography Using Digital … 307

32. Lehmann EL, Casella G (1998) Theory of point estimation, 2nd edn. Springer, New York.
ISBN 0-387-98502-6. MR 1639875
33. Huynh-Thu Q, Ghanbari M (2008) Scope of validity of PSNR in image/video quality
assessment. Electron Lett 44(13):800. https://doi.org/10.1049/el:20080522
Spine Diseases Detection Using SVM

Jyoti M. Waykule and V. R. Udupi

Abstract This paper addresses the use of HOG (Histogram of Oriented Gradi-
ents) and SVM (Support Vector Machine) for boundary regression for biomedical
image segmentation and comparing the output with image processing segmenta-
tion techniques. Here we will use MATLAB for computation of HOG and SVM.
MATLAB image processing toolbox will be extensively used for reading, processing,
visualizing and saving the images.

Keywords HOG (Histogram Oriented Gradients) · SVM (support vector

regression) · NFS: neural foramina stenosis · CT: computer tomography · MRI:
magnetic resonance image

1 Introduction

The nerves are foraminate by the nervous radicals of the neural foramina columns
and can become a strain that can lead to painful numbness, weakening of the arm,
wakening and equilibrium of the leg. Neural foramina stenosis (NFS) is the restricting
of the openings of any vertebral in the spine called foramina. The risk of neural
foramina stenosis rises according to the age factor as we age discs in the spine lose
height, starting to dry excessively because of NFS, over 80% of peoples are infected.
For segmentation, HOG methods are used to detect the right area and examination,
MRI and CT imaging are used to demonstrate the stenosis for the diagnosis and
treatment manual. The segmentation system of the physician should be used because
of size, shape, variance attribute. Segmentation is the automatic boundary detection
process.

J. M. Waykule (B)
E & C Department, Sanjay Ghodawat University, Atigre, Kolhapur, Maharashtra, India
e-mail: [email protected]
V. R. Udupi
Department of E & C Engineering, Gogte Institute of Technology, Belgaum, Karnataka, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 309
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_31
310 J. M. Waykule and V. R. Udupi

2 Literature Review

A number of very motivating and pioneering image segmentation algorithms have

been developed in recent years, and these algorithms can be loosely grouped into
five key categories: thresholding, matching models, clustering, detecting edges and
increasing regions. In many applications, these algorithms have been efficient, but
none of them are universally applicable to all images, and different algorithms are
typically not equally appropriate for a specific application. Histograms of Directed
Gradients for Human Detection were used by Dalal and Triggs to detect human
HOG descriptors. The authors used the SVM (Support Vector Machine) classi-
fier in the paper to identify descriptors obtained by HOG. For SVM preparation,
the authors computed descriptors from R-HOG and C-HOG and checked the effi-
ciency of both descriptors [1]. This paper outlines an approach to edge computation.
The effectiveness of the method depends on the concept of Comprehensive set of
targets for the measurement of edge points [2]. The issue of multiple-input multiple-
output (MIMO) frequency non-selective channel estimation is discussed by Sanchez-
Fernandez ‘SVM Multi regression for Nonlinear Channel Estimation in Multiple-
Input Multiple-Output Systems.’ For multiple variable regression estimation, authors
create a new method [3].
This paper present the novel approach to develop a new segmentation algorithm
considering multiple anatomic planes (sagittal and axial plane), multiple anatomic
structures (disc, vertebral), and multimodality approach (MRI, CT). This approach
formulates the segmentation task innovatively as a boundary regression problem and
is fulfilled by the advancement of sparse kernel machines, regression segmentation
multi-dimensional support vector regression (MSVR). This segmentation approach
was comprehensively tested on images from 113 clinical subjects with getting high
dice similarity index (DSI) 0.912 and a low boundary distance (BD) 0.928 mm [4].
Intervertebral disc degeneration is an age-associated condition related to chronic
back pain, while its consequences are responsible for over 90% of spine surgical
procedures worked toward 2-D semiautomatic segmentation of both normal and
degenerated lumbar intervertebral discs from T2-weighted mid sagittal MR images
of the spine. This task is challenged by partial volume effects and overlapping gray-
level values between neighboring tissue classes. The best overall performance, when
considering the tradeoff between segmentation accuracy and time efficiency, was
accomplished by the atlas-robust-fuzzy c-means approach, which combines prior
anatomical knowledge by means of a rigidly registered probabilistic disc atlas with
fuzzy clustering techniques incorporating smoothness constraints [5].
Spine Diseases Detection Using SVM 311

2.1 HOG (Histogram Oriented Gradients)

HOG is a feature identifier used in image processing and computer vision for object
detection. A feature descriptor is an image or image patch interpretation that simpli-
fies the image by extracting useful information and discarding foreign information.
In most cases, a feature descriptor converts an image of size width * height * 3 (chan-
nels) to a feature vector/array of length n. In the case of the HOG feature descriptor,
the input image is a size 64 * 128 * 3, and the output feature vector is of length 3780.
HOG descriptor can be calculated for other sizes, but the above numbers presented
in the original paper so one can easily understand the concept with one concrete
example. The histogram of directions of gradients (oriented gradients) is used as a
feature in the HOG feature descriptor. The gradients (x and y derivatives) of a image
are useful because the magnitude of the gradients is intense around the edges and
corners (regions of rapid intensity changes) and we know that the edges and corners
provide plenty of key information about the shape of the object than the flat regions.
Gradient Calculation: Mostly in HOG feature extractor, the input image is divided
into small parts called cells, normally 8 × 8 pixels, as shown in Fig. 1. The gradient
in both vertical directions is then estimated for each pixel in the cell. Simple [−1, 0,
1] and [−1, 0, 1] T gradient filter is applied to the pixel value f(x, y) to obtain both
fx (x, y) and fy (x, y) which is defined as the gradients in both x and y directions have
been determined, the gradient Magnitude m(x, y) and the gradient direction (x, y)
could be determined as:

fx (x, y) = f(x + 1, y) − f(x − 1, y) (1)

Fig. 1 Block size and cells size in HOG

312 J. M. Waykule and V. R. Udupi

fy (x, y) = f(x, y + 1) − f(x, y − 1) (2)


M(x, y) = f2x (x, y) + f2y (x, y) (3)

fy (x, y)
θ(x, y) = arctan (4)
fx (x, y)

In the case of color image, the gradients at each pixel would be calculated for all
color components separately and the gradient with the largest magnitude would be
selected to be used in next steps of feature extraction.
Histogram Formation: The first steps is to produce an orientation histogram for each
cell in the image based on the values of m(x, y) and (x, y) derived for the pixels
in the cell. The interval of [0, π] is evenly divided into the number of orientations
for the development of orientation bins to be considered in the calculation of the
histogram. The original HOG feature extraction introduced by Dalal and Triggs [1]
has a value 9 bins to be expected for a number of directions since it has shown better
result in the detection [6] for each pixel inside the cell as reference in Paper [1].
Block Normalization: The key element in the extraction of the HOG function is
the process of standardization across blocks. Once the gradient for each pixel of
the cell is determined, the neighboring cells are grouped together to create a more
spatial block. In general, blocks are characterized as overlapping numbers of adjacent
cells. Usually, each block consists of four cells, i.e. collection of 2 × 2 neighboring
cells [1]. Based on the normalization scheme adopted for this step, the four cells
vectors within a block are combined to produce a normalization factor. Considering
normalized vector function as show in paper [1], we are going to have a small constant
and L2-Hys is L2-norm followed by clipping and restricting the maximum values
and then renormalizing. Dalal and Triggs have shown that L2-Hys, L2-norm and
L1-sqrt perform equally well for normalization, while L1-norm decreases efficiency
by 5% [1, 6].

2.2 Block Diagram

In order to measure the HOG descriptor, we first need to measure the horizontal and
vertical gradients; however, we want to calculate the gradient histogram (Fig. 2). The
gradient of the image is determined in both the “x” and “y” directions (Fig. 3c). And
in Fig. 3d. Then angle Fig. 3f and magnitude Fig. 3e of the gradients is calculates
using following two equations.
Pre-processing: Here the preprocessed step considers the conversation of Di.com
image into uint8 image by using the MATLAB function.
Spine Diseases Detection Using SVM 313

Fig. 2 HOG descriptor block diagram

Fig. 3 a Original image, b normalized image, c X-gradient, d y-gradient, e magnitude image,

f angles image
314 J. M. Waykule and V. R. Udupi

Compute Gradients with its Magnitude and Angles: The 2D filter [−1 0 1] is used
for calculation in x direction and 2D filter transpose of [−1 0 1] is used to calculation
of gradient in y-direction. The filter then calculates the horizontal difference between
the next and previous pixels and stores the result as “dx” in the current pixel; similarly
calculates the vertical difference between the next and previous pixels and stores the
result as “dy” in the current pixel.√Calculated the magnitude and angles by using the
following formula. Magnitude = dxˆ2 + dyˆ2 and Angle = tan − 1(dy/dx).
The angle and magnitude of gradients is then used for HOG calculations. The gradi-
ents of the picture are sensitive to complete illumination. If you darken the image by
dividing all pixel values by 2, the gradient magnitude will change by half, and the
histogram values will change by half.
Histogram of Gradients (HOG Descriptor): In the HOG function, the input image
is divided into small parts called cells, typically 8 × 8 pixels. The gradient in both
horizontal and vertical directions is then determined for each pixel inside the cell as
shown in the earlier point. Now the values of angles and magnitudes in each cell [64
of angles and 64 of magnitudes] are to be shaped into a histogram of 9 bins, making
9 values for each cell. The x-axis bins of the histogram are the angles of the gradients
measured above. The angles are divided into 9 values scale from zero to 180° in 20°
measures. Thus the 9 bins are [0, 20, 40, 60, 80, 100, 120, 140, 160]. Then the angles
between the bin values are divided into the left and the right parts. The left part is the
allocation to the closest bin value less than the angle and the right part is the nearest
bin value greater than the angle [180 is not a part of this bins because largest will
is similar to zero]. The magnitude of the section is the difference between the angle
and the bins. Now the gradient magnitude corresponding to the bins are inserted and
the histogram is created (Fig. 3c). The matrix is now transformed to a single array of
cell-wise histogram values and this is going to send the final HOG function vector
of matrix size M × N × 9, where M and N are the number of vertical and horizontal
cells respectively.
Boundary Representation: Image editing software (GNU Image Manipulation) is
used and the boundary of the image taken for segmentation training is traced. The
analyzed boundary then divides the image into two regions, the inside boundary
region and the outside boundary region. The area outside the boundary is filled
with Black and the area inside the boundary is filled with White. And the image is
converted to Binary image. White pixels are the border pixels, and Black pixels are
the non-boundary pixels. The generated labels are used for the formation of the SVM
classification.
Support Vector Machine (SVM Classifier): Support Vector Machine (SVM) is
supervised machine learning technique. SVM is binary classifier used for classifying
data into two classes. The SVM classifier is generally a hyper plane that divides the
data into two sections. The bins separated by the HOG are divided into two sections
Spine Diseases Detection Using SVM 315

[boundary region and non-border region]. In SVM Training the first step is to choose
a nonlinear function (kernel function) that maps the input to a higher dimensional
space. This option will also be influenced by understanding of the problem domain.
In the absence of such knowledge, one could prefer by using polynomials, Gaussians
or other basic functions. The dimensionality of the configured space may be random
high.
Classify Boundary and Non-boundary Pixels: SVM model is trained with HOG
feature descriptor and gradient magnitude characteristics are given. SVM model
offers a set of labels for border and non-border cells. SVM labels are stored in an
array. The array is “N * M” where ‘N’ is Horizontal cell numbers and ‘M’ is vertical
picture cell count. The labels array is now iterated with the length “N” segments of
“M.” if you find boundary pixels in the array then corresponding element in matrix
represent by one. The matrix is then transformed to a binary image.
Boundary on Original Input Image: The Boolean image obtained from the SVM
output label array is resized to the original image size. The pixels in the original
image equivalent to those in the boundary Binary image are labeled with red color
pixels.

3 Result and Discussion

Spine image segmentation methodology explained in this paper is implemented using

MATLAB software. Figure 4 shows the results obtained by segmentation of Spine-
MRI image using different methods. Figure 4a shows segmentation results of HOG-
SVM based image segmentation method explained in this paper. The segmentation
results are then compared with standard edge detection methods available readily
with MATLAB. Prewitt image segmentation (Fig. 4b) gives similar results to our
method, but our method removes edges of the image which are out of spine section
(advantage of machine learning) and hence is better. Sobel operator (Fig. 4c) produces
same results as of Prewitt, with somewhat less fine parts inclusion inside spine image.
Canny segmentation (Fig. 4d) method includes all fine details but also adds high
frequency noise.
The number of 1s (represented by red color dots in the output image Figs. 4, 5,
6 and 7) for each method of segmentation as shown in Table 1. According to this
method, HOG SVM obtains the greater number of 1s as shown in Table 1 and best
performance result in as compared with Sobel, Prewitt and Canny edge detection as
shown in Figs. 4a–d, 5a–d, 6a–d and 7a–d. The Canny approach reveals the smaller
number of 1s but gives the blurred output picture. HOG SVM method shows the best
performance (Tables 2 and 3).
316 J. M. Waykule and V. R. Udupi

Fig. 4 Spine MRI image segmentation using HOG SVM, Prewitt, Sobel and Canny resp

4 Conclusion

HOG-SVM-based image segmentation provides better results than most regular

image segmentation techniques as shown in Figs. 4, 5, 6 and 7. The approach allows
a wide group of SVM training images to be more correctly classified as opposed
to traditional methods of segmentation of images, so that no new class of images
can be worked on. SVM provides a more precise and noiseless regression of limits
to the phase and is thus better than the traditional methods of image segmentation.
The grading scheme for lumber spinal foramina stenosis has been almost perfectly
inter-observed and intra-spective justification for analysis.
Spine Diseases Detection Using SVM 317

Fig. 5 Spine MRI image segmentation using HOG SVM, Prewitt, Sobel and Canny resp

Fig. 6 Spine MRI image segmentation using HOG SVM, Prewitt, Sobel and Canny resp
318 J. M. Waykule and V. R. Udupi

Fig. 6 (continued)

Fig. 7 Spine MRI image segmentation using HOG SVM, Prewitt, Sobel and Canny resp

Table 1 Shows the execution time segmentation methods

Sr. No. Test image size Count of ones Count of ones Count of ones Count of ones
320 × 320 by HOG SVM by Prewitt by Sobel by Canny
1 Figure 4 1146 1134 1134 1099
2 Figure 5 855 839 839 827
3 Figure 6 861 838 838 830
4 Figure 7 951 938 938 926
Spine Diseases Detection Using SVM 319

Table 2 Shows the graphical representation of Table 1

Table 3 Shows the execution time for different segmentation method

Sr. No. Test image size Execution by Execution by Execution by Execution by
320 × 320 HOG-SVM (s) Prewitt (s) Sobel (s) Canny (s)
1 Figure 4 73.376430 70.337076 72.745216 71.329501
2 Figure 5 72.550456 68.643676 66.772196 68.175625
3 Figure 6 69.175966 67.321464 66.875997 71.342935
4 Figure 7 69.974786 69.929276 67.543475 69.660774

References

1. Dalal N, Triggs B (2005) Histograms of oriented gradients for human detection. In: 2005 proceed-
ings of the IEEE computer society conference on computer vision and pattern recognition, vol
1, pp 886–893
2. Canny J (1986) A computational approach to edge detection. IEEE Trans Pattern Anal Mach
Intell PAMI-8(6):679–698
3. Sánchez-Fernández M (2004) Member, IEEE, SVM multiregression for nonlinear channel
estimation in multiple-input multiple-output systems, vol 52, no 8
4. Wang Z, Zhen X, Tay KY, Osman S, Romano W, Li S (2015) Regression segmentation for M3
spinal images. IEEE Trans Med Imaging 34(8)
5. Michopoulou SK, Costaridou L, Panagiotopoulos E, Speller R, Panayiotakis G, Todd-Pokropek
A (2008) Atlas-based segmentation of degenerated lumbar intervertebral discs from MR images
of the spine
6. Sakthivel K, Nallusamy R, Kavitha C (2014) Color image segmentation using SVM pixel
classification image. World Acad Sci Eng Technol Int J Comput Inf Eng 8(10)
BCI Integrated Wheelchair Controlled
via Eye Blinks and Brain Waves

Sredha Prem, Jeswin Wilson, Shelby Mathew Varghese, and M. Pradeep

Abstract People who suffer from complete or partial paralysis often require the
help of a second individual to move around from one place to another even with
automated traveling mechanisms. This dependence on another person takes a toll on
the individual. Thus, a framework is proposed wherein eye blinks and brain waves
are used to decide the decisions made by the individual. This requires no extra phys-
ical exertion. These signals will be collected by an Electroencephalogram Headset
(Neurosky Mindwave Mobile) and will then be passed to an android application
which will then transfer necessary data to the Arduino, which will use these data to
run the device. Such a framework can be executed on mechanized wheelchairs to
uphold the impaired and help in their transition of becoming independent.

Keywords Brain-computer interface · Eye blink detection · Brainwaves ·

Electroencephalogram · Neurosky Mindwave Mobile · Arduino · Wheelchair
control

1 Introduction

Wheelchairs give versatility to individuals who experience the ill effects of halfway or
complete loss of motion and have been around for an exceptionally significant time-
frame. Early Chinese models of wheeled goods utilized work carts to move both
substantial items and impaired individuals. After some time, the plan and useful-
ness of wheelchairs have improved with the development of the main lightweight,
folding, and mass-created wheelchair in 1933 by Harry Jennings and Herbert Everest
filling in as the presentation of the advanced wheelchair. Besides, the wheelchair has
developed with the headway of innovation. Notwithstanding the conventional labor
driven wheelchairs, electric and battery-controlled wheelchairs are currently like-
wise accessible. Additionally, wheelchairs can in any case be made a stride further
with the intensity of brainwaves.

S. Prem (B) · J. Wilson · S. M. Varghese · M. Pradeep

Department of Mechanical Engineering, Mar Baselios College of Engineering and Technology,
Trivandrum, Kerala, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 321
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_32
322 S. Prem et al.

Electroencephalography is the center of brainwave innovation. It is a method

of recording and checking mind exercises with the utilization of anodes appended
to an individual’s head. The anodes record action through electrical driving forces
that the cerebrum’s neurons radiate to speak with the remainder of our body. Up
until the most recent couple of years, electroencephalography has just been acces-
sible in clinics and other clinical foundations where experts utilize costly EEG gear
that can cost a great many dollars. These are generally inaccessible for standard
purchasers and engineers [1–6]. Notwithstanding, the previous years have hardly
seen the progression and improvement of more moderate EEG-related items, for
example, Neurosky’s Mindwave Mobile—conceivably the most reasonable EEG-
sensor and Brain-Computer Interface accessible to engineers in the market today.
The broad accessibility of moderate EEG sensors has made ways for the boundless
prospects in the field of brainwave innovation.
The joining of brainwave innovation into advanced wheelchairs will give
specialists and patients the same new choices intending to motor-related impairments.

2 Literature Survey

Keerthana et al. [7] of VIT built up a method of turning electrical machines on

and off utilizing brainwaves. The venture utilizes a Neurosky Mindwave Mobile
for brainwave detecting and transmission. Aside from the turning on and off of
the electrical apparatus, this task likewise tried the Mindwave Mobile’s exactness
regarding interpreting brainwave information into usable ones. A progression of
tests for changing forces of Beta and Gamma waves was made and the Mindwave
Mobile figured out how to pass every one of them. This venture didn’t utilize different
highlights of the Mindwave Mobile, for example, squint location. Constraints of
this undertaking incorporate the maximum transmission scope of Bluetooth and
the restricted scope of brainwave-related information that the Mindwave Mobile
can detect. Manuel Adrian et al. [8] at De La Salle University Philippines built up
a short informing framework utilizing Emotiv’s super-costly $800 EPOC headset.
Emotiv’s EPOC headset can distinguish explicit facial developments, giving the
designer admittance to a wide scope of potential controls. The greatest constraint
of this task is its expense. While Emotiv’s EPOC headset can distinguish a bigger
scope of brainwaves and facial developments, it is amazingly costly and along these
lines blocked off to most understudy engineers. Dominguez [9] built up a Bluetooth
distant controlled RC vehicle for his mid-year study’s last undertaking. His venture
utilizes a Bluetooth module to send and get signs to and from the Android application
that fills in as the distant regulator of the RC vehicle. Achkar et al. [10] built up a
cell phone-controlled wheelchair with an auto-development highlight that permits
you to spare a predefined way that the wheelchair will take. Impediments of this
undertaking incorporate the absence of an obstruction detecting include that would
make its auto-development more compelling.
BCI Integrated Wheelchair Controlled via Eye Blinks and Brain … 323

3 Definition of Terms

3.1 Brain-Computer Interface (BCI)

A brain-computer interface is an immediate correspondence pathway between an

improved or wired cerebrum and an external gadget [11]. A brain-computer interface
allows users to directly control an external device or gadget using only brain signals.
The Neurosky Mindwave Mobile [12] is an example of a BCI system.

3.2 Electroencephalography (EEG)

Electroencephalography or EEG is a method of recording the brain’s electrical

activity. EEG is done by placing electrodes on the subject’s scalp. It measures voltage
fluctuations within the neurons of the brain. EEG was confined to medical institu-
tions, but the development of cheaper, more consumer-friendly EEG devices have
put it in the mainstream market.

3.3 Electrooculography

Electrooculography or EOG is a technique for recording movements of the eye. EOG

uses the electrical changes due to muscle-related movements in the ocular region to
detect events such as blinks.

3.4 Bluetooth

Bluetooth is a global wireless communication standard for connected devices wire-

lessly over a certain distance. At present, there are about 8.2 billion Bluetooth devices
in use worldwide. Bluetooth devices, depending on the class, can transmit up to
100 m. However, the most common transmission distance for Bluetooth devices is
10 m.
324 S. Prem et al.

3.5 Microcontroller

A microcontroller is a very small computer on a single integrated circuit. It has a

processor, a memory module, and programmable input and output ports. Microcon-
trollers are typically used in embedded systems that have specific and dedicated
functions.

3.6 HC-06 Bluetooth Module

The HC-06 Bluetooth module is one of the most common Bluetooth modules used
by hobbyists and professionals alike. It is a serial port protocol Bluetooth module
that only acts as a slave. HC-06 modules can transmit up to 10 m.

3.7 Arduino Uno

The Arduino Uno is one of the most, if not the most, popular microcontrollers around.
It is based on the ATmega238p. The Uno has fourteen digital input/output pins with
six of those doubling as PWM outputs.

3.8 L298N Motor Driver

The L298N Motor Driver enables the control of DC motors by amplifying the low-
current signal from the Arduino into a higher-current signal suitable for motor control.

3.9 Neurosky Mindwave Mobile

Neurosky’s Mindwave Mobile is an Electroencephalography or EEG headset that

measures and transmits brainwave data via Bluetooth. It can monitor attention and
meditation levels as well as to detect blinks.
BCI Integrated Wheelchair Controlled via Eye Blinks and Brain … 325

3.10 Attention eSense Meter

A value based on the user’s beta brainwaves that are calculated by Neurosky’s propri-
etary algorithm. Attention is associated with the focusing of a single thought by the
user.

3.11 Meditation eSense Meter

A value based on the user’s alpha brainwaves is calculated by Neurosky’s proprietary

algorithm. Meditation indicates the level of mental calmness and relaxation.

4 Sensory Features of Mindwave Mobile

4.1 EEG Frequency Bands

To understand how the Mindwave Mobile works in terms of what EEG data is
captured by its sensor and how it can identify the user’s attention and meditation
levels, a basic understanding of the different EEG Frequency Bands that the Mind-
wave Mobile can detect is needed. The five main brainwave frequencies [13] include
Gamma, Beta, Alpha, Theta, and Delta. Figure 1 shows the different EEG bands and
their associated states.
Gamma Waves. Gamma waves are associated with learning and processing new
information.
Beta Waves. Beta waves are associated with intense, logical, and analytic thinking.
They dominate one’s normal conscious state when our attention is directed towards
cognitive and real-world tasks.
Alpha Waves. Alpha waves are associated with quiet and meditative thoughts. They
aid in mental coordination, calmness, and alertness. The Alpha state of the brain is
known as its resting state.
Theta Waves. Theta waves mostly occur when one sleeps, moreover, such waves
are also present in deep meditation.
Delta Waves. Delta waves are associated with absolute relaxation. They are generated
when one is in the deepest of meditations and dreamless deep sleep.
326 S. Prem et al.

Fig. 1 EEG band

frequencies. Source [11]

4.2 Electrooculography Blink Detection

In addition to being able to output eSense meters, the Mindwave Mobile is also able
to detect blinks [13, 14]. It does this by using Electrooculography or EOG. EOG
measures the electrical potential between electrodes placed at points near the eye or
in the ocular region. Blink detection with EOG follows the concept that whenever a
person blinks, a resulting spike in the EOG data will happen.

4.3 eSense Meters (Attention and Meditation)

Two of the three essential yields of the Mindwave Mobile are its eSense meter read-
ings [15]—Attention and Meditation. Attention just alludes to the user’s center level
while reflection alludes to the degree of serenity. Attention levels increment when an
individual spotlights on a solitary idea, for example, tackling a numerical question,
and diminishes when he is diverted. Contemplation levels, then again, increment
when an individual loosens up his psyche and diminishes when he gets focused.
Regarding the calculation Neurosky uses to interpret crude brainwave information
into its eSense meter readings, they, sadly, have not unveiled it. No documentation
examines the points of interest of how Neurosky does this. This might just be a
BCI Integrated Wheelchair Controlled via Eye Blinks and Brain … 327

result of protection and scholarly copyright reasons and there is no way around it.
Neurosky has, in any case, disclosed that the consideration eSense meter depends
on beta waves while the contemplation eSense meter depends on alpha waves. This
bodes well as beta waves are related to serious, legitimate, and logical reasoning while
alpha waves are related to calm and reflective considerations. Just exactly how they
measure the crude alpha and beta waves, Neurosky has not unveiled. With respect
to the precision and dependability of Neurosky’s eSense meters, they are, similar to
blink detection, accurate and reliable to a certain extent. While a portion of this can
be ascribed to the calculation Neurosky utilizes, the more plausible purpose behind
this is our powerlessness to explicitly tune into and control a specific EEG band
recurrence. There is certifiably not a person on earth that has supreme authority over
his cerebrum’s EEG frequencies. It is simply genuinely and intellectually difficult to,
state; just increase your Beta waves and quiet different frequencies. This failure will
prompt confounding and, now and again, out and out baffling outcomes. You might
be zeroing in such a great amount on attempting to tackle an analytics issue, yet
Neurosky’s eSense consideration worth will even now be low. Neurosky delivered
a disclaimer expressing how its eSense meters are intended for diversion purposes
just—confirmation that eSense meter readings are not 100% precise and solid.

5 Design Overview

5.1 Hardware Development

The Android application will be connected by both the Mindwave Mobile and the HC-
06 Bluetooth module simultaneously. The hardware system flow is shown in Fig. 2. It
begins with the Neurosky Mindwave Mobile, worn around the user’s head, that gets
brainwave-related information and cycles it before remotely sending it through its
own inherent Bluetooth module to the Android Application. The Android application
goes about as an agent between the Mindwave Mobile and the Arduino, which is the
microcontroller that lives in the little wheelchair itself. The Android application can
likewise be considered as a security precautionary measure for the user. Since infor-
mation from the Mindwave Mobile isn’t legitimately communicated to the Arduino,
the Android application can get rid of undesirable information, guaranteeing that
only the important data is sent to the Arduino, thereby making it more efficient. The
Mindwave Mobile and the Android application is interfaced via Neurosky’s Android
application, allowing the transmission and reception of EEG and EOG data. The
subsequent information received by the Android application will at that point be
remotely sent to the HC-06 Bluetooth module, which is associated with the Arduino.
The Arduino will then pass the information to the L298N motor driver. The L298N
Motor driver has 4 DC motors associated with it and is liable for both its speed and
heading control. Depending upon the data passed on by the Arduino, the L298N
328 S. Prem et al.

Fig. 2 Hardware system flow

Motor driver at that point sends resulting orders to the DC motors, permitting the
movement of the wheelchair.

5.2 Software Development

After both the Mindwave Mobile and the Arduino’s Bluetooth module is set up in
association with the Android application, the Android application starts receiving the
signal quality worth, which can be identified as poor, medium, or good. The signal
between the temple skin and the dry sensor, and great if the dry sensor connects
properly with the temple. The Software system flow is presented in Fig. 3.
When a good signal value is obtained, the Android application commences
listening for any incoming force blink data from the Mindwave Mobile. When a
forced blink or a blink whose blink strength value is above the value of 80 is iden-
tified, the Android application begins moving between the four direction values
namely, forward, reverse, left, and right for 10 s with a 2-s interval in between the
change in direction value. This window of 10-s direction-cycle is known as command
mode. During command mode, the Android application listens for two consecutive
blinks also known as a double blink event, from the user. When a double blink even
is detected, the cycling of different directions stops and the direction at which the
cycle was at when the double blink event happened, becomes the current direction.
The blinks are considered consecutive if the time interval between two blinks is less
BCI Integrated Wheelchair Controlled via Eye Blinks and Brain … 329

Fig. 3 Software system flow

330 S. Prem et al.

than 4 s. There is a shift to the focus mode in the Android application once the direc-
tion has been chosen. When the attention values (put out by the Mindwave Mobile
once every 1 s) reaches a value of 50 or more, the Android application changes to
the running mode where it sends a command to the Arduino based on the direction
chosen earlier. Each direction has a respective Bluetooth command that will be trans-
mitted to and interpreted by the Arduino which will be present in the wheelchair.
Similar to command mode, the user exits running mode by blinking consecutively
to go back to standby mode.
From then on, the whole operation loop is repeated if the user wants to move
the wheelchair once again. The speed is kept constant all through activity when the
wheelchair is running. This is because of precision and control-gives that are natural
to the brainwave detection in the Mindwave Mobile. Along these lines, the consistent
speed can likewise be thought of as a safety measure for the user.

6 Conclusion

By using brainwaves and eye blinks to control the motion of a wheelchair, people
suffering from paralysis can easily move around from place to place without the aid
of another person and will give them a sense of independence. This is just an idea, on
the potential of implementing a Brain-Computer Interface system in an everyday item
like a wheelchair. Brain wave technology is a relatively new and upcoming field due to
which several limitations are existent. EEG and brainwave technology is effective to
a certain extent but nowhere near perfect. Blink detection isn’t 100% precise and this
issue can be solved if the blink recognition technology shows signs of improvement.
As human beings do not have complete control of their brainwaves, there might be
inconsistencies and fluctuations in the brain wave data received which makes the
brainwave detection not so precise. As far as human beings learn how to control
and manipulate individual brainwave frequencies, complete and absolute control
over brainwave is impossible leading to minor inaccuracies in the algorithm used to
obtain the required brainwave data. Mindwave Mobile has got issues like hardware
bugs that endanger the sudden stoppage of the wheelchair at times, contributing to
various problems that the researchers have no control over. Some of these issues can
be resolved by using a more advanced EEG headset like EMOTIV Epoch but is more
expensive. This is only preliminary research. More extensive researches have to be
conducted to get more conclusive results. The BCI integration applications hold great
promise in the coming years [16]. Brainwave technologies will without a doubt get
better over time and we are optimistic that soon days will come when BCI will be
successfully incorporated into everyday items.
BCI Integrated Wheelchair Controlled via Eye Blinks and Brain … 331

References

1. Carlson T, del Millan JR (2013) Brain controlled wheelchairs: a robotic architecture. IEEE
Robot Autom Mag 20(1):65–73
2. Butt A, Stanacevic M (2014) Implementation of mind control robot systems. In: Application
and technology conference (LISAT), Farmingdale, pp 1–6
3. Postolache O, Viegas V, Dias Pereira JM, Vinhas D, Silva GP, Postolache G (2014) Towards
developing a smart wheelchair for user physiological stress and physical activity monitoring. In:
IEEE international symposium on medical measurement and application (MeMeA), Portugal,
pp 1–6
4. Kumar RM, Lohit HS (2012) Design of multipurpose wheelchair for physically challenged
and elder people. SASTECH 11(1):108–117
5. Takashi G, Ann G (1998) Developing intelligent wheelchairs for the handicapped. Applied AI
System Inc. (AAI)
6. Nicolas-Alonso LF, Gomez-Gil J (2012) Brain computer interface, a review. Sensors
12(2):1211–1279
7. Keerthana et al (2010) Controlling electrical devices with human brainwaves. https://aasrc.
org/conference/wp-content/uploads/2015/04/Controlling-Electrical-Devices-with-Human-
Brainwaves.pdf
8. Manuel Adrian CA, Anne Monique BC, Dominic Even MP, Lawrence Angelo AT (2012)
PC based hands free short messaging system through facial movements using Emotiv EPOC
headset. De LaSalle University, Philippines
9. Dominguez H (2014) Bluetooth remote controlled car. University of California, Riverside,
California. https://alumni.cs.ucr.edu/~hdomi001/docs/car.pdf
10. Achkar R, Haidar GA, Dourgham H, Semaan D, Araji H (2015) Mobile controlled wheelchair.
In: IEEE European modelling symposium (EMS), Madrid, pp 429–434
11. Psychedelic information theory. https://psychedelic-information-theory.com/eeg-bands
12. Neurosky store. https://store.neurosky.com/pages/mindwave
13. Ahsan MR, Ibrahimy MI, Khalifa OO (2009) EMG signal classification for human computer
interaction: a review. Eur J Sci Res 33(3):480–501
14. EEG: Electroencephalography with Lab VIEW and Mindwave Mobile. https://cerescontrols.
com/projects/eeg-electroencephalography-with-labview-and-mindwave-mobile/
15. EEG eSense Analysis. https://store.neurosky.com/products/eeg-analyzer
16. Nicolas-Alonso LF, Gomez-Gil J (2012) Brain computer interfaces, a review. Sensors
12(2):1211–1279
Classification of Soil Nutrients Using IoT

Sandesh Koli, Dhaval Khobare, Amol Salunke, and Ranjeet B. Kagade

Abstract The analysis of soil wetness systems in inexperienced agricultural houses

supported IoT automation management. This type of intelligent soil wetness system
helps manage the wetness level of providing the water. During this analysis, embed-
ding, a sway system into AN automatic pump controller relies upon the soil’s wetness.
To assess the types and amounts of nutrients in food in the region resulting from plant
growth on contaminated land, and to identify such contaminants’ sources. After
classifying soil nutrients, we know how much the soil is getting polluted and how
to recover it—different crops. Farmers and landowners will realize how not to use
more chemical fertilizer which directly affects the growth of crops.

Keywords Arduino Uno microcontroller · Soil wetness sensing element · Attach

wires · PH module · Color sensor · Bread board

1 Introduction

Soil is a natural body consisting of layers that are primarily composed of different
materials. Soil moisture and nutrients content is an essential factor of a crop’s growth.
Moisture must be measured. Soil nutrients content is the single most crucial factor
in determining plant growth initially. The water calculates devices are programmed
such that they detect arrangements of moisture content. No resource will be more
valuable to agriculture in the future. Soil contamination is distinguished by frozen
or liquid hazardous substances sundry with naturally occurring soil. Soil contamina-
tion can arise from several sources, which could be both naturally occurring in soil
and human made. In the Multi-dimensional flow of water, plants will uptake, and
high frequency of water sources increase the complexity in modelling soil moisture

S. Koli (B) · D. Khobare · A. Salunke · R. B. Kagade

Department of Computer Science and Engineering, SVERI’s College of Engineering, Pandharpur,
India
R. B. Kagade
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 333
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_33
334 S. Koli et al.

passage from trickle irrigation. We used all calculated methods to identify certain
boundary conditions to model the insinuate from water source [1].
The system will check every particle of nutrients and soil contamination a sample
is taken from the land. It refers to what will be the solution for it, and every crop
needs is different based on it. The system allows us to verify which land is suitable
for which crop which fertilizer is ideal for crop growth. Many studies have been done
about the contamination of soil due to coal mine waste, indicating particularly on
the influences on the environment, treatment of coal mine waste, but the research on
Attributes of soil heavy metal pollution around coal mine waste piles was rare [2]. In
precision agriculture and organic farming, it is important to continuously monitor the
fields as they are site-specific. Monitoring plant health is essential, which enriches
the productivity of food grains. Soil moisture is one of the primitive factors for plant
health. The water that remains in soil as a thin film aid in supplying nutrients to the
plant growth. The pollution of soils with heavy non-ferrous metals is a world problem.
Many studies have been achieving soil pollution and the portability of the most toxic
pollutants [3]. There have been many reports from several states such as Maharashtra,
Telangana, Karnataka etc. of contamination of soils with Lead, plumbum, Zinc, etc.
The lead, copper and zinc cannery are the main sources of the contamination [3].
The higher percentage of heavy non-ferrous metals in the soil, compared with the
entrance limit values are established. The children are very conscious of the lead
contamination of soil [3]. All crops have different pH preferences. Knowing the
soil’s pH can help in choosing the right crops and allow proper treatment for the
ground. Soil pH can change every year because of elements, including the type of
rain and consumption of certain nutrients. Hence it is necessary to monitor soil pH
before planting and regularly in a random area of land all over the season’s growth.
But so far, many Indian farmers have not paid attention to soil pH on their land.
The local Government Agriculture Service only takes the Soil pH testing by taking
samples of soil to be measured from a field, then measurements are carried out by lab
tests using a pH meter and using different techniques, or by the calorimeter process.
The results of dimension with conventional systems are less effective because only
the soil pH value can be known from the sample. While all models of soil pH value
of a measured land is unknown, and the measurement time needed is very long [3].
To prevent these all contamination, performing the Soil pH test is very useful for
the first progress of the system. The pH sensor can easily give value on the spot
of farmers land. Farmers can use it to know the nutrients of their land easily and
cheaply. The soil test kit can approximate the amount of nutrients like potassium,
nitrogen and phosphorus, and the soil’s pH. The system takes hazardous samples
of soil for the next process going through a colour sensor for the exact percentage
of nutrients value. The Sensor will help in providing standard recommendations to
reduce human errors. Color comparison and looking for crop recommendations was
eliminated to reduce the time analysis of soil samples. The system was collected of
one digital color sensor and an LDR-RGB LED color sensor for the color analysis
of the soil sample and it is microcontroller-based using Arduino. It also had a data
storage unit for the future use and invention. The stored data get processed using
a program made which provides the analysis and the recommendations of fertilizer
Classification of Soil Nutrients Using IoT 335

use which cannot be contained by the system. This also delivers the database of the
whole process. Standard procedures using the rapid test kit were applied to avoid
incorrect results [4]. Only after the procedures of the rapid test kit, the Sensor was
used for the colour sensing procedure. The only connection of the rapid test kit and
the device was that the device was only a replacement in the analysis of colours to
indicate the nutrient of the soil. Recommendations on the type fertilizers used were
based on the analysis of the data.

2 Required Hardware

The hardware used in the system of classification of soil is pH sensor, Color Sensor,
Arduino microcontroller, Temperature Sensor and Soil moisture Sensor.

2.1 pH Value Sensor

pH scale is used to scale the acidity and basicity of a liquid. It has ranged from 1 to
14 where 1 shows acidic liquid and 14 shows basic liquid. It is expected that the pH
scale is argued to range from 0 to 14 or perhaps 1–14, but neither is correct. The pH
range does not have an upper or lower bound since, as defined above, the pH is an
indication of the concentration of H+ . For example, at a pH of zero the hydronium
ion concentration is one molar, while at pH 14 the hydroxide ion concentration is
one molar. One can go somewhat below zero and slightly above 14 in water because
the concentrations of hydronium or hydroxide ions can exceed one molar.

2.2 Colour Sensor

Colour Sensor, related TCS3200, is a colour detector capable of detecting colours.

The output of the Sensor is a digital wave with a frequency proportional to the
intensity of light. It includes a TCS3200 Red Green Blue sensor chip and four white
LEDs. A color sensor detects the color of the material. This Sensor usually detects
color in RGB scale. This Sensor can categorize the color as red, blue or green. These
sensors are also equipped with filters to reject the unwanted IR light and UV light
[4]. To detect the colour of material three main types of equipment are required. A
light source illuminates the material surface, a surface whose color has to be detected
and the receivers that can measure the reflected wavelengths.
336 S. Koli et al.

2.3 Arduino Uno

Arduino Uno is IoT device that can handle various sensors. It is a microcontroller
with 14 digital input/output pins, a USB connection is used to connect Arduino to
power supply and used for reprogramming. Arduino is an open-source electronics
platform based on easy-to-use hardware and software. Arduino Boards are able to
read inputs—light on a sensor, a finger on a button, or a Twitter message—and
turn it into an output—activating a motor, turning on an LED, publishing something
online. You can tell your board what to do by sending a set of instructions to the
microcontroller on the board. To do so, you use the Arduino programming language
and IDE.

2.4 Temperature Sensor

The temperature sensor is used to sense the temperature in the field and is propor-
tional to the centigrade. The sensors have low output friction, the output and precise
inherent scale make interfacing to the control circuitry easy. A thermocouple or
resistance temperature detector that provides temperature measurement in a read-
able form through an electrical signal. A thermometer is the most basic form of a
temperature meter used to measure the degree of hotness and coolness. Temperature
meters are used in the geotechnical field to monitor concrete, structures, soil, water,
bridges etc. for structural changes in them due to seasonal variations. A thermocouple
(T/C) is made from two dissimilar metals that generate an electrical voltage directly
proportional to the temperature change.

2.5 Soil Moisture Sensor

A Soil Moisture Sensor consists of two major components. A two-pointed Lead,

that goes into the soil where water has to be measured. This has two pointed leads
which connects to the circuit, which is in turn connected to the Arduino shown
in the figure. The Soil Moisture Sensor is used to measure the volumetric water
content of the soil. This makes it ideal for performing experiments in courses such
as soil science, agricultural science, environmental science, horticulture, botany, and
biology. Use the Soil Moisture Sensor to Measure the loss of moisture over Time
due to evaporation and plant uptake. Evaluate optimum soil moisture contents for
various species of plants.
Classification of Soil Nutrients Using IoT 337

3 Objectives

To determine the efficiency of a wet instrument. To balance sensors, insulators probe

to live moisture content. Provide index of nutrients accessibility of soil. Estimation
of the nearby nutrients status of soil. Determination of acidity, salinity and pH scale
of soil. Analysis of suitableness of soil. Helping farmers know how their land is
contaminated, to provide services to them based on growth of crops. Describe organic
substance and how it can be used to improve the soil, Identify strategies to reduce
the impact of fertilizer on water quality. To understand all the problems related
to nutrients which are required for crops, to reduce Time for searching facilities
and solutions. Identify the properties of soil and describe how they influence soil
suitability for growing plants, Identify the variety of soil-dwellers, their benefits, and
strategies for promoting their health.

4 Operating of Sensor

The operating of the Soil detector is straightforward. It works on the principle

of voltage regulation comparison between the Sensor. The subsequent circuits are
useful in understanding the operating of a typical soil wet detector. As you’ll see,
one input of the comparator is connected to a 10 k Potentiometer whereas the
opposite input is connected to a potential divider network fashioned by a 10 k
electrical device and therefore the Soil wet Probe. Based on the number of water
within the soil, the physical phenomenon within the probe varies. If water content
is a smaller amount, the physical phenomenon through the probe is also less and
thus the input to the comparator is high. This suggests the output of the comparator
and as a result, the semiconductor diodes are OFF. Similarly, once there’s adequate
water, the probe’s physical phenomenon will increase and therefore the output of the
comparator becomes LOW. The semiconductor diode then starts glowing.

5 Methodology

As Fig. 1 shows first we need to take soil samples from the random location of the
respected land of farmers. Each sample must contain 30–50 g of the Soil sample. The
random location of land is must needed because the contamination and the nutrients
are not same in different locations due to multiple sources of water being given to the
land borewell water is much polluted. Although the fertilizers were given in a proper
way, thus creating various salty parts of land to get average calculated contamination
we must take all random possible locations for soil samples.
After taking the random soil samples they have to mixed for average calculation
this is the way to find possibly accurate contamination and nutrients of the land. The
338 S. Koli et al.

Fig. 1 Workflow of the system

next process is to differentiate the mixed sample’s pH values if random soil samples
were 15 then mixed samples will be divided into 3, 5 per each. Here the 3 mixed
samples get tested for pH value. Some amount of water gets added in mixed sample
for pH value test, the pH sensor’s tip is drowned in the mixed sample after the device
gets on it shows the pH value of that particular mixed sample. Performing the same
task for the remaining mixed sample. After the pH sensor test the mixed samples are
combined with the Rapid test solution for their nutrients and contamination test. The
rapid test kit contains Chemicals for the making solution of the mixed sample. First
it needs to take only 80% of water and 20% of mixed sample for a good amount of
liquid solution it will give better results. Each chemical is different for each nutrients
like phosphorus, nitrogen, zinc etc., however the chemicals get reacted with the
mixed sample it turns into various colors and each color sample is different from
another like pink for nitrogen, blue for phosphorus and orange for Potash. After
adding chemicals into the samples to make a solution it kept 5 min to get a complete
reaction and emit the colors. The level of the darkness and fitness shows the amount
of nutrients is present per cubic feet.
In the color sensing process, each sample solution is kept under the color sensor
TCS2300 unit, configured to emit light and get refracted from the solution. The
refracted light has data of how much nutrients are present and how the sample is
contaminated. For accurate results, the Sensor’s data is passed through the Arduino
Uno that verifies and gives the actual percentage of nutrients.
The dataset was first grouped assuming that there were two clusters and analyzed
for consistency. These two clusters would represent “naturally contaminated” soil
samples and “unnaturally contaminated” soil samples. The datasets were subse-
quently grouped with the assumption that three or more clusters were present. These
cluster centroids would also represent different types of contamination and poten-
tially different sources of contamination. The features were investigated to determine
a hierarchy of significance. The importance of a metal with respect to its impact on
classification was evaluated by comparing the standard deviation of the lognormal
concentration spread to the “distance” between the centroids at that feature. The
closer the concentrations at the centroids, the less “important” it was deemed to
Classification of Soil Nutrients Using IoT 339

be. The data is collected in the all the format for analysis and reporting purpose,
In analysis phase the each of mixed sample results get passed under the method of
percentage of nutrients present in the sample. If potash is more dark in orange, then
the potash amount is more than enough to present in the soil, but it won’t damage
crops. All the nutrients like Nitrogen, Phosphorus and Potash get classified for accu-
rate results. The results show much better than imagination. If the contamination
is more than land there must be a Decontamination process in which land is acidic
then use the basic fertilizer if land is basic then use limestone to maintain pH of the
land. If the nutrients are not enough to grow good crops, use fertilizers suitable for
land like nitrogen is less present in the land then use fertilizers that have nitrogen in
good amounts. Here in the solution phase, the farmers know how much their land is
contaminated and how much nutrients are present in the land.

6 Result

Efficient results are captured from the project. Tables 1 and 2 shows the recorded
knowledge at the field in many days. The hardware implementations are roaring and
reliable, and therefore, the sensors that we tend to exploit are giving idle results and
playing to the belief.
As Table 1 shows the temperature of soil and atmosphere is different for each day
also the moisture. Here we can say the atmosphere’s temperature directly affects the
moisture present in the soil.
As Table 2 shows, the pH values are different for each day because of soil moisture,
the nutrients of soil can be managed by water sources in a quality way.

Table 1 Readings of
Date and time Atmospheric Soil Soil moisture
temperature and soil moisture
(1.30 p.m.) temperature temperature (%)
25/05/2020 29 28 32
26/05/2020 28 27 30
27/05/2020 30 29 28
28/05/2020 32 31 26

Table 2 Readings of pH,

Date and time pH value Contamination Percentage of
percentage of contamination
(1.30 p.m.) nutrients
and nutrients
25/05/2020 8 08 69
26/05/2020 8 10 70
27/05/2020 7 11 65
28/05/2020 8 9 68
340 S. Koli et al.

7 Applications and Future Scopes

A Soil Wetness Sensor has many appeal, especially in agriculture. Irrigation is a key
factor in farming. Detecting the amount of moisture and temperature in the soil and
turning on the system when the wetness level falls below a predefined value helps
keep away a lot of wastage of water and human resources. These classes of sensors
make automatics easier. A database can be formed. It can be used to direct the types
of acids, alkalis or salts present in the soil. Saltiness of soil can also be calculated
by correlating it with the output voltage. Signal carrying of the output data directly
to the user can be done using Zigbee or Bluetooth. We can get the values from a
stored database in PC so that the soil’s moisture holding capacity can be determined.
Indian farmers are facing a lot of problems, but the advancement of wireless sensor
networks in agriculture would be promising in the present scenario of water scarcity
and unpredictable weather conditions. This paper provides implementation of WSN
based soil moisture monitoring system. As WSN is battery operated to enhance the
lifetime. EWMA event detection algorithm is used which generates events only when
threshold conditions are met. Rest of the Time, the nodes are in a sleep state which
can save their energy. This work can also be further extended by considering more
than one sensor module.

8 Conclusion

The soil wet response observation system designed is extremely straightforward

to grasp and handle. It will be operated by all age-groups of farmers. It is often
reprogrammable to add a lot of options. The wet is measured up to the foundation
zone of the crop. So I often want to check the wet worth for any crop. Detector is often
placed vertically in the soil to examine the depth of irrigated water. It will be placed
horizontally at completely different heights in the soil consistent with the crop. It’s
user friendly and may be employed by uneducated farmers. The wet is checked in the
morning and the evening and it is found that wet is linear up to 20% VWC (volumetric
water content) and after output voltage becomes virtually constant, classification, it
can be stated with high confidence that our team has apparent soil Classification.
These Nutrients is characterized by exceptionally high concentrations of Nitrogen
Phosphorus Potash and zinc. The contamination source is yet unknown. It might be
wrong and hazardous fertilizer, given the provided data, but may be related to soil
sample depth or specific site operations.
Classification of Soil Nutrients Using IoT 341

References

1. Angelakis AN, Rolston DE, Kadir TN, Scott VN (1993) Soil-water distribution under trickle
source. J Irrig Drain Eng ASCE 119:484–500
2. Zhang M, Wang H (2009) Characteristics of soil heavy metals pollution around mine waste
piles. In: International conference on ES and IAT. https://doi.org/10.1109/2009.319
3. Prastika YFD, Zani NR, Arifin W, Alasiry AH (2019) Design and development soil pH mapping
portable system for crop selection using fuzzy algorithms on agricultural land. In: IEEE Asia-
Pacific conference on geoscience, electronics and remote sensing technology (AGERS), Jakarta,
pp 43–48. https://doi.org/10.1109/AGERS48446.2019.9034381
4. Regalado RG, Dela Cruz JC (2016) Soil pH and nutrient (nitrogen, phosphorus and potassium)
analyzer using colorimetry. In: IEEE region 10 conference (TENCON), Singapore, pp 2387–
2391. https://doi.org/10.1109/TENCON.2016.7848458
Non-invasive Methodological Techniques
to Determine Health of a Bone

Meghana R. Khare and Raviraj H. Havaldar

Abstract The healthiness of bone is decided by organic as well as inorganic contents.

Inappropriate proportions of organic (collagen) and inorganic (minerals) contents
give rise to fractures and diseases like arthritis, osteopenia, and osteoporosis. Bone
Mineral Density (BMD) is a commonly used inorganic clinical indicator to deter-
mine the quality of bone. There are invasive methods like chemical quantitative
analysis, and noninvasive methods include imaging techniques, to find the compo-
sition of bone. This paper is to review different medical imaging techniques like
X-Ray, DEXA, SR µCT, Photo Acoustic Imaging, Quantitative Ultrasound to do
bone mineral density analysis. The study concludes that the imaging technique Dual
Energy X-ray Absorptiometry (DEXA) will be a more useful diagnostic modality for
rapid investigation of bone health. It is a two-dimensional imaging tool that calcu-
lates a ratio of Bone Mineral Contents (BMC) with an area of bone is known as
areal BMD. But the shortcoming of DEXA technique is that the bones of different
lengths may produce the same BMD results by maintaining the equal ratio thereby
indicating incorrect strength of bones. Hence along with the area, we have to take
into account the thickness of the bone to predict BMD.

Keywords Bone mineral density · DEXA · Invasive · Non-invasive

1 Introduction

Bones are living and continuously developing tissue. Bone strength continually
changes throughout life. In the remodeling process, some bone cells dissolve and
new bone cells get developed. Bone isn’t uniformly solid, but consists of a versatile
matrix and bound the minerals. Bone matrix is collagen that gives a soft framework
to the body, while minerals add strength and hardness to the body framework [1].

M. R. Khare (B)
Department of Electronics Engineering, Walchand College of Engineering, Sangli, India
R. H. Havaldar
Department of Biomedical Engineering, Dr. M. S. Sheshgiri College of Engineering and
Technology, Belagavi, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 343
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_34
344 M. R. Khare and R. H. Havaldar

This mixture of collagen and minerals makes bone strong and versatile enough to
face up to different stresses.
By considering weight as a parameter, 60% of bone tissue is the inorganic
constituent, 30% organic constituents, and the remaining 10% is water. While by
considering volume as a parameter, 40% is the inorganic component, 35% is the
organic component, and the remaining 25% is water [2]. The bone may be a reser-
voir for various minerals. Bone minerals are categorized into two types like macro and
micro. Macrominerals include Calcium (Ca), Phosphorous (P), Magnesium (Mg),
Potassium (K), Sodium (Na), and Sulphur (S). Micro minerals contain Copper (Cu),
Iron (Fe), Manganese (Mg), Zinc (Zn), Cobalt (Co), etc. Both types of minerals
hold immense importance to ascertain the quality of a bone. The requirement of
Macrominerals is more as compared to micro minerals [3]. Calcium is the major
macro mineral present in the body. It plays a vital role in the formation of bones,
teeth, and as a messenger in cell signaling. The minerals Calcium and phosphorous
are present in the complex form of insoluble salt called hydroxyapatite [4]. The
variation in the proportion of Calcium to Phosphorous contents may make the bone
porous or weak. Consequently, this provides rise to osteopenia, osteoporosis, and
the probability of rupture in a bone [5]. Hence, it’s significant to determine the bone
mineral contents. Mineral contents and mineral density can be assigned with each
other through areal measurements. The areal Bone Mineral Density (aBMD) is the
ratio of bone mineral content with a particular area of the bone (ROI). Volumetric
Bone Mineral Density (vBMD) indicates the relation of bone mineral content with
the volume of the particular bone region [6].
The determination of BMD can be done either by invasive method or noninva-
sive method. The invasive method is applied to small animals like mice, rabbits,
hamsters, pigs, and dogs [7]. The method requires an actual bone sample in the form
of ash for mineral content measurement. To obtain ash of animal bones, first, they
are dried at low temperature and then burnt at high temperature so as to get bone ash.
Due to this drying and burning process, organic contents of bone vanish and only
inorganic constituents remain [7]. Scanning electron microscopy–Energy-dispersive
X-ray (SEM–EDX) [8], XRD, XRF, and LIBS [9] are some methods used to quan-
tify inorganic constituents from bone ash. These techniques have high accuracy but
require special sample preparation (in vitro) and are often slow that may take 5–
7 days to get the results [10]. So applying the same invasive technique on human
bone is not as suitable as on animal bones [11].
The noninvasive method measures the BMD, by using different scanning tech-
niques on live animals [7]. Quantification of bone density to diagnose the health of
bone is done by using quantitative imaging techniques. The purpose of Non Inva-
sive imaging techniques is to study the clinical tools for quantification of inorganic
Non-invasive Methodological Techniques to Determine Health … 345

constituents (in terms of BMD) to determine the health of a bone. The computable
imaging techniques are extensively used due to its more accuracy, reusability, and
validation on a huge number of populations.

2 Use of Image for Diagnosis of Health of a Bone

The noninvasive clinical techniques to diagnose the health of a bone include the size of
cortical thickness, quality analysis of trabecular bone, and radiolucency from radio-
graphs. These parameters are used to quantify mineral density. Total body weight is
supported by bones like the femur, hip, forearm, etc. As a result, these bones are more
prone to fracture risk. This section describes different clinical imaging techniques
applied for the clinical analysis of the health of bone.

2.1 Radiography

X-ray source emits two types of radiations as low energy and the other high energy
[12–14]. When radiations are passed through a sample, inorganic contents present
in that sample absorb some energy. The unabsorbed energy is detected by an X-
ray detector. The difference between source energy and unabsorbed energy helps
to determine Bone Mineral Density. The value of BMD (g/cm2 ) is multiplied with
related area to give BMC (g) [14, 15]. Thus a bone mineral density measures how
much calcium and other inorganic contents are available in a particular region of the
bone.

2.2 Dual Energy X-Ray Absorptiometry (DEXA)

DEXA is a standard clinical technique that gives an areal amount of bone mineral
density [12]. In DEXA scan imaging technique, bone is radiated by photon beams of
2 energy levels, high energy of greater than 70 keV and lower one is 30–50 keV [16].
DEXA has good precision, less contact to radiations, and a fast scanning time [17].
Determined BMD is used to diagnose the health of bone. The sample for DEXA scan
of the femur bone is shown in Fig. 1.
The health of a bone is categorized as normal bones, osteopenia bones, and the
one with osteoporosis based on BMD values as shown in Fig. 2. The obtained results
are helpful to determine the strength of bone. DEXA BMD outcomes are classified
from standardized T-score and Z-scores values (Table 1) [18]. It shows a variation of
measured bone density to that of a healthy person. The lesser the T score, the more
fragile bones are.
346 M. R. Khare and R. H. Havaldar

Fig. 1 DEXA image of femur bone

Fig. 2 Result of DEXA scan

Table 1 T-score for

Health of bone T-score
diagnosis of health of a bone
according to WHO Normal bone Greater than −1
Bone with osteopenia Between −1 and −2.5
Bone with osteoporosis Less than −2.5
Non-invasive Methodological Techniques to Determine Health … 347

(patient  s_BMD) − (Mean_young_adult_BMD)

T − score =
SD_of _young_adult_BMD
(patient  s_BMD) − (Mean_age_matched _BMD)
Z − score =
SD_of _age_matched _BMD

The drawback of DEXA scan is its more cost and limited availability [18, 19].

2.3 Quantitative Computed Tomography (QCT)

QCT gives a volumetric quantity of bone mineral density. In this imaging technique,
applied X-rays absorb by tissues of a bone and scatter the same at various directions
across the Region of Interest. To measure BMD values, samples of various bone
mineral densities are used. The received attenuation signal is used (Hounsfield units)
to convert into BMD values [12]. The obtained T-scores of QCT method are lower
than that of DEXA method. Martin-Badosa et al. have investigated the risk of fracture
in terms of volumetric BMD as shown in Table 2 [19, 20].
For QCT imaging technique WHO has given following range of BMD as shown
in Table 3 [21].
There are different types of QCT like Multidetector CT (MDCT), High-Resolution
Peripheral (HR-pQCT), Synchrotron radiation micro Tomography (SR-µCT).
MDCT offers a good spatial resolution. Bone Mineral Density and extracted
microstructures help to distinguish between healthy and unhealthy groups [22, 23].
HR-pQCT helps in getting higher accuracy volumetric images of the bone. A stan-
dardized analysis is used to quantify volumetric BMD. Due to marginal quantifica-
tion, the technique is appropriate for patients paining from spinal and hip fractures
[24, 25]. SR µCT Technique is used to obtain three-dimensional images to quantify
trabecular bone architecture [26]. This paper proposes a technique for measuring
the amount of minerals in bone samples using 3-Dimensional SR µCT images.

Table 2 Fracture risk

Fracture risk BMD
according to volumetric BMD
values Low From 80 to 110 mg/cm3
Moderate From 50 to 80 mg/cm3
High Less than 50 mg/cm3

Table 3 Trabecular BMD for

Health of bone QCT trabecular spine BMD range
diagnosis of health of a bone
according to WHO Normal Greater than 120 mg/cm3
Osteopenia From 80 to 120 mg/cm3
Osteoporosis Less than 80 mg/cm3
348 M. R. Khare and R. H. Havaldar

The captured 3 D image slices are considered to find gray levels of the images.
Variations in grey levels indicate different mineralization. Regions with darker grey
levels indicate low mineralization and newly formed bones. To calibrate the results
of mineral contents, homogeneous water solutions with different concentrations of
Dipotassium hydrogen phosphate is used. The result obtained from the chemical anal-
ysis was compared with those obtained from microradiography. SR µCT technique
provides overall mineral contents as well as the three-dimensional microarchitecture
of trabecular bone.
But Overall Limitation of any QCT method is that it has high radiation as compared
to DEXA technique.

2.4 Magnetic Resonance Imaging (MRI)

It is a non-invasive method to assess bone structure, bone quality, and its microar-
chitecture [12, 27]. MRI is a non-ionizing Imaging technique that helps in volume-
wise checking of trabecular bone as well as cortical bone. When compared to QCT
technique MRI technique has no radiations but more expensive [28].

2.5 Digital X-Ray Radiogrammetry (DXR)

It is a hi-technology texture analysis method used to measure BMD [12]. The tech-
nique is about showing an association of DEXA–BMD values of forearm, femur
and spine. DXR has less possibility of human error and results are comparable with
DEXA However, the volumetric BMD measurement by this technique is not as
precise as compared with QCT technique.

2.6 Quantitative Ultrasound (QUS)

It is a technique of measuring trabecular bone construction [12]. It uses frequented

sound waves of 500 kHz to 1.25 MHz to transmit through the part of a body. The
properties ofa bone are defined by two quantities, speed of sound (SOS) and broad-
band ultrasound attenuation (BUA). For unhealthy or poor bone, the value of BUA
elevates while the value of SOS falls. Other evaluation parameters are amplitude-
dependent SOS and surface combined backscatter. QUS is a low cost, simple, no
emission technique. Calibrating QUS takes the help of DEXA measurements [17].
Other shortcomings of QUS include lesser accuracy and reliability.
Non-invasive Methodological Techniques to Determine Health … 349

3 Conclusion

The invasive and noninvasive methods are available to find BMD. The invasive
method needs actual bones to be processed to create ash of sample which is quite
a time-consuming process and requires ethical permissions. The benefit of imaging
techniques is lesser time consumption compared with invasive techniques. In this
review, different Imaging techniques like radiograph, QCT, QUS, MRI, DXR and
DEXA have been discussed. QCT and MRI are the techniques that offer volumetric
BMD but they are expensive. DEXA is a two-dimensional technique for the determi-
nation of areal BMD which is precise, more accurate but it has disadvantages like low
availability to date. For DEXA, the bones of different lengths may produce the same
BMD results by maintaining the equal ratio thereby indicating incorrect strength of
bones. Hence along with the area, we have to take into account the thickness of the
bone to predict BMD. So it’s the need in today’s situation to develop a mathematical
model to provide a better clinical analysis of bone health without additional cost and
distraction.

References

1. Boskey AL (2013) Bone composition: relationship to bone fragility and antiosteoporotic drug
effects. Int Bone Miner Soc
2. Feng X (2010) Chemical and biochemical basis of cell-bone matrix interaction in health and
disease. Curr Chem Biol
3. Costa e Silva LF, Filho SDCV, Engle T, Rotta PP (2015) Macrominerals and trace element
requirements for beef cattle. PLoS One. https://doi.org/10.1371/Journal.Pone.0144464
4. Swetha M, Sahithi K, Moorthi A, Srinivasan N, Ramasamy K, Selvamurugan N (2010) Biocom-
posites containing natural polymers and hydroxyapat ite for bone tissue engineering. Int J Biol
Macromol 47(1):1–4. https://doi.org/10.1016/j.ijbiomac.2010.03.015
5. Faibish D, Ott SM, Boskey AL (2006) Mineral changes in osteoporosis—a review. Clin Orthop
Relat Res 443:28–38
6. Humbert L, Martelli Y, Fonollà R, Steghöfer M, Di Gregorio S, Malouf J, Romera J, Del Río
Barquero LM (2017) 3D-DXA: assessing the femoral shape, the trabecular macrostructure and
the cortex in 3D from DXA images. IEEE Trans Med Imaging. https://doi.org/10.1109/Tmi.
2016.2593346
7. Kim HS, Jeong ES, Yang MH, Yang S-O (2018) Bone mineral density assessment for research
purpose using dual energy X-ray absorptiometry. Osteoporos Sarcopenia 4
8. Ellingham STD et al (2017) Scanning electron microscopy–energy-dispersive X-ray
(SEM/EDX): a rapid diagnostic tool to aid the identification of burnt bone and contested
cremains. J Forensic Sci. https://doi.org/10.1111/1556-4029.13541
9. Tariq U, Haider Z, Tufail K, Hussain R, Ali J (2016) Determination of calcium to phosphate ratio
in hydroxyapatite extracted from bovine bone using libs, vol 2, pp 48–53. eISSN 2504-8546
10. Sluiter A, Hames B, Ruiz R, Scarlata C, Sluiter J, Templeton D (2005) Determination of ash
in biomass, laboratory analytical procedure
11. Khalid M, Bora T, Al Ghaithi A, Thukral S, Dutta J (2018) Raman spectroscopy detects changes
in bone mineral quality and collagen cross-linkage in staphylococcus infected human bone.
Sci Rep 8:9417
12. Areeckal AS, Sumam David S (2019) Current and emerging diagnostic imaging-based
techniques for assessment of osteoporosis and fracture risk. IEEE Rev Biomed Eng 12:254–268
350 M. R. Khare and R. H. Havaldar

13. Chuaychunu N, Pititheerapab Y, Chanwimalueang T, Lertprasert P, Pintavirooj C (2007) Bone

mineral density and bone mineral content estimation using low-cost X-ray detector. In: IEEE
ICICS 2007
14. Khan SS, Jayan AS, Nageswaran S (2017) An image processing algorithm to estimate bone
mineral density using digital X-ray images. IEEE. 978-1-5090-3239-6/17/2017
15. Promworn Y, Pintavirooj C (2012) Development of bone mineral density and bone mineral
content measurements system using a dual energy X-ray. IEEE. 978-1-4673-4892-8/12
16. Shankar N, Babu S, Simon LA, Philip G, Vanaja N (2017) Comparison of X-ray and DXA for
evaluating osteoporosis. Int J Res Sci Innov IV(III). ISSN 2321 – 2705
17. Choksi P, Jepsen KJ, Clines GA (2018) The challenges of diagnosing osteoporosis and the
limitations of currently available tools. Clin Diabetes Endocrinol 4:12. https://doi.org/10.1186/
s40842-018-0062-7
18. Shaikh AB, Sarim M, Raffat SK, Khan M, Chinoy A (2013) Bone mineral density correlation
against bone radiograph texture analysis: an alternative approach. Res J Recent Sci 2(3):87–91.
ISSN 2277-2502
19. Michael Lewiecki E (2006) Bone density testing in clinical practice. Arq Bras Endocrinol
Metabol 50(4):586–595
20. Link TM (2012) Osteoporosis imaging: state of the art and advanced imaging. Radiology
263(1):3–17
21. ACR–SPR–SSR practice parameter for the performance of musculoskeletal quantitative
computed tomography (QCT), revised 2018 (resolution 9)
22. Chen C et al (2018) Quantitative imaging of peripheral trabecular bone microarchitecture using
MDCT. Med Phys 45(1):236–249
23. Mei K et al (2017) Is multidetector CT-based bone mineral density and quantitative bone
microstructure assessment at the spine still feasible using ultra-low tube current and sparse
sampling? Eur Radiol 27(12):5261–5271
24. Jiang H, Yates CJ, Gorelik A, Kale A, Song Q, Wark JD (2018) Peripheral quantitative computed
tomography (pQCT) measures contribute to the understanding of bone fragility in older patients
with low trauma fracture. J Clin Densitometry
25. Ohlsson C et al (2016) Cortical bone area predicts incident fractures independently of areal
bone mineral density in older men. J Clin Endocrinol Metab 102(2):516–524
26. Nuzzo S, Peyrin F, Martín-Badosa E, Lafage-Proust MH, Boivin G (2001) Assessment of bone
mineral content from 3-D synchrotron radiation microtomography images. IEEE Trans Nucl
Sci 48(3):859–863
27. Chang G, Boone S, Martel D, Rajapakse CS, Hallyburton RS, Valko M, Honig S, Regatte
RR (2017) MRI assessment of bone structure and microarchitecture. J Magn Reson Imaging
46(2):323–337
28. Deniz CM, Xiang S, Hallyburton RS, Honig S, Cho K, Chang G (2018) Segmentation of the
proximal femur from MR images using deep convolutional neural networks. Sci Rep 8:16485.
https://doi.org/10.1038/s41598-018-34817-6
Face Detection and Recognition Using
Raspberry Pi

P. R. Dolas, Pratiksha Ghogare, Apurva Kshirsagar, Vidya Khadke,

and Sanjana Bokefode

Abstract In Today’s world, security frames the most essential segment of our lives.
Face Recognition is an important part of the purpose of the security and surveillance
field. A small project which does face detection using OpenCV library on Raspberry
Pi. Face Recognition/Facial Recognition is a category of biometric software that
identifies people by their faces. The face is captured by the digital camera and the
system is trained and then it is capable of identifying the person. This paper focuses
on the implementation of a face detection system for human identification based on
the open-source computer vision library (OpenCV) with python. We also proposed a
hierarchical image processing approach to reduce the training or testing time while
improving recognition accuracy.

Keywords Face recognition · Raspberry Pi · PI camera

1 Introduction

The Concept of image processing through the python OpenCV platform has been
used for human identification through face detection. Human Identification means
to recognize particular people through their unique structures like fingerprint, palm,
and iris, and face detection. This paper is based on the implementation of the face
detection system. source library in which the source code is open and it is useful in
the visual field such as image processing. The main motto of this work is to take
and manage attendance using face recognition. The testing of this technique has
proceeded through Raspberry Pi devices [1].
Face recognition is an active and important research topic from the 1970s. Given an
input image with 1 face, face recognition systems typically first run face detection to
isolate the faces. Each face is preprocessed and then a low-dimensional representation
(or embedding) is obtained.

P. R. Dolas (B) · P. Ghogare · A. Kshirsagar · V. Khadke · S. Bokefode

Department of Computer Science and Engineering, SVERI’s College of Engineering, Punyashlok
Ahilyadevi Holkar Solapur University, Solapur, Maharashtra, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 351
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_35
352 P. R. Dolas et al.

2 Literature Survey

In the proposed system, we use the camera to accomplish the pictures when a motion
detects via the PIR sensor. Subsequently, we will apply a computer vision module to
the caught pictures to discover the faces. This system is extremely helpful and vital if
we want to protect the area. The application can be divided into two parts which are
motion detection and face detection. The system will not go to face detection if there
is no motion discovered. But, if a movement has been detected, then the detected
movement of the current frame will be processed by the algorithm.
A review of Face Recognition System Using Raspberry Pi in the Field of IoT Jain
et al. algorithm of face detection. To explicitly define a low-dimensional face repre-
sentation based on ratios of distances, areas, and angles. An explicitly defined face
representation is desirable for an intuitive feature space and technique. However,
in practice, explicitly defined representations are not accurate. Later work sought
to use holistic approaches stemming from statistics and Artificial Intelligence (AI)
that learn from and perform well on a dataset of face images. Statistical techniques
such as Principal Component Analysis represent faces as a combination of eigenvec-
tors. However, these deep neural network-based techniques are trained with private
datasets containing millions of social media images that are orders of magnitude
larger than available datasets for research [2, 3].

3 Methodology

Raspberry PI is a small computer that has been used to develop an embedded system
to perform a specific particular task. This electronic module has been operated with
the use of the Raspbian operating system and is based on the LINUX platform.
This module has an SD card slot, Inbuilt WIFI, and Bluetooth Connectivity, 40
GPIO pins for Input–output operations, PI Camera Port, PI Display port, Audio Port,
HDMI cable Port, 4 USB port for connecting pen drive, Mouse, Keyboard, USB
Camera, etc. It also has an Ethernet port for data sharing as well as network sharing
between computers and raspberry PI [4].
The Functionality of this system is mainly categorized in the following steps:
To enroll and detect faces using a camera connected to the ARM Cortex of
Raspberry Pi board.
To display the match status on the LCD as well as the terminal running on the
VGA (Video graphics array) monitor.
To program for the same using python language. The code imports certain modules
that enable functions such as face recognition, GPIO modules. The identification
and authentication technology operate using the following four stages shown in
Figs. 1 and 2 clear explanations in the modules are presented.
Face Detection and Recognition Using Raspberry Pi 353

Fig. 1 Block diagram of implementation of face detection and recognition system using Raspberry
Pi

Fig. 2 Flow of operation

3.1 Figures and Tables

1. Capture: A physical or behavioral sample is captured by the system during

enrollment and also in the identification or verification process.
2. Extraction: Unique data is extracted from the sample and a template is created.
3. Comparison: The template is then compared with an existing sample.
4. Match/non-match: The system decides if the features extracted from the new
samples are a match or a non-match and accordingly accept/reject.
354 P. R. Dolas et al.

Fig. 3 Actual result

4 Conclusion and Results

Efficient human Identification through face detection has carried out relevant infor-
mation. The performance of the system is based on three steps which are datasets,
trainer, and detector python script. An algorithm that has been used for image
processing in OpenCV and especially for face detection as shown in Fig. 3. As
we talk about future modification of this project, it will be used in a high-security
system and face detection-based attendance system. Also, we can develop projects
based on image processing.

References

1. Patel T, Shah B (2017) A survey on facial feature extraction techniques for automatic face anno-
tation. In: IEEE international conference on innovative mechanisms for industry applications
(ICIMIA), Bangalore
2. Foster I, Kesselman C (1999) The grid: blueprint for a new computing infrastructure. Morgan
Kaufmann, San Francisco
3. Yang MH, Kriegman DJ, Ahuja N (2002) Detecting faces in images: a survey. IEEE Trans PAMI.
Face recognition based on elastic template. Beijing University of Technology, China
4. Mathur S, Subramanian B, Jain S, Choudhary K (2017) Human detector and counter using
Raspberry Pi microcontroller. In: IEEE international conference on innovations in power and
advanced computing technologies i-PACT2017, Vellore
Human Tracking Mechanism
for Institutions Using RFID and Facial
Recognition

Rameez Shaik and L. V. Patil

Abstract The Adoption of digital content by the institutional members and students
have seen rapid growth in recent years. The students could access the content across
various devices, platform and applications, which has a direct implication on the
physical presence of the student. The institutions follow old and very traditional
base approach like a manual record of attendance to track the company of students
which consumes a lot of time and efforts from the staff members. The study looks at
the various technologies available in the market and present the implementation of
the best possible solution. The latest use of technology of Facial Recognition with
a combination of RFID will enhance the tracking process and also provide valuable
insight into student behaviour. The data collected by the system can further utilize
to improve the efficiency and effectiveness of student behaviour patterns and predict
the learning trend, which will help the institutions to make the correct decisions.

Keywords Application · Big data analytics · Bio-metric · Bluetooth · Face

recognition · Iris · Tracking system

1 Introduction

Physical participation is one of the ways to showcase if someone is interested or to get

involved. In the educational institutions, the physical appearance of the students plays
a vital role to showcase the interest in educational activities. The research proved
that the physical participation of students has a direct impact on their academic
results. The additional case expressed that the student who has less involvement in
the institutions tends to have poor performance in academics.
The faculty members of an Institute tracks the attendance of each participated
student. The existing way to maintain manual records is very time-consuming, not
effective and require someone to orchestrate the process. The paper records have
then been store into the digital system for further references. There is a fair chance
of data quality and manipulation issues for the data which is entered manually by the

R. Shaik (B) · L. V. Patil

Department of I.T, SKN College of Engineering, Pune, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 355
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_36
356 R. Shaik and L. V. Patil

institution employees. Most of the Indian educational institution does not have the
correct framework to address these problems [1, p. 3]. Some of the common issues
faced in the traditional use of tracking the presence of student are manual errors, loss
of recorded copies which will not help the students and well as institutions. So there
is a need to have a mechanism which could automate the process of capturing the
presence of student without any manual intervention.
The existing use of technology must introduce an electronic framework that can
oversee and assist the faculty members with taking the attendance rapidly and error-
prone. The institution can quickly implement the automated framework digitally
without manual intervention. The other concern of having the attendance records in
the copy form is that the faculty member may lose it and would be difficult to retrieve
it [2, p. 2].
The digital framework is generating server related activities, which are reaching
in terabytes. The student activities and digital educational system modules receive
an enormous amount of data daily. Traditional analytical programs prolong to meet
the analysis requirements as the data requirements are exploding. To identify the
behaviour correctly, we do need a mobile base application that makes easy to access
the digital format on all the available platforms and necessary devices. In the current
days, mobile base application on various platform like Android, IOS or Windows
helps to capture correct data is provide a basis for a huge number of simultaneous
processing and analysis to get the valuable insights. The institutions for studies
are attaining valuable insights based on the processing of analytical tools to make
improvement on the education sector on the larger different culture base population
of students. While the improvement with respect to digital content are progressing
quicker, biometrics innovation are progressing at faster pace. These innovations
address the students personality through various features like fingerprints, face detec-
tion, Iris detection, retina scan samples, palm prints, voice recognition and so on. The
techniques used related to these features helps to identify and validate user authen-
tication that could be more trustworthy than other type of security system, which
involves password or identity cards. The biometric data are more trustworthy and
error prone and can contain it for more duration of time [3, p. 6].
It is atmost critical to validate the correct devices and instruments to use in logical
test cases. Common standard validation methods like radio frequency device frame-
work, Bluetooth devices, and NFC devices are few cases which can be utilise in
the process. The upfront cost of these techniques are very costly to do the imple-
mentation and has lot of restrictions to use it. In today’s generation digital related
validation process contains number of biometric related devices and systems. The
improvements in these areas and the techniques usage depends on physiological or
social features. Biometrics denotes as the “robotized distinguishing proof or confir-
mation of human personality through the estimation of repeatable physiological and
conduct attributes”. Different kinds of biometric methods are listed (Fig. 1).
Human Tracking Mechanism for Institutions Using RFID and Facial … 357

Fig. 1 Biometrics techniques

2 Related Work

The section eventually reveals some facts based on thoughtful analysis of many
authors work as follows.
Core Research recommended the utilization of a mechanized participation frame-
work, which can lessen human contribution, manual information section mistake, and
monotonous work. This framework is going to improve profitability, reduce finance
mistake, and brought down compensation swelling, decreased extra time, the retire-
ment of heritage frameworks, Elimination of paper expenses, and which can give all
the reports on request. In this framework, the workforce needs to gauge participation
physically, and just these records must be gone into the advanced framework [4, p. 2].
Another similar kind of project proposed, but as part of use case, the student should
register themselves using client-server socket program from their own devices such
as laptops. A session of the application creates an execution process for the list
of students for a specific course. These courses display when the faculty members
approve the request or start the application. The student presence is marked when
clicked to checkbox of the student name, and afterwards, click the section button to
stamp their presence.
In 2013, Bhalla et al. [5], have recommended the participation framework, which
can gauge participation in applying Bluetooth. In this task, presence marked using
teacher’s cell phone. Application installed in faculties mobile device which triggers
the notification on student mobile devices using Bluetooth and the data exchange
by PDA Media Access Control (MAC). The student appearance marks once they
acknowledge the report.
358 R. Shaik and L. V. Patil

In Ayu [6] suggested the online participation framework using NFC innovation
which named as Tochin System. In the framework, two methods of activities used
where one of them is writer mode and other is a peer to peer mode. In this framework,
each room of the institution has NFC enabled, which is a program to connect to the
faculty member’s computer.
In 2012, Josphineleela and Ramakrishnan [7, p. 4] presented a framework in
which participation has done using one of the biometric features of finger impres-
sion. This framework can apply for students and faculty members of the institu-
tion. In this method, the unique finger impression acknowledged for the presence
of the subject, and it is orchestrated into the accompanying modules Pre-preparing,
Minutiae Extraction, Reconstruction, Fingerprint Recognition and report. This new
calculation reproduces the stage picture from Miniature.
In 2010, Kadry and Smaili suggested using the iris-based framework, which is one
of the techniques of bio-metric. A remote iris base organization framework executed,
utilizing Daugman’s calculation. This biometric base framework is challenging to
implement due to the complex process of data. The implementation of the framework
is too expensive and time-consuming authentication. The use of the iris base tech-
nology is implemented in the financial sector, such as banks and financial institution,
where complex authentication is the highest priority.

3 Implemented Framework

The human tracking system is a global application framework to improve and auto-
mate the manual intervention to write down and transmit in near real-time. The proper
way of implementation consists of two options which are related to authentication
used by the RFID base device and another option is to use the biometric technique
such as facial recognition which matches with the saved images. Once the match
is successful, then presence is marked against the authenticated subject, or else it
is marked as absent. Above implementation helps to automate the subject tracking
mechanism with fraud tolerance. Figure 2 demonstrates the system workflow by
using biometric facial recognition along with the RFID authentication to track more
efficiently.
The functionality of the system, includes registration, near real-time tracking, are
as follows.
A. Student and Faculty Registration Platform
During admission of a student in an educational institution, registration form and
process is explained to register into the system. During the process, the institu-
tion will issue an RFID Tag whose unique code data set saves into the database
concerning its name. The student has to undergo through the biometric fingerprint
scanner where fingerprint information stored in the database respective to their terms
at the same time. The exact process is followed by the faculty members of the insti-
tution to register themselves with necessary information into the system. The faculty
Human Tracking Mechanism for Institutions Using RFID and Facial … 359

Fig. 2 Processing system

members can search for the specific student either by the name or ID. The system
automatically calculates the attendance percentage as per the pre-configuration set
by the authorized member of the institution. The teachers can later check the atten-
dance percentage of the student for a particular period of days. It is a mandate for
faculty members to update the courses related information into the system regularly.
Different authorization roles are created while setting up the system where the faculty
team needs to update their profile, attendance and leave base information.
Figure 3 demonstrates the data link of the process flow or system. Another critical
functionality in the system that the student who registers for the different semester
can track their daily class routine for the ongoing courses.
The administration of institution can also frequently monitor their faculty
member’s performance metrics which is calculated by the system automatically based
on the student presence corresponding to each course. In this way, institutions can
improve the quality of education for students.
B. Near Real-Time Tracking
The tracking mechanism/framework can mark the daily attendance of enrolled
students, teachers, and other institutional staff automatically without any manual
intervention. The RFID readers are installed at various places like entrance, exist
gates and hotspots of the institution. Whenever a student enters the premises, the
RFID automatically reads the unique code of RFID passive tag from their identity
card and data is transferred to the server and get saved in the database. When the
student present for an enrolled course, the RFID reader reads the unique code of the
tag and cross-checks that information present in the system database. The checks
against the attributes like semester number, subject code, classroom information and
360 R. Shaik and L. V. Patil

Fig. 3 Data flow diagram

so on. After scanning the RFID, student photo is captured, and the image information
is transfer later to the server for processing the image against the database copy. If
the match is found, then the presence of student for that enrolled course is marked as
a success. The facial recognition device will only be active at the scheduled time the
course. The teacher who is delivering the lecture can decide on the activation time
of the system where he can set the rule to mark the attendance as absent as penalty
if the student visits the class late base on the threshold value.
The installation of the system completes at a faster pace due to less complexity.
The facial recognition unit can install on the entrance of each classroom but not at all
places, which can help to reduce the cost for the institution. The information which
is present in the database can later put in use to analyse the data to take the necessary
measure to improve the quality of education for students.
C. Mathematical Expression/Algorithm
A model can explain its effects on a different level and predict the correct implemen-
tation of the system. The following equation could show the mathematical expression
of Face validation:

X = { , E, δ, D}.
Y = Face Recognition.

= set of input values = {audio content, video content, characters}.
F = set of output symbol content = {Match concerning inform r, Match Not Found}.
Human Tracking Mechanism for Institutions Using RFID and Facial … 361

– Starting the flow

– No. of training dataset N * N images
– Resize the dimensions of the image to Nˆ2 * 1
– Test set of Nˆ2 * M Dimensions, M: number of images used for testing
– Finding the average faces, subtract from the faces in the training sets, create matrix
A.

1 
M
ψ= i
M i=1

where
 = average no. of images, M = number of image objects,
Gi = denotes the images vector. i = Gi − .
Where, i = 1, 2, 3… M.
A = [1, 2, 3…M].
– Computation of covariance matrix: (AA ).
– Compute eigenvectors of the c covariance matrix.
– Compute Eigenfaces = No. of training image − no. of classes (total number of
people) of eigenvectors.
– Create reduced Eigenface space. The selected set of vectors are then multiplied
by the A matrix to generate the reduced result set.
– Evaluate the Eigenface of an image object determines in the question.
– Next step would be the calculation of Euclidean distance for the Eigenfaces and
the image object.
– Now the action is to search the insignificant Euclidian distance.
– The output is as follows: Image object with the less Euclidian space or image is
not recognizable.
D = {the system cannot process the voice dataset, Eigen generates the grey image
object; the key algorithms will execute till frames.}.

4 Results and Discussion

This section depicts the outcome of the system which intends to track human partic-
ipation which can analyze the behaviour. There are many technologies which are
present for monitoring the students like manual, Bluetooth base, infrared, Wi-Fi-
enabled application. The solution is to create feasible and affordable devices which
can remove the corns of the previous implementation. The framework uses RFID
which can read several labels and process it in parallel, which is one of the most
crucial features. The RFID works on the theory named Automatic Identification and
362 R. Shaik and L. V. Patil

Data Capture (AIDC). AIDC functionality is to track objects automatically, collect

data inputs about them, and that information is transmitted to the computer system.
This method uses radio waves which transfer the data from RFID reader to RFID
tag.
The implementation of Facial biometric technique with RFID ensures that we
have a fail-proof system which can be trusted and works accurately. The position of
the camera directly impacts on the outcome due to various external factors like light,
the position of the object and distance of the image object as there is two module in
the framework which connects between each other by the application software. With
the help of application software, one can start, log, access the menu and mark the
attendance by selection attendance list. The enrolment process is designed to allow
individuals to enroll their details along with their images. The Facial recognition
system is intended for an individual to attest his/her presence in the enrolled course
but also track their presence across the campus.
Table 1 demonstrates the number of test cases performed to track student based on
the implementation of human tracking mechanism using RFID and Facial recognition
techniques. The initial test was conducted on 50 subjects to validate the effectiveness
of the identity device. For the first test run, the external parameters are set as accu-
rate as possible or at higher configurations like lighting condition, the distance of
image object, position of the image object. The device can identify 47 test subjects
successfully and only three items were unsuccessful to validate.
For the second test run, the external parameters are adjusted to the actual real
scenario of an institution classroom where 48 students are accurately identity from
a total of 50 test subjects. After analyzing the failed subject data logs, we identify
the root cause of the unsuccessful tracking is due to one of the external parameter is
time. The subject must spend a minimum amount of time near the tracking device to
get detected by the system.
Figure 4 illustrates, graphical presentation of the accuracy of subject tracking by
the implementation of a solution using RFID and Facial recognition. The improve-
ment metrics show, in the second test run, the accuracy was improved by two

Table 1 Tracking test scenario

Test No # Students Successful track rate Unsuccessful track rate Accuracy %
1 40 38 2 95
2 40 38 2 95
3 40 40 0 100
4 45 40 5 88
5 45 44 1 97
6 45 45 0 100
7 50 49 1 98
8 50 50 0 100
9 50 50 0 100
Human Tracking Mechanism for Institutions Using RFID and Facial … 363

Fig. 4 Tracking metrics in graphical representation

Table 2 Time elapsed for

Type of No of students Total time (s) Average time
tracking attendance
system (s)
Paper 50 580 11.6
based
NFC 50 300 6.0
cards
Facial + 50 437.11 8.7
NFC

percentage, which is 96% in total. After tweaking the external parameters, we see
the increase in ratio to 100 rates, also considering the time spent by the subject with
the tracking devices.
Table 2 demonstrates the elapsed time by the system concerning the previous
technology-based implementation. The average time consumed by the solution is
far less as compared to the survey performed on the other tracking technologies and
devices. The average time is 8.7 s as compared with 11.6 s, which consider as to be
the most promising results.

5 Conclusion

The implementation of the solution in educational institutions will increase the

chances of tracking and managing students without any manual intervention or
efforts. The adoption of the system can be suitable for addressing the issues related
to manual steps and computation for managing student’s presence in the enrolled
courses. The system is a web-based system that allows the lecturer to compute
student’s existence via the web browser. This system can be quickly deploy in a
various number of educational institutions. It can be scalable as per the need, which
364 R. Shaik and L. V. Patil

makes the affordable solution to the educational sector. The data captured by the
system can feed into various analytical tools to create different learning patterns as
well as a retention plan for the students. The data can also be utilized to develop
a predictive model of student behaviour which helps students and institutions to
provide better and quality education.

Acknowledgements I am incredibly thankful to my guide Dr. L V. Patil for suggesting the topic
for research and providing all assistance to complete the work. His guidance and discussion with
him are valuable in the realization of the work.

References

1. Srinidhi MB, Roy R (2015) A web-enabled secured system for attendance monitoring and real
time location tracking using biometric and radio frequency identification (RFID) technology. In:
International conference on computer communication and informatics, 978-1-4799-6805-3/15
2. Eridani D (2015) Simulation of attendance application on campus based on RFID (radio
frequency identification). In: Proceedings of 2015 2nd international conference on information
technology. 978-1-4799-986-3-0/15
3. Sundararajan M, Mahalakshmi M (2015) Tracking the student’s performance in web-based
education using scrum methodology. In: 2015 international conference on computing and
communications technologies. 978-1-4799-7623-2/15
4. Nandyal S (2015) An automatic attendance system using image processing. Int J Eng Sci. ISSN
2319 – 1813 ISSN
5. Noguchi S (2015) Student attendance management system with Bluetooth low energy beacon.
In: 2015 international conference on network based information system. https://doi.org/10.1109/
NBiS.2015.109
6. Benyo B, Doktor T (2015) Student attendance monitoring at the university using NFC. In: 2015
wireless telecommunication symposium. https://doi.org/10.1109/WTS.2015.6266137
7. Gunawan TS (2015) Design and development of portable classroom attendance system based
on Arduino and fingerprint biometric. In: The 5th international conference on emerging
technologies. https://doi.org/10.1109/ICT4M.2015.7020601
Monitoring Power Consumption
and Automation Using IOT

Riya Jain, Revati Awale, Neha Kothari, Swarali Shah, and Amit Kore

Abstract Today, the entire world is frequently facing a challenging and speedy
environment everywhere. The main problem we all are going through is the energy
crisis. Actual energy demand made on existing energy supply is Electric energy
consumption. Early India has recorded a swift growth in electricity generation since
1985 which is increasing from 179 TWh in 1985 to 1057 TWh in 2012. The rapid
increase of this came from non-conventional renewable energy resources (RES) and
coal-fired plants. In India gross utility of electricity generation was about 1384 billion
kWh in 2019–20 which represents 1.0% growth yearly as compared to 2018–2019. A
country’s electric power development is an important measure by annual electricity
consumption per capita. It reached 3084 kWh, up 42.3% from 1990. Sensor networks
have been incorporated in the management of buildings for organizations and cities.
In recent it has led to an exponential increase in the volume of data available, and
monetary savings. For this purpose, new approaches, alignments and techniques are
required to investigate and analyze information in big data environments. For this
problem, having a relevant system to monitor the power usage is the only solution.
This paper proposes an analytical and prediction model by using different energy
profiles, which will provide the user power consumption chart of a consumer over a
period of time, to perform quantitative analysis using smart meters that automatically
acquire context information. There are devices which are capable of measuring a
customer’s energy consumption, for example smart meters. In this paper, there are
two modules. The first one focuses on receiving the data from the smart meter and
also transfers it to the data analyst. The second module is the predictive module which
uses consumption data and information of the consumer in order to understand the
behavioral patterns of the consumption of electricity. These models can be used to
predict energy consumption and also identify irregularities and outliers. The customer
gets acknowledged about abnormal usage. On getting acknowledged the consumer
will get to know where and when exactly the power usage is increasing and can
control power consumption which will ultimately result in reduction of expenses.

R. Jain · R. Awale (B) · N. Kothari · S. Shah · A. Kore

AISSMS’s Institute of Information Technology, Pune, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 365
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_37
366 R. Jain et al.

Keywords Data analysis · Electricity prices · Power consumption · Smart meter

data

1 Introduction

Electricity is time dependent phenomena, it is generated, transmitted and distributed

to satisfy customer’s demands and needs in certain moments. Power is the soul of
world which is related to the electricity and “electricity” is the word which now
rules the world [1]. At distribution level power consumption in households has a
significant influence on total consumption. Huge amount of energy is consumed by
various large scale industries like steel, iron, petro-chemical etc. Power consumption
can be defined as power which is the rate of doing work or transferring heat that is
the amount of energy transferred or converted per unit time and consumption means
the action of using up the resources. Also, power consumption in electrical terms is
defined as the amount of energy used per unit time. It is important in industries and in
digital systems where we use electricity in huge volumes. Digital systems draw both
dynamic current and static current. Dynamic power is used to charge capacitance as
signals change between 0 and 1. Static power is used even when signals do not change
and the system is idle. Power consumption is usually measured in units of Watts (W)
or kilo-Watts (KW). Mathematically, it is the product of voltage and current supplied
to the appliance. Every appliance or electronic device that we use is reflected in
our monthly electrical bills. But, some of those devices cost us more than other
appliances. To figure out such devices analysis of power consumption is necessary.
With this technique we can understand which device has consumed how much power
and thus control the energy hogs and lower down our bills. Even small adjustments
can help whittle down our expenses. We only need two numbers to get started: The
device’s wattage and the number of hours we use it per day. It is also estimated that
the price of electricity and demand is going to increase in upcoming years. This
paper mainly deals with smart meter’s data, which utilizes the features of embedded
systems i.e. the combination of hardware and software in order to implement the
desired functionality. It also discusses the use of SmartMeter-Elite-440-445, ethernet
module and IOT Gateway with SIM-card. Smart meter is an electronic device which
records the consumption of electric energy and communicates the information to the
electric supplier for billing. Elite 440 is a multi-line-three phase panel for accurate
and reliable measurement of electric parameters. Using this we can achieve a cost
effective online monitoring. Similarly, IOT Gateway is a device which provides the
bridge between IOT devices in the field, cloud and other user equipment. It provides
critical functionality i.e. device connectivity. Whatever data is collected through
the smart meter is connected to the server or the cloud using IOT gateway. Hence
using these devices this paper basically aims to analyze the consumption of electricity
consumed by a particular appliance, which in order notifies the user about the increase
in consumption and thus reduces the bills and electricity consumed shown in Fig. 1.
Monitoring Power Consumption and Automation Using IOT 367

Fig. 1 Consumption of electricity

1.1 Factors Influencing Electricity Prices

Power sector in India is governed by the Ministry of Power Supply. It is not

only a question of electrical or financial engineering but also about informing the
user/consumer that how much they have consumed and the amount they need to pay
for the consumed power
1. Fuel: Cost of Fuel differs during time or span of high demand such as natural
gas which results in higher price for fuel and it gives elevated price of electricity.
2. Transmissions Systems: These huge systems provide power on a very large
scale which are able to repair the damage of the system from extreme weather
conditions.
3. Seasons: Usually in summers electricity prices tend to be higher, when there’s
more consumer demand (presumably for air conditioning). When people require
more power the generation and delivery of electricity increases with the cost
simultaneously.
The grid of the power reflects the real-time cost of supplied electricity as the price
and power supply of electricity differs minute by minute. As we discussed earlier
that demand of electricity is mostly high in afternoon and evening which is directly
proportional to the cost. Many consumers pay costs according to the seasonal average
cost, they do not experience such types of fluctuations. So, the problem of electrical
use may increase and we may not use less electricity in our day to day life but we have
to access some features which helps us to reduce the amount of money. People can
save money by using power saving products. The electricity that moves is optimized
368 R. Jain et al.

by these power saving modes, makes sure that whether it is not wasted and that all
is utilized by the consumers.
Various Factors influencing Electricity Prices are: Fundamental Factors, Fuel
Prices Temperature, Weather Conditions Time indices such as day of the week, month
of the year, season of the year Cost of production of electricity per unit Operational
Factors Power Transmission Congestion Power System Operating condition, Elec-
tricity production (deficit/surplus), Network Maintenance, Electricity Load, Strategic
Factors Power Purchase Agreements, Bilateral Contracts, Power Exchange, Bidding
Strategy, Market Design, Historical Factors, Price, Demand.

2 Literature Survey

Monitoring and Control of Electricity Consumption Using Raspberry Pi

Through IoT: In this paper researchers tested functionality of the system by using
different appliances. They were able to design a system which allows the user to
monitor home energy consumption. The system is connected to the database where
users can view the past readings of energy consumption. As it monitors and controls
electrical lines it does not analyse the data and predict the data consumed.
Big Data Analytics for Discovering Electricity Consumption Patterns in Smart
Cities (Ruben Perez-Chacon, Luna Romera): In this system the model based on the
k-means algorithm was designed for the purpose of using the distributed computing
advantages of Apache. The studies about their CVI’s optimized for parallelization-
the DB Dunn was carried out. From these numbers majority voting strategy was
applied in order to choose optimal number clusters. In this system the depth analysis
of clusters was performed but they were not analysed daily or hourly basis.
Energy Consumption Analysis Based on Energy Efficiency Approach (A. M.
Leman, M. Faris Mubin). In this paper the author have studied the detailed suffi-
cient data about power consumption and analysed it in order to develop appropriate
electricity-saving measures. Their main motive is to increase awareness about EE
(Energy Efficiency) approach and the energy usage of each house/business. So the
implementation part is focused on how to reduce consumers monthly electricity bill
by EE approach.
Survey and Analysis of Energy Consumption in Universities Campuses (Jae
Woong Jung). This paper surveyed and analysed energy utility in universities as
leading research to reduce power consumption. The data were analyzed by dividing
it into general information related energy consumption and energy consumption data.
Then this data was derived from the uniform range due to differences in the range
of available campus data. The data range and reliability are secured for the monthly
electricity consumption pattern analysis were selected and analyzed.
Monitoring Power Consumption and Automation Using IOT 369

Advanced IoT-Based System for Intelligent Energy Management in Buildings

(Vangelis Marinakis, Haris Doukas). In this system implementation of web based
and integrating the above mentioned architecture has been done. Tool has performed
a very important role of immediate and complete virtual distribution on the Internet
of the energy consumption in buildings and commercial buildings. Users might get
updated on the energy consumption and other indicators like energy, CO2 emissions
on the website. The system analyses data amongst four major groups and displays
the result.
Average Power Consumption Estimation and Momentary Power Consumption
Profile Generation of Softcore Processor (Berna Ors). In this module post imple-
mentation simulation and power consumption was performed using different outputs
and different work environments. A momentary power consumption profile genera-
tion method has been introduced to be later used in channel attack resistance analysis,
with some graphical results presented.

3 Proposed Methodology

Figure 2 explains us the flow of our project. The Smart Meter Data is stored in the
form of an excel sheet. The data is programmed to be received by the database on
every minute basis. The data sending procedure is called the auto generator in the
diagram. The data is then received by the Mysql database and stored. As the location
to store data is the database, no extra technique or method is used to do the process
[2]. The data is then validated and cleaned. Based on this data, the prediction on
future consumptions is done. In this system, one can get predicted data within a span

Fig. 2 Flow chart

370 R. Jain et al.

of a week. The predicted data directly shows the overall expected consumption in
units and in price. A user can store much wise data upto a month in the system. So
that the user can easily compare the electricity bill he paid and the bill on the app.
A report is generated of electricity consumed allowing the user the choice to select
the timespan of the report. The entire data is easily accessible and visible on the app.
The app, made in android studio, help the user to view data in graphical form and the
price of consumption [3]. The app is the interface to view the graphical form of the
content stored in the database. The Smart meter data used here is from a pre-installed
meter which reports reading on a regular basis of one minute. This data is in excel
sheet format. The sheet involves many parameters of electricity, like, voltage, current,
kvarh, kWh, pf and many more. This data with its every minute detail is being stored
in our database of application. The database is stored in the system. The database is
accessed using MySQL. There are 2 tables in the database, one of the users details,
second of the every minute detail of electricity. These use simple and few complex
queries of MySQL. The app also provides a feature of viewing prediction to the
user. The prediction is available for all parameters as well as for total consumpy and
bill. The restriction for prediction is upto 5 days. Prediction is basically a method
of forecasting based on data analytics which requires a trained model. The model of
prediction needs to be trained on the data set. The data set is then divided as a test
data set to check how appropriate it predicts. Various techniques are used and the
one which gives the maximum percentage of approximate value is usually chosen.
As prediction is only the forecast based on the past data, it shouldn’t be considered
that this is the exact value one will get in future as it could be influenced with many
unknown and unexpected factors. The interface of the app shows many pages which
can give a good idea of data. The main page has graphical bar representation of the
current data. The graph is the price of the consumed electricity bill along with the
units consumed. Next page has the historical data. It shows all the past consumption
a user has done right from the data has started to store. It shows data in parameters.
Each parameter is shown with its consumption till date. Next page is the prediction
page where the user can watch his prediction of the selected date. Each parameter
prediction is visible to the user. Next page in the app is report generation. This
feature helps the user to compare or see the record of his consumed electricity. The
generation of reports requires starting and ending dates. The total consumption unit
wise and price wise is visible to the user as it is in the electricity bill paper. This
will help the user to compare the bill received from MahaVitaran and the one the
app shows. The entire functioning of the system as a prototype is done using xampp.
When the xampp MySQL servers are switched on, the device like mobile or laptop
can access it using the localhost. For localhost connection all devices need to be in
the same network/router. Only then the devices will be able to access the database on
the server. A virtual network interface is created from which the user can request data
using the port 80. The virtual network interface revives the request and then sends
the request to the local web server services. The request is answered back in the same
flow, i.e. the data requested is sent in reverse flow. The tool used for communication
is the IP address. In this prototype, the above said method is used as developers use
the same technique for testing applications which work over the internet.
Monitoring Power Consumption and Automation Using IOT 371

4 Result

Then the user selects the time scale and finally a graph according to his requested
data is plotted. For a clear understanding we can refer to the pictorial representations
shown. The system also enables the user to keep a check on the past consumptions.
Users can also predict their consumption for the upcoming days and can get a detailed
report of any day including prices of their consumption.

5 Graphical User Interface

Here, in the screenshots, Fig. 3 is the signin/signup page where the user has to enter
his information to login in into his account. The user has a choice to select the
page that should be visible. Figures 4, 5 and 6 show the data stored in the database
according to the day. Figure 7 shows the historical data stored in the database in bar

Fig. 3 Sign-up, login and navigation option

372 R. Jain et al.

Fig. 4 Present day’s graph

Fig. 5 Yesterday’s graph

Monitoring Power Consumption and Automation Using IOT 373

Fig. 6 Tomorrow’s graph

chart form. Also Fig. 8 shows the facility of comparing 2 parameters of electricity
as a time. Figure 9 shows the prediction of the selected date to the user, which shows
prediction of all parameters. Figure 10 shows the report generation page where the
user can fetch his data of electricity consumption selecting from date and last date.
The generated report is in PDF format.
374 R. Jain et al.

Fig. 7 Historical data

comparison(a)

Fig. 8 Facility of comparing

2 parameters of electricity as
a time
Monitoring Power Consumption and Automation Using IOT 375

Fig. 9 Prediction of selected

date

6 Conclusion

Smart meters are used to produce considerable volume of data presenting the oppor-
tunity to enhance utility end customer service, improve energy efficiency and lower
the cost and to reduce the bill and save energy for consumers. The system aims
to analyze the data, understand the energy profiles and appliances that affect the
consumption profiles to group similar consumers together. Setting standards for effi-
cient energy consumption is aided by the analysis and it also identifies the outliers
and irregularities in the consumption patterns. The system also calculates the energy
consumption of devices and even makes the energy unit reading to be handy. Hence
it also reduces the wastage of energy and brings awareness among them. Even it can
reduce manual intervention. The proposed system will overall help all the industries
to lower down their electricity bills and will give a detailed analysis of the electricity
consumed by appliances during the particular duration of time.
376 R. Jain et al.

Fig. 10 Report of selected

date

References

1. Nezhad AJ, Wijaya TK, Vasirani M, Aberer K. SmartD: smart meter data analytics dashboard
2. Medina CC, Joe Mel U. Monitoring and control of electricity consumption using Rasberry-Pi
through IoT
3. Borle P, Saswadhar A, Hiwarkar D, Kali RS (2013) Automatic meter reading for power
consumed. Int J Adv Res Electr Instrum Eng 2(3):982–987
Light Fidelity (Li-Fi): Future 5G Wireless
Connectivity to Empower Rural India

Prajakta A. Satarkar and Girish V. Chowdhary

Abstract Visible Light Communication (VLC), also known as Optical Wireless

Communication (OWC) has emerged as a viable candidate with a wide range of
applications due to its promising features like license-free channels, large band-
width, immunity to interference and security and, less power utilization. Light
Fidelity is a potential application of OWC mainly used for indoor applications to
provide high-speed internet up to 10 Gbps. Shadowing effect and data loss in Li-Fi
is a major concern that can be solved by using the hybrid network and different
methods like access point selection (APS) and resource allocation (RA). In this
paper, we have discussed a hybrid network of Li-Fi and carried out a comparative
study of different techniques used for access point selection and resource allocation.
Finally, we concluded that Reinforcement learning provides better throughput with
less computational complexity.

Keywords Access point selection · Exhaustive search · Fuzzy logic · Li-Fi ·

OWC · Reinforcement learning · Resource allocation · VLC

1 Introduction

The congested radio frequency spectrum is the main restraint on the increasing
demand for connectivity with high speed. The latest Cisco Visual Networking Index
predicts projects global IP traffic also called as web traffic to approximately triple
from 2017 to 2022. The expected growth of overall IP traffic is 396 EB per month
up to 2022. In 2017, overall IP traffic observed was 122 EB per month. It is esti-
mated to increase with a Compound annual growth rate (CAGR) of 26% [1]. This
increase in CAGR is due to the growing share of mobile traffic over the total IP traffic.

P. A. Satarkar (B)
Computer Science and Engineering Department, Swami Ramanand Teerth Marathwada
University, Nanded, MH, India
e-mail: [email protected]
G. V. Chowdhary
School of Computational Sciences, Swami Ramanand Teerth Marathwada University, Nanded,
MH, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 377
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_38
378 P. A. Satarkar and G. V. Chowdhary

This rapid increase in the number of mobile devices and growth in social services
like Facebook, Video on demand and Twitter, etc. are the primary reason for this
tremendous growth in mobile traffic. Along with the limitation in spectrum band-
width, interference due to electromagnetic wireless devices and power inefficiency
and security are other important issues in RF wireless communication. To overcome
these drawbacks, Visible Light Communication (VLC), also called as Free Space
Optics (FSO) networks could become a promising candidate.
The visible spectrum has a wide range for wavelengths from 380 to 750 nm
and supports a wide frequency range of 430–790 THz [2]. The advantages of Li-Fi
network over RF are outlined below [3]:
• License-free unused spectrum
• Immune to electromagnetic interference
• Support broadband, peer to peer communication
• High bandwidth
• Low power consumption
• Inexpensive optical components
• Secure as light cannot penetrate through walls.
VLC has a wide range of applications which include indoor Li-Fi network to
provide high-speed internet up to 10 Gbps, vehicle to vehicle communication to
avoid road accidents, in hospitals and in aircraft to avoid interference with RF signals
of other machines, underwater communication for submarine or remotely operated
vehicles and signboards in the areas like railway station, traffic, airports.
With all the above advantages and applications of VLC, some challenges are
observed during the actual implementation of VLC. They can be described as:
(a) Interference with normal light sources
(b) Data loss in case of shadowing
(c) Fluctuation and dimming of light
(d) Integration with other available technologies.
In this paper, we start with a survey of the potential of Li-Fi, working principle,
system overview, indoor applications, and challenges are discussed. The remaining
paper is arranged as follows. We start by providing system components and system
overview in Sect. 2. In Sect. 3 hybrid Li-Fi network is discussed with designing
issues which are followed by Applications of Li-Fi in Sect. 4. Section 5 contains
Challenges and concluding remarks.
Light Fidelity (Li-Fi): Future 5G Wireless Connectivity … 379

2 System Components and Overview

2.1 System Components

VLC uses intensity modulation and direct detection for information transmission.
Li-Fi uses the physical layer of VLC for communication. This section is composed
of an overview of the VLC system with its two primary components: transmitter and
receiver.
1. VLC Transmitter
In VLC, the transmitter consists of single or multiple LED lamps, a driver circuit,
modulation techniques, and other components as shown in Fig. 1 [4]. For example,
during transmission, simple On–Off Keying modulation uses two separate levels of
light intensity representing the data bit “0” and “1”. LED serves two purposes illu-
mination and communication. While designing care has to be taken that illumination
should not get affected by communication. The performance of the VLC system is
dependent on the design of LEDs. White light can be generated with three different
LEDs: The dichromatic, tri-chromatic, and Tetra-chromatic modes [2]. RGB combi-
nation is preferred for communication as it can use Color Shift Keying to modulate
the data with three LEDs using different color wavelengths.

Fig. 1 VLC system components

380 P. A. Satarkar and G. V. Chowdhary

2. VLC Receiver
The VLC receiver has 3 components: optical amplifier circuit, optical concentrators,
and filter as shown in Fig. 1 [4]. The optical concentrator is used to compensate for
the beam divergence of LEDs which results in attenuation in the case of a large area.
VLC receiver is responsible for light detection and conversion into photocurrent. In
VLC we can use silicon photodiode, PIN diode, or avalanche photodiode as receiver
[5]. After a comparison between these three types, it is observed that cost and gain
are higher for avalanche photodiode than a PIN photodiode. To resolve the inter-
ference effect in VLC from sunlight and other light sources, optical filters should
be introduced in the design. Two types of receivers to receive the signal from the
transmitter:
(1) A Photodetector: It is also referred to as a photodiode or Non-imaging receiver.
This device converts the received light signal into an electric current. It is better
suited for stationary receivers.
(2) Imaging Sensor: It is also called a camera sensor used to receive the transmitted
light signals and has a larger Field Of View (FOV) in the case of mobile
applications. They are slow and energy expensive.
Based on the type of application we can choose the proper type of VLC receiver.

2.2 System Overview

Depending on the transmitters and receivers location and network range supported,
these networks can be categorized into:
(i) Satellite Networks (SNs)
(ii) Outdoor Networks (ONs)
(iii) Indoor Networks (INs).
Indoor networks are applicable for wireless communications inside offices, build-
ings, and houses. Usually, a base station is connected with multiple users through
optical wireless links i.e. Light Emitting Diode (LED). For indoor communication
OWC links are divided into two types based on the way of propagation:
1. Line Of Sight (LOS) links: An LOS link requires both the transmitter and
receiver should be inline. It is again subdivided into three types: (a) Both the
transmitter and receiver face towards each other as shown in Fig. 2a [6]. This
method improves power efficiency and resists to distortion caused by ambient
and artificial light sources. (b) The transmitter and receiver are not facing towards
each other as shown in Fig. 2b [6]. Wide beam transmitters and wide FOV
receivers are required for the transmission. It needs high power levels to cope up
with the high optical loss and the multipath-induced distortions. Here multipath
fading issue is resolved. (c) Hybrid design method where the transmitter and
receiver can have different levels of directionalities, such as a narrow beam
Light Fidelity (Li-Fi): Future 5G Wireless Connectivity … 381

Fig. 2 Propagation methods in VLC

transmitter which is not aligned with a wide FOV receiver as shown in Fig. 2c
[6]. LOS links achieve higher capacity because of a better power budget and the
absence of multipath propagation effects.
2. Non-LOS (diffused) links: In the Non-LOS links, signals transmitted from the
source do not directly arrive at the receiver. They are reflected from different
surfaces like walls, ceiling, floor, or other objects and arrive at time intervals to
the receiver. This causes multipath distortions and makes the path loss estimation
more difficult. The Non-LOS architecture with non directed transmitter–receiver
pair as shown in Fig. 2f [6] is called diffused systems which is the most robust
system and easy to implement for mobile communication systems. Non-LOS
links are more robust when coming across obstacles in the light path.

3 Hybrid Li-Fi Network

In an indoor environment, Li-Fi can be used for localization and communication

between multiple users. As light cannot penetrate through walls it provides better
security as compared with Wi-Fi. But, if an opaque object comes between trans-
mitter and receiver, due to blockage, throughput gets affected. This issue can be
solved using a combination of Li-Fi and Wi-Fi networks which can be referred to
as a hybrid network. They can co-exist due to different spectrum range. A hybrid
network gives a good data rate with ubiquitous coverage as compared to a single Li-Fi
or Wi-Fi network. For hybrid network, Access Point Selection (APS) and Resource
382 P. A. Satarkar and G. V. Chowdhary

Allocation (RA) are major concerns as Wi-Fi gets overloaded whereas Li-Fi near it
remains underloaded. Many researchers have focused on these issues and came up
with different solutions. APS and RA are resolved using 3 approaches: fuzzy logic,
optimization, and machine learning. Aspects like handover and mobility are consid-
ered in [7] to improve system throughput. Further to optimize the throughput, [8]
used load balancing, and power allocation is considered in [9] to improve energy effi-
ciency. Demir et al. [10] used particle swarm optimization but all of these optimization
techniques introduced high computational complexity.
Fuzzy logic techniques are implemented in [11] to improve data rates with a
dynamic load balancing approach. In [12] APS is carried out in 2 steps, the first
user gets associated with Wi-Fi and the remaining users to Li-Fi. It is observed that
similar results are obtained with quite less computational overhead.
In [13] Reinforcement learning with knowledge transfer is implemented to handle
network selection. The decision probability distribution is used in [14] for APS. In
[15], the Reinforcement learning load balancing approach is used by considering the
following parameters: throughput, user satisfaction, and outage probability along
with computational complexity. Trust Region Policy Optimization (TRPO) is used
in this paper to find an optimal policy at each step.
The comparative study for average network throughput and computational
complexity is done for all different methods for APS and RA which is presented
in Table 1.
Here the worst case performance in terms of computational complexity (Big-O
notation) is compared for four types of methods, namely, SSS, Iterative, Reinforce-
ment learning and Exhaustive search. NAP represents number of access points, Nu
represents number of users, I represent number of iterations for iterative method.

Table 1 Comparison of APS

Sr. No. Method Throughput Complexity
methods for average network
(Mbps)
throughput and complexity
1. Strongest signal 8.2 O(NAP Nu )
strategy
2. Load balancing 9
3. Fuzzy logic SSS 9.4
4. Fuzzy logic LB 9.5
5. Instantaneous LB 10.88
6. Iterative 12.3 O(NAP Nu I)
7. Mobility aware 13.41
LB single Tx
8. Mobility aware 14.16
LB multiple Tx
9. Reinforcement 15.53 O(NAP Nu 2 +
learning Nu 2 + NAP Nu )
10. Exhaustive 17 O((NAP )Nu )
Light Fidelity (Li-Fi): Future 5G Wireless Connectivity … 383

4 Applications of Li-Fi

Li-Fi applications are vast as they are dependent on the key characteristics, such as
EMI free, energy efficiency, security, improved data rate, availability, license-free,
and integrated networking capability. We can use Li-Fi for satellite communication,
dense urban environment, aircraft and petroleum stations, Outdoor communication
like vehicle and underwater communication, indoor communication for localization,
augmented reality, and cellular communication.

5 Challenges and Conclusion

Based on the literature review we will discuss the important challenges that can
be open issues for the researchers. The solution to these issues will help in the
deployment of Li-Fi practically in near future.
A. LOS Alignment: To maximize the channel response and achieve high data rates,
the transmitter and receiver should align their FOV. However, practically due
to mobile user position and orientation of the receiver may get changed. So
the receiver’s FOV will not be aligned with the transmitter. It causes a drop in
received optical power. Designing the techniques to handle the LOS problem
is extremely challenging and is an important direction for future research.
B. Shadowing: When any object or mobile user blocks the LOS, an immediate
reduction in the observed optical power is witnessed; this causes data rate
reduction and data loss. So a large scope is there to understand and work on the
phenomenon of shadowing in VLC. As the shadow effect is for a small duration
of time we have to utilize the reflected optical power promptly.
C. Design and Energy-Related Issues of Receiver: For receiving a light signal in
current VLC we either use a photodiode or an imaging sensor. For the stationary
users FOV can be fixed using photodiode but for mobile users FOV of the
imaging sensor is sufficiently large for communication. Due to a large number
of photodiodes imaging sensors can become slow and consume a large amount
of energy. Hence the achievable data rate will get reduced. Thus, it is challenging
to design a receiver that can give a high data rate while moving and should work
with low energy consumption.
D. LED to Internet Connectivity: LEDs are required to connect to the internet
for broadband applications. Large numbers of LEDs are used for illumination
and connecting all of them to the internet via Ethernet is a costly process. The
advantage of reusing the LED infrastructure for communication is no more
useful. A large number of LEDs cause interference, which hampers the achiev-
able data rate. Power line communication is suggested [16] which provides
available power lines for communication. It only introduces the requirement of
Ethernet to power the modem and power to the VLC modem. The performance
and coverage issue [17] is resolved in this method.
384 P. A. Satarkar and G. V. Chowdhary

E. Dimming Control: During the daytime, if LEDs are switched off then no
communication is possible. LED should be “on” to use the communication
link. Hence similar to RF communications, power consumption for data trans-
mission is not free. The solution can make the LED brightness level low so that
LED is considered as “off”. Further work is required to solve this problem.
F. Inter-Cell Interference: If within a small cell area, LEDs are deployed densely
then it certainly leads to inter-cell interference. This reduces performance due to
low SINR. The solution to this problem can be network MIMO or rearrangement
of LEDs. In the network, MIMO LEDs synchronize the transmissions to ensure
high SINR. In the rearrangement method, LEDs are arranged to get minimum
interference. Research is required for design and analysis to get an optimized
solution.
G. Uplink Issues: More focus is given on the downlink performance problem but
less attention is given to the uplink problem. Mobile users with a smartphone
can use imaging sensors to downlink effectively. But for uplink, if a flashlight
or notification indicator is used then it will cause battery consumption with
user visual irritation. Also, the alignment and orientation of the smartphone are
important for the uplink connection. The solution is, to use RF for uplink [18].
It creates the issues of hybrid VLC with Wi-Fi.
H. User Mobility: For a mobile user the coverage of the VLC network is required
at every point to provide a high data rate and avoid interruption in indoor
communication. The link-layer techniques are designed to facilitate data rate
adaptation, frame aggregation, etc. If RF is used along with VLC then handover
related issues will arise.
I. Security: Light cannot penetrate through walls but the light leaking from
windows, doors, etc. will become a concern for security issues.
J. Interference from Sunlight: For outdoor applications, this problem is associated
with transmission beams which cause low SNR. The solution can be optical
filters but more research is still going on to solve this issue completely.
This survey paper has provided an introduction to visible light communica-
tion, its merits over the current RF system, and the need for VLC to facilitate the
increasing demand for data rate. The shadowing effect followed by the data loss
can be handled effectively using different strategies for APS and RA in a hybrid
network. The comparative study shows SSS gives the weakest performance while
Exhaustive search gives the best performance with complexity overhead. RL gives
better performance with moderate complexity.
Light Fidelity (Li-Fi): Future 5G Wireless Connectivity … 385

References

1. Cisco (2019) Cisco visual networking index: forecast and methodology, 2017–2022, 27 Feb
2019
2. Khan LU (2017) Visible light communication: applications, architecture, standardization and
research challenges. Digit Commun Netw 3:78–88
3. Son IK, Mao S (2017) A survey of free space optical networks. Digit Commun Netw 3:67–77
4. Lee CG (2011) Visible light communication, advanced trends in wireless communications.
In: Khatib M (ed) InTech. ISBN: 978-953-307-183-1. Available from https://www.intechopen.
com/books/advanced-trends-in-wireless-communications/visible-light-communication
5. Kharraz O, Forsyth D (2013) Performance comparisons between PIN and APD photodetectors
for use in optical communication systems. Opt Int J Light Electron Opt 124(13):1493–1498
6. Pathak PH, Feng X, Hu P, Mohapatra P (2015) Visible light communication, networking, and
sensing: a survey, potential, and challenges. IEEE Commun Surv Tutor 17(4):2047–2077
7. Wang Y, Haas H (2015) Dynamic load balancing with handover in hybrid Li-Fi and Wi-Fi
networks. J Lightw Technol 33(22):4671–4682
8. Wu X, Safari M, Haas H (2017) Joint optimization of load balancing and handover for hybrid
LiFi and WiFi networks. In: Proceedings of IEEE wireless communications and networking
conference (WCNC), Mar 2017, pp 1–5
9. Kashef M, Ismail M, Abdallah M, Qaraqe KA, Serpedin E (2016) Energy efficient resource
allocation for mixed RF/VLC heterogeneous wireless networks. IEEE J Sel Areas Commun
34(4):883–893
10. Demir MS, Sait SM, Uysal M (2018) Unified resource allocation and mobility management
technique using particle swarm optimization for VLC networks. IEEE Photon J 10(6):1–9
11. Wang Y, Wu X, Haas H (2016) Fuzzy logic based dynamic handover scheme for indoor Li-Fi
and RF hybrid network. In: Proceedings of IEEE international conference on communications
(ICC), May 2016, pp 1–6
12. Wu X, Safari M, Haas H (2017) Access point selection for hybrid Li-Fi and Wi-Fi networks.
IEEE Trans Commun 65(12):5375–5385
13. Wang C, Wu G, Du Z (2018) Reinforcement learning based network selection for hybrid VLC
and RF systems. In: Proceedings of MATEC web of conferences, vol 173. EDP Sciences, Les
Ulis, p 03014
14. Wang J, Jiang C, Zhang H, Zhang X, Leung VCM, Hanzo L (2018) Learning-aided network
association for hybrid indoor LiFi-WiFi systems. IEEE Trans Veh Technol 67(4):3561–3574
15. Ahmad R, Soltani MD, Safari M, Srivastava A, Das A (2020) Reinforcement learning based
load balancing for hybrid LiFi WiFi networks. IEEE Access 8:132273–132284
16. Ma H, Lampe L, Hranilovic S (2013) Integration of indoor visible light and power line
communication systems. In: Proceedings of 17th IEEE ISPLC, Mar 2013, pp 291–296
17. Yousuf M, El-Shafei M (2007) Power line communications: an overview—part I. In:
Proceedings of 4th international conference IIT, Nov 2007, pp 218–222
18. Rahaim M, Vegni A, Little T (2011) A hybrid radio frequency and broadcast visible light
communication system. In: Proceedings of IEEE GC Wkshps, Dec 2011, pp 792–796
Optimized Dynamic Feature Matching
for Face Recognition

Ganesh Gopalrao Patil and Rohitash Kumar Banyal

Abstract Since the last three decades, face detection and recognition have become
very active and a huge part of image processing research. In real-time applications
like video surveillance, front views cannot be guaranteed as input. Hence the failure
rates can degrade the performance of the face recognition system. The proposal aims
to introduce a novel PFR method termed as DFM that combine Sparse Representation
Classification (SRC) and FCN for resolving the partial face recognition issues. As
the major contribution, this proposal aims to tune the sparse coefficient of DFM in
an optimal manner, such that the reconstruction error should be minimal. Moreover,
this proposal introduces Jaccard Similarity Index measure to calculate the similarity
scores among the gallery sub feature map and probe feature map. For optimization
purpose, this work deploys a hybrid algorithm that hybrids both the concepts of Grey
Wolf Optimization (GWO) and Sea Lion Optimization (SLnO) algorithm.

Keywords Grey Wolf Optimization · Sea Lion Optimization algorithm · Sparse

Representation Classification

1 Introduction

“Face recognition is the process of identifying the faces from the images during
human computer verification” [1]. Under the controlled condition, the face recog-
nition algorithm [2, 3] obtains more optimal results, whereas, under uncontrolled
condition, various issues are faced during the recognition process. Mostly, the face
recognition techniques [4, 5] are classified into two types based on the occlusion
problem: they are part-based and holistic. The human face [6] is presented in the
captured image and it focuses on noisy environment as non-frontal pose, shading,
facial association and overexposing, thus it leads to a problem known as partial face

G. G. Patil (B)
Department of Computer Science and Engineering, SVERI’s College of Engineering, Pandharpur,
India
R. K. Banyal
Department of Computer Science and Engineering, Rajasthan Technical University, Kota, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 387
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_39
388 G. G. Patil and R. K. Banyal

recognition [7]. Traditional face recognition systems usually fail to address the prob-
lems of partial face detection. Moreover, the partial face recognition system does
not provide proper run time effectiveness. Introducing a realistic face recognition
system [8] can handle partial faces directly without any synchronization, and it will
be resilient to occlusions, and overcomes the issues regarding pose and illumination.
Partial faces can be determined from the images and accordingly, detection can
be effectively performed from various facial segments in to a complete face [9]. The
important concern is to develop the advanced face detectors for the identification of
occluded and partially noticeable faces. The similarity [10] among the probe partial
faces and gallery faces can identify the improved partial faces and hence the face part
is occluded. To identify the facial landmarks [11], the alignment method is provided
and then occluded.
The conventional face recognition methods use the deep learning methods by
the Convolutional Neural Network (CNNs). The benefits of deep learning method
require trained datasets to identify the better characteristics of data representation
[12]. The collection of large-scale datasets of faces including real-world variation
provides the faces-in-the wild on the web. The trained datasets in CNN based on face
recognition methods obtains better accuracy so that one can know the features [13,
14] whether they are robust in the real-world conditions. The exploitation of deep
learning approaches based on computer vision assists the facial recognition task in
an effective manner [15]. However, certain machine learning approaches like Fully
Convolutional Networks (FCNs) does not contain a permanent input as in the spatial
feature maps similar with the input size of images [16]. Therefore, it turns out to be
essential to develop a more advanced PFR system for detecting the partial faces in
an optimal manner [17].

2 Related Works

In 2018, Lahasan et al. [18] proposed an Optimized Symmetric Partial Facegraphs

(OSPF) for identifying the facial expressions of the human faces. Here, the hybrid
intelligent single particle optimizer and improved harmony search algorithm (HSA)
has offered better searching capabilities for identifying the optical landmarks and
moreover, the facial occlusion problems were investigated and resolved in an effective
manner. In 2019, Mahbub et al. [19] introduced “Deep Regression-based User Image
Detector (DRUID)” algorithm for detecting the partial faces. The proposed method
was compared to three other traditional facial methods and hence the classification is
done based on binary values. In 2018, Greening et al. [20] have implemented a linear
SVM algorithm for detecting the facial expressions in humans. Here, contestants with
partial face samples containing five emotional categories were taken into account for
a rapid “event-associated” analysis. In 2018, Duan et al. [21] proposed a “Topology
Preserving Graph Matching (TPGM)” technique to detect the features present in
the facial images. The proposed method was provided with higher order structure
and it included partial face recognition methods like “Robust Point Set Matching
Optimized Dynamic Feature Matching for Face Recognition 389

(RPSM) and Multi-Key point Descriptor with Gabor Ternary Pattern (MKD-GTP)”.
In 2016, Weng et al. [22] have implemented a new PFR method for finding the
concentration of an individual from their partial faces. “Robust Point Set Matching
(RPSM)” method was presented to match its geometrical and textural information
of the facial characteristics. In 2019, He et al. [23] have proposed Dynamic Feature
Matching (DFM) with Sparse Representation Classification (SRC) and FCNs for
resolving the issues persisting in partial face detection. Here, the DFM does not
necessitate the previous information regarding the position of partial faces. More-
over, the feature maps were computed from the whole image and it resulted in a
speedy process. Finally, the experimental outcomes have shown better effectiveness
and advantages over the traditional models. In 2016, Lei et al. [24] have presented
efficient 3D face recognition methods for identifying all the challenges in PFR. The
proposed method introduced the “Two-Phase Weighted collaborative Representation
Classification (TPWCRC)” framework for rectifying the issues found in the samples
images. The proposed model solved the following parameters like lost data, defor-
mations, occlusions and accessibility of limited training samples and thus it offered
higher efficient results. In 2019, Aminu and Ahmad [25] have established a “Locality
Preserving Partial Least Square Discriminant Analysis (LPPLSDA)” with conven-
tional “Partial Least Squares Discriminant Analysis (PLS-DA)” for face recognition.
The LPPLS-DA model hold better capability of capturing local structures and thus it
was found to be well suited for detecting the facial images. Finally, the outcomes of
proposed method were provided with improved consistency and reduced computa-
tional complexity. In 2019, Patil and Banyal [26] have presented techniques of deep
learning for image recognition. In 2020, Patil and Banyal [27] have presented uncon-
strained feature matching technique for face recognition. Table 1 show the reviews
on the partial face recognition systems.

Table 1 Reviews on conventional partial face recognition techniques

Author Adopted methods Features Challenges
Lahasan et al. [18] OSPF • High accuracy rate • Needs improvement
• Robust approach on run time
complexity
Mahbub et al. [19] DRUID • Better performance • Needs improvement
• Increased precision in real time
applications
Greening et al. [20] FMRI • High accuracy • No consideration on
• Better reliable socio-emotional
decoding features
Duan et al. [21] TPGM • Minimal • It does not exploit the
computational cost 3D space
• High accuracy
Weng et al. [22] RPSM • Better efficiency • Have to focus on real
• Increased consistency time appliances
(continued)
390 G. G. Patil and R. K. Banyal

Table 1 (continued)
Author Adopted methods Features Challenges
He et al. [23] DFM • Improved efficiency • No consideration on
• Better accuracy partial RE-ID and
iLIDS databases
Lei et al. [24] TPWCRC • Minimal false rates • Requires more
• High-quality facial deliberation on
data computational
efficiency
Aminu and Ahmad [25] LPPLSDA • Enhanced specificity • Have to solve the
• Increased accuracy sample size issues

3 Methodology

PFR is the process of verifying or finding the facial emotions of human faces from
the video frame, video source, digital images or still images. Multiple methods are
involved in these systems in which the features of selected facial images are compared
within the images in database. This proposal aims to introduce a novel PFR method
termed as DFM that combine Sparse Representation Classification (SRC) and FCN
for resolving the partial face recognition issues. As the major contribution, this
proposal aims to tune the sparse coefficient of DFM in an optimal manner, such
that the reconstruction error should be minimal. Moreover, this proposal introduces
Jaccard Similarity Index measure to calculate the similarity scores among the gallery
sub feature map and probe feature map. For optimization purpose, this work deploys a
hybrid algorithm that hybrids both the concepts of Grey Wolf Optimization (GWO)
and Sea Lion Optimization (SLnO) algorithm. “The GWO algorithm mimics the
leadership hierarchy and hunting mechanism of grey wolves in nature [28]. The
SLnO algorithm imitates the hunting behaviour of sea lions in nature. Moreover, it
is inspired by sea lions whiskers that are used in order to detect the prey [29]”. Thus,
the partial faces will be detected in an effective manner. The overall architecture of
the implemented work is given by Fig. 1.

4 Conclusion

The proposed system regarding partial face recognition will be implemented in

MATLAB and the experimental investigation will be carried out. The performance
analysis will be done by comparing the proposed model over several state-of-the-art
models through the Type 1 measures and Type 2 measures.
Optimized Dynamic Feature Matching for Face Recognition 391

Fig. 1 Overall Architecture of the proposed work

References

1. Li P, Chen K, Wang F, Li Z (2019) An upper-bound analytical model of blow-out for a shallow

tunnel in sand considering the partial failure within the face. Tunnel Undergr Space Technol
91:Article 102989
2. Trofimov A, Drach B, Kachanov M, Sevostianov I (2017) Effect of a partial contact between
the crack faces on its contribution to overall material compliance and resistivity. Int J Solids
Struct 1081:289–297
3. Fang C, Zhao Z, Zhou P, Lin Z (2017) Feature learning via partial differential equation with
applications to face recognition. Pattern Recogn 69:14–25
4. Porpiglia F, Amparore D, Checcucci E, Fiori C (2019) Parenchymal mass preserved after
partial nephrectomy and “global renal damage”: two faces of the same coin. Eur Urol Oncol
2(1):104–105
5. Elmahmudi A, Ugail H (2019) Deep face recognition using imperfect facial data. Future Gener
Comput Syst 99:213–225
6. Werghi N, Tortorici C, Berretti S, Del Bimbo A (2016) Boosting 3D LBP-based face recognition
by fusing shape and texture descriptors on the mesh. IEEE Trans Inf Forensics Secur 11(5):964–
979
7. Zheng W, Gou C, Wang F-Y (2020) A novel approach inspired by optic nerve characteristics
for few-shot occluded face recognition. Neuro Comput 3761:25–41
8. Kryza-Lacombe M, Iturri N, Monk CS, Wiggins JL (2019) Face emotion processing in pedi-
atric irritability: neural mechanisms in a sample enriched for irritability with autism spectrum
disorder. J Am Acad Child Adoles Psychiatry (in press)
9. Yu N, Bai D (2020) Facial expression recognition by jointly partial image and deep metric
learning. IEEE Access 8:4700–4707
392 G. G. Patil and R. K. Banyal

10. He M, Zhang J, Shan S, Kan M, Chen X (2020) Deformable face net for pose invariant face
recognition. Pattern Recogn 100:Article 107113
11. Meinhardt-Injac B, Kurbel D, Meinhardt G (2020) The coupling between face and emotion
recognition from early adolescence to young adulthood. Cogn Dev 53:Article 100851
12. Trigueros DS, Meng L, Hartnett M (2018) Enhancing convolutional neural networks for face
recognition with occlusion maps and batch triplet loss. Image Vis Comput 79:99–108
13. Grati N, Ben-Hamadou A, Hammami M (2020) Learning local representations for scalable
RGB-D face recognition. Expert Syst Appl (in press)
14. García E, Escamilla E, Nakano M, Pérez H (2017) Face recognition with occlusion using a
wireframe model and support vector machine. IEEE Lat Am Trans 15(10):1960–1966
15. Young SG, Tracy RE, Wilson JP, Rydell RJ, Hugenberg K (2019) The temporal dynamics of
the link between configural face processing and dehumanization. J Exp Soc Psychol 85:Article
103883
16. Kim H, Kim G, Lee S-H (2019) Effects of individuation and categorization on face
representations in the visual cortex. Neurosci Lett 70824:Article 134344
17. Iranmanesh SM, Riggan B, Hu S, Nasrabadi NM (2020) Coupled generative adversarial
network for heterogeneous face recognition. Image Vis Comput 94:Article 103861
18. Lahasan B, Lutfi SL, Venkat I, Al-Betar MA, San-Segundo R (2018) Optimized symmetric
partial facegraphs for face recognition in adverse conditions. Inf Sci 429:194–214
19. Mahbub U, Sarkar S, Chellappa R (2019) Partial face detection in the mobile domain. Image
Vis Comput 82:1–17
20. Greening SG, Mitchell DGV, Smith FW (2018) Spatially generalizable representations of facial
expressions: decoding across partial face samples. Cortex 101:31–43
21. Duan Y, Lu J, Feng J, Zhou J (2018) Topology preserving structural matching for automatic
partial face recognition. IEEE Trans Inf Forensics Secur 13(7):1823–1837
22. Weng R, Lu J, Tan Y (2016) Robust point set matching for partial face recognition. IEEE Trans
Image Process 25(3):1163–1176
23. He L, Li H, Zhang Q, Sun Z (2019) Dynamic feature matching for partial face recognition.
IEEE Trans Image Process 28(2):791–802
24. Lei Y, Guo Y, Hayat M, Bennamoun M, Zhou X (2016) A two-phase weighted collaborative
representation for 3D partial face recognition with single sample. Pattern Recogn 52:218–237
25. Aminu M, Ahmad NA (2019) Locality preserving partial least squares discriminant analysis
for face recognition. J King Saud Univ Comput Inf Sci (in press)
26. Patil GG, Banyal RK (2019) Techniques of deep learning for image recognition. In: 2019 IEEE
5th international conference for convergence in technology (I2CT), Bombay, pp 1–5. https://
doi.org/10.1109/I2CT45611.2019.9033628
27. Patil GG, Banyal RK (2020) A dynamic unconstrained feature matching algorithm for face
recognition. J Adv Inf Technol 11(2):103–108. https://doi.org/10.12720/jait.11.2.103-108
28. Mirjalili S, Mirjalili SM, Lewis A (2014) Grey wolf optimizer. Adv Eng Softw 69:46–61
29. Masadeh R, Mahafzah B, Sharieh A (2019) Sea lion optimization algorithm. Int J Adv Comput
Sci Appl 10:388–395
Novel Secure Routing Protocol
for Detecting and Presenting Sybil Attack

S. M. Sawant, S. M. Shinde, and J. S. Shinde

Abstract Secure communication network is the demand of today’s transportation

due to excess number of vehicles and limitation to the existing communication
protocol. This Paper discusses the secure communication protocol between the
vehicles and the detection and prevention of the Sybil attack. Fake node request,
multiple routing entries these are the common example of Sybil attack which can be
detected and prevented using existing proposed method. Virtual nodes are created
and messages are sent to nodes. The proposed method detects the identity of received
message and number of fake entries of that particular massage in the routing buffer.
If messages are validated then communication is possible otherwise request is termi-
nated. Results showed that the existing proposed method is able to detect and prevent
Sybil attack by maintaining single node identity.

Keywords VANET · Ad hoc network · SRAN · Sybil attack

1 Introduction

Vehicular Ad-hoc Network is the technology [1] which can forms a secure network
between vehicles, i.e. Vehicles communicate to every alternative and transfer infor-
mation to another vehicle. VANET provides safety to driver of vehicles by exchanging
messages between vehicles. VANET is not secure because many types of attacks can
be appeared in it so it can leads to insecurity of drivers of vehicles. Vehicular Ad Hoc
Network (VANET) is an emerging area for research. Vehicular Ad-Hoc Network is
a challenging topic because of its mobility and link disruption. Many researchers [2]
have been working on specific issues of VANET like routing, broadcasting, Quality of
Service, security, architectures, applications, protocols, etc. Security is a main issue
in VANET because malicious drivers in the network disrupt the system performance.
Sybil attack creates multiple identities which lead to subvert the computer System

S. M. Sawant (B)
CSE Department, SVERI’s College of Engineering, Pandharpur, India
S. M. Shinde · J. S. Shinde
SVERI’s College of Engineering, Pandharpur, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 393
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_40
394 S. M. Sawant et al.

[3]. Novel secure routing protocol detects and prevents the Sybil attack particular on
Vehicular ad- hoc network. The proposed Secure routing protocol is based on ad-
hoc on demand (AODV) distance vector secure routing (DVSR) for Mobile ad-hoc
network (MANET) or other wireless ad-hoc network (WAN). This protocol main-
tains routing information and route discovery that detects and prevents Sevier Sybil
attack and each node have unique identity and entry in route table [2]. Inter vehicle
communication is passing and receiving the information to increase traffic efficiency,
detection of road conditions, avoid collisions, detect emergency situations and overall
increase of the efficiency of network. In a MANET or VANET, mobile or vehicular
nodes are making huge impact on the performance of routing protocols because of its
varying mobility characteristics. So many of the scholars and researchers [4] devel-
oped Ad hoc On-demand (AODV), Dynamic Source Routing (DSR), Destination
Sequence Distance Vector (DSDV) routing protocols for MANET but they cannot
directly used in VANET. Because of its apparently, widely varying mobility charac-
teristics of mobile or vehicular nodes are expected to have a significant impact on the
performance of routing protocols. Therefore even though researchers [4] have devel-
oped routing protocols like Ad hoc On-demand Vector (AODV), Dynamic Source
Routing (DSR), Destination Sequence Distance Vector (DSDV) etc. for MANET,
these protocols cannot be directly adopted in VANETs, efficiently, because of the
rapid variation in link connectivity, high speed and extremely varied density of vehic-
ular nodes in VANET. Researchers have developed special routing protocols for
VANET [3], and these are aimed to adapt rapidly. However, these recent protocols
are not fully secure and able to prevent attacks on VANET. These attacks are fraud
information, denial of service, black hole, alternative attack and Sybil attack. Sybil
attacks become a serious threat which can affect the functionality of VANETs for
the benefit of the attacker. The Sybil attack is the case where a single faulty entity,
called a malicious node, can present/create multiple identities known as Sybil nodes
or fake nodes. Sybil attack has been found in mostly peer-to-peer network where a
node in the network can operate as multiple identities at the same time it can gains
the authority and power in reputed systems. The main purpose behind this attack
is to gain the majority of influence in the network actions in the system. In a Sybil
attack a single entity that is computer has the capability to create and operate multiple
identities such as user accounts or IP address based accounts. For the other observers
these multiple attacks look like a real unique identity.

2 Literature

Sybil attack model consists of following parts and is as shown in Fig. 1:

E entities = c (correct) entities + f (faulty) entities

(i) Correct—entities that follow the protocols and rules setup in the network
honestly (whose honesty is verified).
Novel Secure Routing Protocol for Detecting and Presenting … 395

Fig. 1 Formal model of Sybil attack

(ii) Faulty—entities whose behavior are arbitrary and can’t be predicted. They
don’t honestly follow the protocols and rules in the network.
(iii) A communication cloud: A very general cloud through which messages
between different entities travel.
(iv) Pipe: to connect an entity with the communication cloud.

3 Methodology

The proposed novel Secure Routing protocol for Ad hoc Network (SRAN) is a routing
protocol for detecting and preventing Sybil attack. It is based on AODV and does
not allow Sybil node into Route discovery by eliminating the node from the route
table. Important views of SRAN protocol are route request packet format (RREQ),
route reply packet format (RREP) and route error packet format (RERR).
Route Request Packet Format
In SRAN routing protocol, if source wants to send message to destination then it first
broadcasts the route request (RREQ) to its neighbors. Neighboring node receives
route request packet format, if receiving node is not destination and does not have
396 S. M. Sawant et al.

route to the destination then it rebroadcast the route request packet format and same
time backward route is created to the source. If the receiving node is destination node
or it has current route to the destination then Route Reply (RREP) is generated.
1. RREQ ID: A sequence number uniquely identifying the particular RREQ when
taken in association with the source node’s IP address.
2. Source IP Address: The IP address of the Source.
3. Source Sequence Number: The Sequence number of Source.
4. Source Unique ID: The Unique Identification of Source.
5. Destination IP Address: The IP address of the destination for which a route is
selected.
6. Destination Sequence Number: The latest sequence number received in the
past by the source for any route towards the destination.
7. Destination Unique ID: The Unique Identification of Destination.
8. Hop Count: Number of hops needed to reach destination.

Route Reply Packet Format

RREP is unicast and it is hop by hop fashion to source.
1. Destination IP Address: The IP address of the destination for which a route is
given.
2. Destination Sequence Number: The Destination sequence number associated
to the route.
3. Destination Unique ID: The Unique Identification of Destination.
4. Source IP Address: The IP address of the Source.
5. Source Unique ID: The Unique Identification of Source.
6. Lifetime: Time to reach to the next Destination.
7. Hop Count: Number of Hops needed to reach the Destination.

Route Error Packet Format

When link break down is detected, RERR is generated and send to the source node in
hop by hop fashion. When each intermediate node invalidates route to an unreachable
destinations or Sybil node is detected then RERR is sent towards source node. When
source node receives RERR then it starts reinitiates route discovery.
1. Unreachable Destination IP Address: The IP address of the destination that
has become unreachable due to a link break.
2. Unreachable Destination Sequence Number: The sequence number in
the route table entry for the destination listed in the previous Unreachable
Destination IP Address field.
3. Sybil Node: This information about Sybil node which detected.

Route Maintenance
Once route is defined then route maintenance is also required. It is to provide infor-
mation about link of the route as well as route to be modified due to movement of one
Novel Secure Routing Protocol for Detecting and Presenting … 397

Table 1 Performance of SRAN protocol with source id and destination id

Source ID Destination ID Source IP Destination IP UID
2 4 192.168.10.10 192.168.10.16 1
4 3 192.168.10.16 192.168.10.15 2
4 1 192.168.10.16 192.168.10.17 1

Fig. 2 Difference of
performance of AODV
protocol and SRAN protocol

or more nodes in the route. Every time route is used to send packet then its expiry
time is updated by adding current time and Active Route Timeout (ART). ART is a
constant value that defines how long new route is kept into routing table of node after
last transmission done. ART defines both source and intermediate node. If route is
not used in the predefined period then node can’t be sure that route is still valid or not
and then this route is removed from routing table. It ensures that no any unnecessary
packet loss.

4 Result

Secure Routing protocol (SRAN) can improve the performance of routing in secure
communication. It means that this protocol will detect and prevent Sybil attack.
In SRAN total numbers of packets sent by source are received successfully at
destination. Table 1 shows the result and depicted in Fig. 2.

5 Conclusion

Security is the important challenges in VANET. SRAN directly rejects fraud

messages introduced by the malicious nodes, misguiding nodes in the network. This
avoids accidents and traffic jam on the road and saves vital life and time. SRAN
routing protocols are providing security for data transmission. On the road and saves
vital life and time. SRAN routing protocols are providing security for data trans-
mission. It provides techniques for attack detection and prevention in routing well
before it become malicious, suspicious and harmful. SRAN is better than VANET
398 S. M. Sawant et al.

especially in the performance that successfully detects and prevents Sybil attack.
SRAN routing protocol maintains unique identity of the node in order to establish
higher secure communication between vehicles to vehicle. SRAN routing protocol
is the best solution for the secure communication.

References

1. Haque MM, Misic J, Misic V et al (2013) Vehicular network security. In: encyclopedia of
wireless and mobile communications, 2nd edn
2. Gadkari MY, Sambre NB (2012) VANET: routing protocols, security issues and simulation tools.
J IOSR J Comput Eng (IOSRJCE)
3. Fonseca E, Festag A (2006) A survey of existing approaches for secure ad hoc routing and their
applicability to VANETS. J NEC Netw Lab, Version 11:1–28
4. Bai R, Singhal M (2006) DOA: DSR over AODV routing for mobile ad hoc networks. IEEE
IoT Model for Heart Disease Detection
Using Machine Learning (ML)
Techniques

Madhuri Kerappa Gawali and C. Rambabu

Abstract From the last decade, a tremendous spotlight is on the giving quality
medical services because of the exponentially growing of life threatening illnesses
of the patients. There are numerous components that influence the health condi-
tion of each person and a few illnesses are more dangerous and cause death of the
patient. And in the present age, the most common reason of death is heart disease.
This research work presents the IoT based system for heart disease detection using
the Machine Learning (ML) technique. It consists a novel preprocessing stage that
provides more accurate classification of the ECG signal. Also, this novel prepro-
cessing method removes the noise effectively from the raw ECG data. The classifi-
cation performance was evaluated using the various classifiers such as KNN, Naïve
Bayes and Decision tree that detects the normal and abnormal heart-beat rhythms.
With the obtained results, we have observed that the preprocessing has improved
the classification performance. This technique further proves that the decision tree
has good performance over the KNN and Naïve Bayes with respect to the accuracy,
sensitivity and precision.

Keywords KNN · Naïve Bayes · Decision tree · Preprocessing · ECG signals ·

IoT · Machine learning (ML)

1 Introduction

The health care industry has been adopting new technologies for providing better
and smart healthcare facilities [1]. With the IoT, remote and real-time monitoring of
patients is made possible and this unleashes the potential to continuously monitor
the health and helps the physicians to give suggestions or treatment in a timely
manner. As a larger community of people are suffering from heart disease, it is
vital to carry out diagnosis at the early stage to save lives and help to support a
healthy lifestyle of people. The health care monitoring has improved tremendously
due to the development of different IoT capabilities and instruments to track patient’s

M. K. Gawali (B) · C. Rambabu

University of Technology, Jaipur, Jaipur, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 399
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_41
400 M. K. Gawali and C. Rambabu

health conditions regularly [2]. The patients can also interact with the doctor more
easily which gives the satisfaction of treatment and it also reduces the hospital stay
and healthcare expenses. The main focus of employing IoT in healthcare system
is to set up a fully automated environment for patient monitoring and providing
assistance and care to patients in real-time. There is a rise in the need for a portable
system that can be used at home by the patient for measuring their ECG profiles and
diagnose their disorder in real-time. So in this paper, an extensive review is carried
out to find the existing technologies that are available for monitoring heart related
diseases. It is understood from the analysis, that the collected raw data contains noise
and irrelevant contents. These are irrelevant and incorrect data that are not useful for
diagnosis. This noise and huge variation in data leads to reduction in the classification
accuracy, sensitivity and precision. Therefore, in this paper a novel pre-processing
approach is used to remove noise and unrelated data from ECG signals. Relevant
attributes are identified using correlation technique to enhance data efficiency. The
machine learning classifier algorithm such as KNN, Naïve Bayes and Decision tree
are used for classifying the ECG signals based on waveforms [3]. The classifier that
obtains better performance metrics can be used for diagnosing the variation in the
ECG waveform and identify the type of abnormality and disorders.

2 Literature Review

Most of the time people go to the hospital only when after they suffer from cardiac
disease. In the traditional ECG setup, the medical instruments are housed in the
hospital, patients need to visit the hospital to check their heart disorders and study
their physiology of heart. During this process, the patient’s activities are limited.
Frequent visit to hospital increases the medical expenses and puts a burden on hospital
authorities. Early intervention is essential for the survival of patients, there has been
a lot of focus and attention on building an automated system for the detection of
abnormalities of heart signals. In [3] an IoT based wearable architecture was proposed
to measure the ECG signals. This system provides a portable platform where a non-
intrusive wearable sensor is used to collect the patient’s ECG signals and send them
to IoT cloud via the smart phone enabled Bluetooth or ZigBee technologies. The data
stored in the cloud can be retrieved by the specialist for further processing using data
analytics to find the disease. The data analytics procedure of data cleaning, storage,
analysis and generation of warning alerts to the concerned specialist in a real-time
manner can be performed by having access to the remote server. To facilitate the
early diagnosis of heart disease, machine learning techniques are employed. From
the health dataset, the investigation were performed to study the abnormal functions
of the heart. To classify the signals, the amplitude and interval periods of the cardiac
waves were analyzed using machine learning classification algorithms such as SVM,
Adaboost, ANN and Naïve Bayes [4]. Identifying accurate classifiers will assist the
physician in making quality decisions on diagnosis and timely treatment. There are
different types of arrhythmia diseases that are related to cardiac rhythm disorders. To
IoT Model for Heart Disease Detection Using Machine Learning … 401

ensure proper diagnosis, statistical and dynamic features extraction of ECG signals
is necessary [5]. So in this paper, heart rate variability is computed to generate alerts
when the patient is affected by arrhythmia disease.
In [6], to reduce the time consuming process of manually checking the ECG data,
a new classifier was proposed to distinguish normal and abnormal heartbeat rhythm.
This classifier removes noise and extracts ECG features. This classifier provided
better performance when compared with other machine learning classification algo-
rithms. The time computation is comparatively reduced and helps in identifying
arrhythmia disease. Early detection of abnormal pulse rates is also crucial for the
survival of the patient. So, to improve survival, a mechanism for the automatic detec-
tion of cardiac arrest was proposed. The ECG based pulse detection system uses the
random forest classifier (RF) [7]. The ECG data were processed to remove noise and
extract the features. The features were fed to random forest classifier and compared
with other existing classifiers. The RF classifier resulted in improved performance
helps the practitioners in making quick decisions for providing appropriate treat-
ment. The pre-processing are widely used in various fields for data cleaning, data
transformation, data integration and data reduction [8]. The identification of missing
values, noisy data and detecting outliers are performed on the data to perform data
cleaning. This data cleaning process provide significant improvement in the perfor-
mance of the classifier. Different pre-processing techniques are available that can be
applied to the dataset for improving the performance metrics. From [9], it can also
be understood that the preprocessing helps in better performance of the classifier. In
[10], a data driven approach used the outlier based alert system for identifying the
anomaly data of patients to reduce the measurement errors. When trained dataset
was tested in real-time system, the system proved to be effective. In the following
section, the novel pre-processing is proposed for classifying ECG data.

3 Proposed Methodology

3.1 Design of IoT Model

The portable IoT system is designed to work with sensors and microcontroller. The
components that are used for setting up the portable system are:
• LM35 Temperature Sensor
• Pulse Sensor
• AD8232 ECG Sensor
• Arduino Uno.
These 3 sensors are connected to the Arduino Uno microcontroller to collect
the body temperature, heartbeat rate and ECG signals. The different reading of the
patient’s vital signs are gathered and send for testing by the classifier model which
are using the dataset for detecting the abnormalities. The prototype is shown in Fig. 1.
402 M. K. Gawali and C. Rambabu

Fig. 1 Prototype of working IoT model

The fundamental concept behind the proposed methodology is to enhance the pre-
processing of ECG data. The proposed model has two main steps: Pre-processing
and classification of heart disease data. The block diagram of the proposed system
is shown in Fig. 2.

Fig. 2 Block diagram of proposed system

IoT Model for Heart Disease Detection Using Machine Learning … 403

3.2 Dataset

The historical health data of the patient was taken from the heart disease health dataset
for training the classifier. The dataset used for training the classifier for testing the
accuracy, sensitivity and precision of classification is the heart disease dataset which
was created by V.A. Medical Center, Long Beach and Cleveland Clinic Foundation.
This dataset is available as an open source from the UCI machine learning repository
[11]. In this dataset, 75 sets of attributes are available but 14 of these attributes are
considered for the prediction of heart disease. The dataset comprises records of 303
patients.

3.3 Correlation

To train the classifier model, heart disease dataset is used. Correlation is used to
identify the relationship between two continuous, quantitative variables. The iden-
tification of relevant attributes is implemented using the correlation technique [12].
All the irrelevant attributes are not considered for training the classifier model. The
correlation coefficient is computed to determine the relationship between the dataset
attributes. This improves the performance of the classifier algorithm by removing
weakly correlated attributes. This improves the performance of the classifier algo-
rithm by removing weakly correlated attributes. To better understand the correlation
between the attributes, the correlation graph is plotted is shown in Fig. 3.

3.4 Preprocessing

This helps to identify the association between the attributes. Correlation values range
from −0.25 to +1. Positive correlation represents that the column attributes either
increases or decrease together. A negative correlation indicates that one attribute
will increase and other one decreases or vice versa. The correlation graph obtained
is shown in Fig. 3 and is depicted in different colors. The dark maroon states that the
attributes are weakly correlated with another and orange color represents a strong
correlated with one another.

3.5 Processing

In the pre-processing phase, missing values in the dataset are identified before using
them for classification. The collected data in the dataset might contain erroneous
entries with noise, missing values, null values and incorrect values. The handling of
404 M. K. Gawali and C. Rambabu

Fig. 3 Correlation of the attributes

missing values is very important in machine learning algorithm implementation as

it would cause errors. The dataset with missing values degrades the performance of
classification. Therefore, in the pre-processing phase, the missing values are replaced
by NaN. Then the NaN values are replaced with mean values. Sometimes the values
may deviate drastically because of measurement errors. So to identify the wide
variation in the attribute values, outlier detection method is used. Outliers are used
to find the mistakes during data collection and variance in the data. Those values or
observations that go beyond the interquartile ranges (IQR) has great impact on the
processing of data. The Interquartile range is given as

IQR = Q3 − Q1

where Q3 represents the third quartile and Q1 represents the first quartile. Therefore,
it is essential to eliminate them so that errors of classification algorithm outcome is
reduced.
IoT Model for Heart Disease Detection Using Machine Learning … 405

3.6 Classification

The classification algorithm used are performing the pre-processing and classification
of dataset values are Naïve Bayes, K nearest neighbour (KNN) and Decision tree
[13, 14]. The KNN uses the Euclidean distance for computing the nearest neighbour
attribute.


D(P, Q) = ( pi − qi )2

pi represents the training set of attributes with a given class qi attributes. To determine
the class from nearest neighbour list, the majority vote of class labels among the K
nearest neighbours are computed using the formula

y = arg max Wi ∗ I (v = yi )
( pi ,qi  Dz )

where the distance of z is the set of k closest training examples to z and v represents
the class labels.
Naive Bayes performs classification by using the formula
  E
H p p(H )
p = H
E p(E)

H represents the hypothesis and E represents the evidence.

 
p HE is the posterior probability of class (target) given predictor (attribute).
 
p HE is the likelihood which is the probability of the predictor given the class.
(H) is the prior probability of the class.
(E) is the prior probability of predictor.
The attributes and predictors are independent. One particular attribute doesn’t
affect other attributes. The variable has two outcomes (yes or no) i.e., whether they
are affected by heart disease or not. The maximum probability of predictor variable
q is obtained using the formula


n
q = arg max P(q) P( pi /q)
i=1

where pi represents the different attributes.

The decision tree algorithm follows a top down approach to classify the attributes
in the dataset. The decision tree works by constructing a tree like graph. There are
two ways of finding the root attribute for the tree. The Gini chooses the attribute
randomly and entropy uses the logarithmic approach for finding the root attribute.
406 M. K. Gawali and C. Rambabu

The formula for Gini and Entropy is as follows.


Entropy = − pi log2 pi
i

The dataset is divided into two subsets, one for training and another for testing.
The training and testing dataset size are varied to identify a best classifier model.
The Naïve Bayes works well for large datasets when compared with other algo-
rithms. Also the decision trees algorithm is used because it can handle categorical
and numerical data. The classifier identifies normal and abnormality from the heart
disease dataset. The Naïve Bayes algorithm considers each of the attributes indepen-
dently to the probability that the patient is suffering from heart disease. It identifies
the maximum likelihood of the normal or abnormality from the patient’s health data.
The accuracy is computed by comparing the test value and predicted values. The
performance of models is evaluated in terms of accuracy with different test sizes.
The decision tree accuracy is high when compared with the Naïve Bayes.

4 Results

The real-time ECG signals of patients were collected from the sensors are fed for
testing via the microcontroller. The retrieval of ECG data from the experimental
setup is shown in Fig. 4. The classifier model is tested using Python.
The classifiers used are Naïve Bayes, KNN and decision trees. The classifier
is trained and validated using the dataset from UCI database. This pre-processing
procedure improves the performance of the classifier algorithm. Figure 5 shows the
Outliers computed using the IQR score. Those values that have large variations are
identified and removed.

Fig. 4 Retrieval of ECG

data
IoT Model for Heart Disease Detection Using Machine Learning … 407

Outliers for each attributes

Attributes

Fig. 5 Boxplot of outlier detection

Figures 6, 7 and 8 depicts the classifier algorithms performance in terms of Accu-

racy, sensitivity and Precision. The results of the three classifiers compared are KNN,
Naïve Bayes and Decision tree.
The testing of data from sensors was tested without pre-processing by the classifier
algorithm and then the same was tested by applying the pre-processing steps. The

Fig. 6 Comparison of
accuracy with and without
pre-processing

Fig. 7 Comparison of
sensitivity with and without
pre-processing
408 M. K. Gawali and C. Rambabu

Fig. 8 Comparison of
precision with and without
pre-processing

test was conducted by changing the training the dataset size. When more data was
used for training the classifier, the classification was improved. From the results, it is
found that the best performance was achieved by Decision tree classifier algorithm
when the preprocessing was performed to the data in terms of accuracy, sensitivity and
precision. So from the analysis, it is found that pre-processing improves classification
performance and this can be used for testing of ECG signals for abnormality and
disorders.

5 Conclusion

In this research work, we have discussed the major breakthrough that can be brought
about in the healthcare industry by the IoT platform. The benefits and challenges faced
by the healthcare system are presented. In this paper, a novel pre-processing technique
was proposed to improve the classification of ECG data using machine learning
algorithms. The classification algorithm used for training the system are KNN, Naïve
Bayes and decision trees. With the introduction of pre-processing technique, the
performance of decision tree outperformed other classification algorithms in terms
of accuracy, sensitivity and precision. The accuracy of the predicted is evaluated.
This method can be used for identification of normality and abnormality of ECG
signals and assist in making early and accurate diagnosis. This technique further
proves that the decision tree has good performance over the KNN and Naïve Bayes
with respect to the accuracy, sensitivity and precision.

References

1. Mora H, Gil D, Terol RM, Azorín J, Szymanski J (2015) An IoT-based computational

framework for healthcare monitoring in mobile environments. J Sens (Switzerland) 17(10)
2. Yang Z, Zhou Q, Lei L, Zheng K (2016) An IoT-cloud based wearable ECG monitoring system
for smart healthcare. J Med Syst
3. Celin S, Vasanth K (2018) ECG signal classification using various machine learning techniques,
pp 1–11
IoT Model for Heart Disease Detection Using Machine Learning … 409

4. Kalaivani RLDV (2019) Machine learning and IoT-based cardiac arrhythmia diagnosis using
statistical and dynamic features of ECG. J Supercomput 0123456789
5. Hammad M, Maher A, Wang K, Jiang F, Amrani M (2017) Detection of abnormal heart
conditions based on characteristics of ECG signals. J Meas 125:634–644
6. Elola A, Aramendi E, Irusta U, Del Ser J, Alonso E, Daya M (2018) ECG-based pulse detection
during cardiac arrest using random forest classifier
7. Bhaya WS (2017) Review of data preprocessing techniques in data mining
8. Zhu C (2016) Influence of data preprocessing, vol 10, no 2, pp 51–57
9. Hauskrecht M, Batal I, Valko M, Visweswaran S, Cooper GF, Clermont G (2013) Outlier
detection for patient monitoring and alerting. J Biomed Inform 46(1):47–55
10. https://archive.ics.uci.edu/ml/datasets/Heart+Disease
11. Rawashdeh M, Al Zamil MG, Hossain MS, Samarah S, Amin SU, Muhammad G (2018)
Reliable service delivery in tele-health care systems. J Netw Comput Appl 115:86–93
12. Alsheikh MA, Lin S, Niyato D, Tan H (2015) Machine learning in wireless sensor networks.
J Algor Strateg Appl 1–23
13. Qi J, Yang P, Min G, Amft O, Dong F, Xu L (2017) Advanced internet of things for personalised
healthcare systems: a survey. J Pervas Mob Comput 41:132–149
14. Gelogo YE, Hwang HJ, Kim H-K (2015) Internet of things (IoT) framework for u-healthcare
system. Int J Smart Home 9(11):323–330
Security Threats and Their Mitigations
in IoT Devices

Saurabh Gupta and N. Lingareddy

Abstract Internet of things (IoT) is another worldview converging with the social
networks, permitting data sharing between the individuals and electronic gadgets.
Likewise, it is expected for omnipresent connectivity among different entities or
things using Internet. Anyway, security and privacy issues are the major concerns
for IoT. The heterogeneous technological advancements, inherent vulnerabilities of
IoT devices, poor design IoT standard invites the cyber attack. This research work
mainly aims to address the security threats and issues on different layers of IoT archi-
tectures and their possible mitigations. Also, it provides a taxonomic representation
of the major 3-layers of IoT architecture with their protocol stack. Finally, we have
highlighted the most challenging security threats and their mitigations with some
future research work proposals.

Keywords IoT · Security · Privacy · Cloud computing security · Security threats

and mitigations

1 Introduction

Internet of Things (IoT) is one of the key components of digital and transformation of
digital world along with Social, Mobile, Analytics and Cloud (SMAC). It is otherwise
called as Internet of Everything or Industrial IoT. IoT, Big Data and SMAC can
helps as a numerous possibilities that were unheard earlier. It takes the absolute
center stage for the product vendors, system integrators, software companies and
IT sector companies. Today’s Industry analysts says, there can be around 26 billion
devices on the IoT (Cisco estimate 50 billion) by the end of 2020 and the data
exchange will be 40 Zettabytes over the networks [1]. According to the McKinsey
Global Inst., IoT market will have the major impact of $3.9–$11.1 trillion per year by
2025 over various applications i.e. smart cities, smart industries, home, offices, retail
environment, worksite, logistic and navigation, smart vehicles, etc. [2]. The abstract
level of IoT model contains various physical devices, or sensors i.e. controllable

S. Gupta (B) · N. Lingareddy

University of Technology, Jaipur, Jaipur, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 411
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_42
412 S. Gupta and N. Lingareddy

Fig. 1 Abstract level of three layer architecture of IoT

sensors, RFID (Radio Frequency Identifications), IoT getaways, web servers [3] as
depicted in Fig. 1.
The term “things” from IoT comprises both the physical and cyber worlds [4].
IoT brings up many challenges and holds much promise i.e. the data generated,
stored or transmitted through IoT devices, so many security issues and privacy of
the users can have serious consequences. Every challenge to the IoT system must
be secured, controllable and privacy to the smart users, only when the IoT systems
are built up with security. The general architecture of an IoT system is divided with
three layers i.e. (1) Physical layer, (2) Network Layer and (3) Application Layer.
The three layer architecture of IoT system is depicted in Fig. 1. The deployment
process involves various technologies i.e. WSNs, RFID, Bluetooth, NFC [5], IP,
EPC (electronic product code), Wi-Fi and various actuators [6]. The intelligent and
smart applications of IoT interconnected devices helps personal as well as economic
benefits to the society [7].

2 Security Issues on IoT

In IoT many sensor devices and smart peoples have the connectivity through Internet
to provide services at anytime, anywhere, and any types of services. IoT also provide
services at any business, anybody, any one, any context, any device, anything, any
path and any Network. Owing to the wide range of impact on daily life, all the sensor
devices are connected through Internet and all are also defenseless to all privacy and
security issues. Security in the field of Information Technology (IT) [8–10] considers
three features: confidentiality, integrity, and availability as the prime objectives and
are called as CIA triad [11]. Security is defined as a process by which unauthorized
access to the system state is prevented and thus the privacy is not compromised.
Confidentiality refers to the secrecy of data, whereas integrity confirms that the data
is not changed in transmission [12, 13]. Further, availability provides the smooth
transmission of data whenever it is required. In a secure network some of the required
capabilities are:
Security Threats and Their Mitigations in IoT Devices 413

i. Data Authentication: The sensed data and related information collected from
secured, authenticated devices must be followed some technical mechanism
and allow to transmit.
ii. Resilience to Attacks: During data transmission if the system crashes, it should
be automatically recover itself as same data uses in different domain. A cloud
server must be protect smartly and intelligently himself from an intruders or
eavesdropper.
iii. Client Privacy: At client side, the used data and information must be secure and
safe. The private data should be protected i.e. no other types of clients can’t be
access the private data from the client.
iv. Access Control: Only authenticated and authorized person can access the
control. The general user can access the system by providing user name and
password and their access rights, which will be controlled by the system
administrator. Different user can access the specific portion of the database
or programs to smoothen running the system.
Security issues are divided into different sub-categories, viz., data confidentiality,
monitoring and tracing of activities, avoidance of malicious insiders, hijacking
of services or processes including phishing, lack of transparency into providers’
service provisions and procedure environments, fraudulent activities and exploita-
tion, management of multi- instance in multi-tenancy etc. [14, 15]. Moreover, prob-
ability of attack through side channel, escaping the sandboxed environment, can
access the virtual machine and hence unauthorized or spoofed access to the host are
also a possibility [16–18]. Encryption techniques are also the most important tool in
providing multidimensional security services for IoT [3].

2.1 IoT Security Issue Challenges

In IoT mostly, applications data are personal related or industrial. These applications
data must be provided with security and confidentiality against any kind of unautho-
rised access [19]. The biggest challenging factor in IoT is security. Information is to be
transmitted over the communication channel with more security in the network. The
IoT improve the communication between devices but still there are so many issues for
time (response time), scalability, and availability. In IoT though machine to machine
technology is the first phase, but it enables new applications and to bridge diverse
technology by connecting physical objects together in support of intelligent deci-
sions. Open Web Application Security Project (OWASP) has defined ten top security
issues associated with IoT devices i.e. (1) Insecure network services, (2) Insecure
web interface, (3) Insufficient authentication or authorization, (4) Non availability
of transport encryption, (5) Privacy concern, (6) Lack of secured cloud interface, (7)
Lack of secured mobile interface, (8) In-sufficient security configuration, (9) Lack
of secure software/firmware and (10) Less physical security [20].
414 S. Gupta and N. Lingareddy

The major security challenges are:

i. During the transmission, data is to be secure from thefts/hackers.
ii. The insurance companies installing IoT devices on any application oriented a
device which collects data to take decisions about insurance.
iii. The rapid increase of IoT devices generates more traffic in the network commu-
nication. Therefore, there exist a requirement of the networks with more
capacity to store and process of huge data.
iv. As there exists many standards for IoT devices and many IoT companies,
authorized and unauthorized devices connected to the IoT system are most
challenging factor.
v. IoT systems primarily focus on different security issues, to propose the appro-
priate guidelines to achieve the secure communication over the network. For
dealing the IoT applications we require application security and network
security helps in securing the IoT communication network.

3 IoT Reference Model

The reference model of IoT can be represented by seven levels. The levels are:
i. Physical devices and controllers
ii. Connectivity
iii. Edge computing
iv. Data accumulation
v. Data abstraction
vi. Application
vii. Collaboration and processes.
Broadly different levels of an IoT system can be described as: Level 1: IoT system
consists of single device that performs sensing/actuation, performs analysis storing
data and host the applications. These levels are suitable for modelling less complexity
and less cost solutions where the data is medium and analysis requirements are
computationally extensive. Mostly these are used in home automation. A level 2:
IoT system has same as level 1 IoT system but it included local analysis. Information
storage is at the clouds and applications are generally cloud-based. These are more
appropriate for solution where the data usage is more and the requirement is same as
level 1 IoT. Mostly these are used in smart irrigation. But in Level 3 IoT system has
same as level 1 and 2 but data can be stored and processed in clouds and cloud-based
applications.
These system levels are involved in big data analysis and computationally inten-
sive requirements as depicted in Fig. 2. Level 4 IoT system consists various nodes
that performs local processing, i.e. cloud-based applications where data’s are stored
in the cloud and observer nodes (not performing any control functions). The observer
node can subscribe to and receive information collected in the cloud from the IoT
devices and also process information. These are more applicable for big data, and used
Security Threats and Their Mitigations in IoT Devices 415

Fig. 2 IoT level 1, 2, 3

in requirements are computationally intensive with requirement of multiple nodes.

Mostly we used these levels for noise monitoring. Level 5 IoT system consists many
nodes with a single coordinator which can collect information from the endnodes and
send to the cloud. The end node helps in sensing/actuation. These are mainly imple-
mented in forest fire detection. Information is stored, analyzed in a cloud database
and applications are cloud-based as in level 4. These level 5 systems data are big and
requirement analyses are computationally intensive.
But in level 6 IoT systems have various independent endnodes that performs
sensing/actuation and the information sends to cloud. Information is stored in cloud
(cloud database) and supports various cloud applications as depicted in Fig. 3. In
Level 7, these results will be seen using cloud applications. The main controller is
known to all end nodes and sends control commands to the nodes. Mostly these are
used in weather monitoring. The end node contains various sensors. An endnode
sends these information to the clouds using the Web Socket service. These informa-
tion analyses are done in the clouds by using cloud databases. To make a predic-
tion we aggregate the data in a cloud-based application which are visualized in the
cloud-database.
416 S. Gupta and N. Lingareddy

Fig. 3 IoT level 4, 5, 6 and 7

4 Attacks and Threats in IoT System Model

In IoT system model different layers of IoT are facing various attacks including active
and passive attacks. The attacks can be of two types depending on their behavior in
network i.e. active and passive attacks [21]. Network behaviors are
i. Active Attack: The attacker disturbs the performance of the network by stealing
the data at the time of communication in active attack.
ii. Passive Attack: The communications channels are observed and form its usage
history the passive attacker steal the information [21].
As discussed earlier, IoT framework model can be spoken to by three principle
layers for example Physical, system and application layer. Every one of these layers,
has its own innovations and some security shortcomings. The security problems
of each layers are discussed with possible threats and next part of this section, we
discussed the feasible solutions to that threats.
Attacks in Physical/Perception Layer
The physical layer incorporates sensors and actuators to perform estimation of temp,
speeding up, mugginess and functionalities like questioning area [22]. The funda-
mental security dangers in physical layer are because of constrained hub assets and
appropriated composed structure. The main threats are:
i. Tampering: These attacks generally focused on hardware components and an
attacker is to be physically present into the IoT systems and continue its process
Security Threats and Their Mitigations in IoT Devices 417

to make system busy. A few models are hub altering and vindictive code infu-
sion. In hub altering assailant can harm the sensor hub by electronically denies
to get to, alter the sensitive information or physically replace the part of its
hardware or entire node. But in malicious code injection, the attacker injects
its malicious code physically on to a node and access to that node from that
IoT system.
ii. Impression: Authorization/Validation in the disseminated systems are so trou-
blesome. So enabling malevolent hub to make a phony distinguishing proof for
noxious assaults.
iii. Denial of Service: An attackers can adopt the finite processing abilities of each
node, to make the system unavailable.
iv. Routing Attackers (WSN, RSN): In the Data collection and forwarding
processes, intermediate malicious node modify the routing path and make the
system busy.
v. Data Transit Attack (DTA): Various attacks like sniffing, Man in the middle
attacks on the integrity and confidentiality during data transit.

Attacks in Network and Transport Layer

System layer gives universal access condition to the physical layer. It get the informa-
tion from physical layer and transmit the assembled data to a specific data framework
through Internet or access systems [23]. The important security threats are:
i. Routing Attacks: During the data collection and forwarding process inter-
mediate malicious nodes may modify the routing path and get system
infected.
ii. DoS Attacks: In network layer a vulnerable attack is because of heterogeneity
and complexity of IoT networks. Exhaustion, collision and unfairness are the
three important methods in DoS attacks.
iii. Data Transit Attack (DTA): In core networks, during data transit, various attacks
occur on the data integrity and confidentiality.
iv. Spoofed Routing Information: Attackers spoof, alter or replay IP address to
disturb the traffic in the networks, resulting routing loops, fake error message,
and shortened routes etc.
v. Selective Forwarding: A malignant or altered hub may change the IP of the
traffic by dropping some message and sending others, subsequently debased.
Man in the Middle attacks: When the attackers jamming to access the
information for his advantage. It mainly contains three types of attacks like:
(a) Eavesdropping: It’s a passive attack, where attacker can access the
communication channel and alter the received packets and send to all.
(b) Directing Assault: Assailants may change the steering data and make
steering circle to essentially decay the nature of administrations.
Replay attacks: Attackers capture a signed packet and gain the trust of the destined
entity by resending later to the sender. It changes the message sequence numbers and
authentication code and also acts as real sender.
418 S. Gupta and N. Lingareddy

Attacks in Application Layer

Application layer provides the services as per request by the customers. Its signifi-
cance for the IoT is the capacity to provide superb brilliant administration. Different
IoT environments can be implemented in their application layer. The Application
Support Sub layer (ASS) underpins a wide range of business administrations, asset
assignment, canny calculation and can be implemented through specific middleware
as well as cloud computing platforms [24]. In this layer the main attacks are:
i. Data Leakage: The interloper/assailants can without much of a stretch take the
secret phrase or mystery information with an aware of vulnerable services or
applications.
ii. DoS Attacks: The interloper/attacker can demolish the accessibility of admin-
istrations or applications.
Malicious code injections: The intruder/attacker uploads their malicious code into
the software application, to get system infected and exploiting the layer vulnerable
to get attacked.

5 IoT Layer Attacks and Their Possible Solutions

Sometimes active attacks/vulnerable attacks can prevent the IoT devices smartly.
Prevention of IoT devices from the vulnerable attacks can be done by deploying
some security constraints [25, 26]. According to the behavior, different categorized
attacks are:
(a) Low level attack—When network is attacked by intruders and that and it’s
attack is not secure)
(b) Medium Level attack—When intruders are listening to the medium while
changing the data integrity.
(c) High level attack—When intruders is entered into the network and it can alter
the intensity as well as modify the original data.
Extremely High level attack—When Intruders attack on the network with the
adoption of unauthorized access and doing the illicit operations leading to suspended
or unavailability of networks or congestion of the network (Tables 1 and 2).
Solving the Security Challenges in Device Level
In designing phase of products the security aspect should be incorporated. The secu-
rity aspect can be introduced in operating system level and it should be extended
through the device stack to the implemented applications and having hardware secu-
rity capabilities. Generally IoT devices are having 70% security threat and 25%
security aspect concerns per device as depicted in Fig. 4. As most of IoT devices
may not be designed with security concern and it leads to susceptibility and config-
uration management problem. Figure 5 depicts about the organization’s controls for
controlling IoT devices.
Security Threats and Their Mitigations in IoT Devices 419

Table 1 IoT layer protocols: issues and their solution

Layer Protocol Issues Solution
Physical layer IEEE802.15.4 DTA AES-CCM algorithm [25]
BLE DTA: header type Black network solutions [27]
Wifi, LTE DTA WEP, WPA2 protocol [27]
EEA and EIA algorithm [28]
Network layer IPv4/IPv6 Threat to NDP protocolS SEND protocols in IPv6 [29]
LoWPAN DTA Compressed DTLS [25]
RPL Routing and DOS attack SVELTE IDS solutions [26]
Application layer MQTT DTA, scalable key SecuredMQTT solutions with
management ABE, Sec Kit solutions [30]
CoAP DTA Lithe solution [31]
AMQP Switching, reliability, Subscriber or publisher
message orientation, and models [32]
queuing
XMPP Gaming, chatting and Client–server and
multimedia callings server–server communication
path [32, 33]
DDS Publish/subscribe model Real-time communication [34]

6 Conclusion

Now a day’s IoT seems to be unbeatable and the overwhelming use of smart devices
cannot be reversible. Unless and until the security issues are addressed, organiza-
tions require to be attentive, putting appropriate controls, to ensure security risks
against the applications and focus on IoT devices those are performing well and who
are connected to their networks. The main idea behind this paper is to highlight the
security issues and their challenges to the different layers of IoT and deliberating
the security concern in various protocols and their possible corrective measures. IoT
devices became soft targets as they are deprived of security mechanisms. Security
mechanisms should be incorporated to all IoT related devices along with the commu-
nication networks. To protect from introducers or threats, we should have used default
password and for first time user, install all the security enabled requirements for all
the smart devices.
Table 2 Security threat in IoT automation and the mitigations
420

Layers Threat types Mitigations

Physical Tampering Tamper-resistant packaging
Denial of service (DoS) Spread-spectrum technique
Physical attacks Shared cryptographics and keys or a routing table
Impersonation AES-CCM algorithms
Routing attack and DTA WEP, WPA2 protocols, black network solution
Firmware alterations Use physical access control for updating procedures
Jamming Channel surfing, priority messages, and spatial retreats
Radio interferences Delayed disclosure of key
Tampering Tamper proofing, hiding
Collisions Error-correcting codes
Exhaustion Rate limitation
Unfairness Small frames
OS/software vulnerabilies Educating the peoples about security and conduct product tests
Networking (data processing) Denial of service Traffic control, link authentication, active firewalls, and passive monitoring (probing)
Eavesdropping Encryption, authorization
DTA Compressed DTLS
Back door attack At entry point in all system must be properly configured firewalls
Social engineering Awareness about security and its mechanism to the employee
Exhaustion Traffic monitoring
Malware Malware detection
Desynchronization Authentication
(continued)
S. Gupta and N. Lingareddy
Table 2 (continued)
Layers Threat types Mitigations
Flooding Client puzzle
Sinkhole Georouting protocol
Wormhole, blackhole Authorizations, monitoring redundancy
Homing Encryption
Mis-direction Authorization, Egress filtering and monitoring
Phishing or pharming Using SSLs to guarantee genuine display of sites
Data wire-tapping Protect communications using IPSEC, SSL/TLS
Application level Client app Anti-virus filtering
Data leakage Lithe solution
DoS attacks Secure MQTT solution with ABE
Malicious code injection Used virus protected S/W and handled the new vulnerabilities
Comm. channel Authorization, proper authentication and integrity verification
Security Threats and Their Mitigations in IoT Devices

Integrity, multiuser access Testing, planning and process design

Modifications Validation
Data access Traceability
User impersonation Using memory card, as a certificate mechanism
Device impersonation
Overwhelm Rate-limiting
Reprogram Authentication
Service interruption Control access mechanism through network
Data alteration Introducing certificate and access control mechanism
421
422 S. Gupta and N. Lingareddy

Fig. 4 IoT security

vulnerabilities on
IoT-devices

Fig. 5 Top controls for

securing the IoT

References

1. Van Der Meulen R (2015) Gartner says 6.4 billion connected ‘things’ will be in use in 2016,
up 30 percent from 2015. Stamford
2. Manyika J, Chui M, Bisson P, Woetzel J, Dobbs R, Bughin J, Aharon D (2015) Unlocking the
potential of the internet of things. McKinsey Global Institute
3. Lamba A, Singh S, Singh B, Sai S, Muni R, Islands C (2018) Quantum computing technology
(QCT)—a data security threat. Int J Emerg Technol Innov Res 5(4): 801–806
4. Ning H (2013) Unit and ubiquitous internet of things. CRC Press, Inc., Boca Raton
5. Sarangi M, Singh D, Khuntia M (2019) A potential solution for man-in-middle security issues
through near field communication (NFC). Int J Eng Adv Technol (IJEAT) 8(4):492–498
6. Al-Fuqaha A, Guizani M, Mohammadi M, Aledhari M, Ayyash M (2015) Internet of things:
a survey on enabling technologies, protocols, and applications. IEEE Commun Surv Tutor
17:2347–2376
7. Khan R, Khan SU, Zaheer R, Khan S (2012) Future internet: the internet of things architecture,
possible applications and key challenges. In: Proceedings of the 10th international conference
on frontiers of information technology, Islamabad, 17–19 Dec 2012, pp 257–260
8. Hashizume K, Rosado DG, Fernández-Medina E, Fernandez EB (2013) J Internet Serv Appl
4(5)
9. Ochani A, Dongre N (2017) Security issues in cloud computing. In: International conference
on I-SMAC (IoT in social, mobile, analytics and cloud) (I-SMAC)
Security Threats and Their Mitigations in IoT Devices 423

10. Tianfield H (2012) Security issues in cloud computing. In: IEEE international conference on
systems, man, and cybernetics
11. Sammy F, Vigila S (2016) A survey on CIA triad for cloud storage services, vol 9, pp 6701–6709
12. Paruchuri VL (2016) Data confidentiality in cloud using encryption algorithms. Int J Cloud
Comput Super Comput 3(2):7–18
13. Lamba A (2019) SR-MLC: machine learning classifiers in cyber security—an optimal
approach. Int J Res Inf Sci Appl Tech (IJRISAT) I(3)
14. Chandrika Sai Priya A (2016) Integrated framework for multi-user encrypted query operations
on cloud database services. Int J Cloud Comput Super Comput 3(2):1–6
15. Aich A, Sen A (2015) Study on cloud security risk and remedy. Int J Grid Distrib Comput 8(2)
16. Bhattacharyya D (2018) Space and security issues in cloud computing: a review. Int J Secur
Appl 12(6):37–46
17. Thirupathi Rao N, Sravani A, Bhattacharyya D, Kim T (2018) Security and assurance aspects
to be observed in cloud computing based data centers: a study. Int J Secur Appl 12(4):1–14
18. Thirupathi Rao N, Bhattacharyya D (2019) Security aspects to be considered in cloud
computing based data centers: a tutorial. Int J Database Theory Appl 12(1):27–42
19. Sicaria S, Rizzardia A, Griecob LA, Coen-Porisinia A (2015) Security, privacy and trust in
internet of things: the road ahead. Comput Netw 146–164
20. Frustaci M, Pace P, Aloi G, Fortino G (2017) Evaluating critical security issues of the IoT
world: present and future challenges. IEEE Internet Things J
21. Kocakulak M, Butun I (2017) An overview of wireless sensor networks towards internet of
things. In: IEEE 7th annual computing and communication workshop and conference (CCWC),
pp 1–6
22. Bhardwaj I, Kumar A, Bansal M (2017) A review on lightweight cryptography algorithms
for data security and authentication in IoTs. In: 2017 4th international conference on signal
processing, computing and control (ISPCC). IEEE, pp 504–509
23. Puthal D, Nepal S, Ranjan R, Chen J (2016) Threats to networking cloud and edge datacenters
in the internet of things. IEEE Cloud Comput 3(3):64–71
24. Pongle P, Chavan G (2015) A survey: attacks on RPL and 6LoWPAN in IoT. In: 2015
international conference on pervasive computing (ICPC). IEEE, pp 1–6
25. Hennebert C, Santos JD (2014) Security protocols and privacy issues into 6LoWPAN stack: a
synthesis. IEEE Internet Things J 1(5):384–398
26. Raza S, Wallgren L, Voigt T (2013) SVELTE: real-time intrusion detection in the internet of
things. Ad Hoc Netw 11:2661–2674
27. Adnan AH et al (2015) A comparative study of WLAN security protocols: WPA, WPA2. In:
Proceedings of international conference on advances in electrical engineering (ICAEE), Dhaka,
pp 165–169
28. Sulaiman AG, Al Shaikhli IF (2014) Comparative study on 4G/LTE cryptographic algorithms
based on different factors. IJCST 5
29. Gelogo YE, Caytiles RD, Park B (2011) Threats and security analysis for enhanced secure
neighbor discovery protocol (SEND) of IPv6 NDP security. Int J Control Autom 4(4):179–184
30. Neisse R, Steri G, Fovino IN, Baldini G (2015) SecKit: a model-based security toolkit for the
internet of things. Comput Secur 54:60–76
31. Raza S, Shafagh H, Hewage K, Hummen R, Voigt T (2013) Lithe: lightweight secure CoAP
for the internet of things. IEEE Sens J 13(10):3711–3720
32. Brendel J (2003) World-wide-web server that finds optimal path by sending multiple syn +
ack packets to a single client. U.S. Patent 6,587,438, issued 1 July
33. Mishra N, Swagatika S, Singh D (2020) An intelligent framework for analysing terrorism
actions using cloud. In: New paradigm in decision science and management. Springer,
Singapore, pp 225–235
424 S. Gupta and N. Lingareddy

34. Lu C, Blum BM, Abdelzaher TF, Stankovic JA, He T (2002) RAP: a real-time communication
architecture for large-scale wireless sensor networks. University of Virginia: Department of
Computer Science, Charlottesville
Age and Gender Classification Based
on Deep Learning

Tejas Agarwal, Mira Andhale, Anand Khule, and Rushikesh Borse

Abstract Now a days, age and gender classification became relevant to growing
applications such as human-computer interaction (HCI), biometrics, video surveil-
lance, electronic customer relationship management, forensic, and many more. The
social networks have a huge amount of data available, but often, people do not
provide some of their data, such as age, gender, and other demographics. Age and
gender of a person from face image is one such significant demographic attribute.
Face recognition is an extremely difficult issue in the field of image analysis and
computer vision on the grounds. Although Face recognition had been done over a
couple of decades with traditional image processing features, recently Deep Convo-
lutional Neural Networks (CNNs) has been proved that it is the descent method for
the classification task, especially of uncontrolled, real-world face images. In this
research, we propose a deep learning CNN method to extract discriminative features
from the uncontrolled, real-world dataset and classify those images into gender and
age group. In this paper, learning representation is shown through the use of the deep
learning-based convolutional neural network (CNN), essential growth in the perfor-
mance can be achieved on these works. To this end, we would like to mention that
our CNN model can be used even when the number of learning features is restricted.
We evaluate our technique on two prominent datasets. Accuracy up to 80% obtained
for the Adience datasets and 95% for the Gender Classification dataset from Kaggle.

Keywords Age classification · Gender classification · CNN · Deep learning

T. Agarwal (B) · M. Andhale · A. Khule · R. Borse

Department of Electronics and Telecommunication, School of Electrical Engineering, MIT
Academy of Engineering, Alandi Pune, India
e-mail: [email protected]
M. Andhale
e-mail: [email protected]
A. Khule
e-mail: [email protected]
R. Borse
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 425
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_43
426 T. Agarwal et al.

1 Introduction

Recently, the age estimation from face image has drawn out large attention in
computer vision due to its significant application in video surveillance, demography,
internet access control and security [1]. Age and gender which are the two facial
key attributes, play very important roles in social interactions, making gender and
age estimation from a one face image is a prime task in intelligent applications [2].
In contrast, the facial age estimation is a difficult process and get easily affected
by many elements such as identity, gender, and extrinsic factors, including artificial
position, environment, human lifestyle and facial expressions. Datasets of the human
facial image have different demographical attributes [3], ex. Regarding the same age
of two persons, one can be a male with a face that appears old, while the other can
be a female with a face that appears young. Hence, enabling a computer system
to categorize the face images based on gender and age of the person is yet a diffi-
cult task. In the last two decades, the facial images classification based on age and
gender has received in many of the research. Past proposals to estimate facial image
attributes have depended on differences in facial feature aspect. Many approaches
have established classification strategies designed specifically for gender or age clas-
sification tasks [4]. Still from a given facial image, accurate physical age estimation
is a difficult problem.
In this research, we approach Deep Convolutional Neural Network, which has
novel face recognition and age, gender classification capabilities. A Deep Convolu-
tional Neural Network CNN is widely used for image recognition and processing.
Convolutional Neural Network has input layers, output layers, and multiple hidden
layers, many of which are convolutional. To check the efficiency of the CNN model,
we approach Adience dataset [5]. This dataset has all categories of images for training
as well as testing purposes.
The rest of the paper is organized as follows. In the next section, we reviewed
prominent techniques associated with age and gender classification based on deep
learning. Our proposal for classification using CNN is detailed in Sect. 3. The dataset
used for the implementation and its associated results can be found in Sect. 4. We
conclude the paper in Sect. 5.

2 Associated Works

Before describing the proposed method, we would give a broad survey of significant
work-related to age and gender classification, which were based on CNN.
In recent years there has been increasing attention, and many methods have been
put forth for the difficulty with automatically extracting age features from the face
images—a detailed survey of many techniques found in [6]. Previous age estimation
methods were based on computing ratios between different measurements of the
facial attributes [7]. The requirement of the input image in methods that represent
Age and Gender Classification Based on Deep Learning 427

the aging process is to be frontage, and well smoothen and thus, current observed
results only on constrained datasets having frontage images [8]. Few have addressed
the difficulties that arise from real-world variations in picture quality, while some
previously have shown fairly good results on constrained input images such as near-
ideal lighting, angle, and visibility [9].
A detailed associated work of gender classification methods with automatically
detected aligned faces can be found in [10]. One of the previous methods for SVM
gender classifiers were used by [11]. The appearance, intensity, quality and texture
features were used with related information in [12]. Deep CNN has additionally been
successfully applied to applications, including facial keypoint detection [13].
A human face contains a large variation in appearance that captured from the
real world conditions. Thus, some of the approaches having difficulty under such
uncontrolled conditions and they unable to provide an ethical way to handle the
outliers. Thus, [14] together proposed a new methodology called Robust Cascaded
Pose Regression to reduce the risks to the outliers by detecting obstruction precisely
and used robust shape-indexed features.
Detail discussion about gender classification and age estimation using CNN found
in [15]. CNN model was used for the gender classification and adience dataset is used
for testing and training in [16].

3 Proposed Methodology

Figure 1 shows a general block diagram of gender and age classification. Initially,
the gender is classified and then it is placed in the appropriate age group. A CNN
based deep learning model is employed. A general understanding of this can be seen
in Fig. 2. The parameter tuning for this CNN model is tabulated in Sect. 4. However,
we explain the general processing inside a CNN model.
Deep Learning has one of the algorithms named Convolutional Neural Networks
(CNN). This algorithm takes the input as an image, and several attributes of the image
are learnt through the kernels. Important objects present in the image are learnt by
CNN, which allows them to detect one image from the other. The convolutional
neural network learns the specific attribute of a person to identify gender easily. One
dominant feature of Convolutional Network is its ability to pre-process the data by
itself as other Machine Learning algorithms need to provide the features manually.
Thus, they may not spend a resource in data pre-processing. They are also able to
learn the features and develop filters of their own.

3.1 Network Architecture

Convolution Neural Network architecture consists of Convolution Layer, RELU

Layer, Pooling Layer, and Fully Connected Layer.
428 T. Agarwal et al.

Fig. 1 The proposed methodology for age and gender classification

Fig. 2 Convolutional neural network architecture

Age and Gender Classification Based on Deep Learning 429

Convolution
Convolution layer performs the extraction of features from the face image and reduces
the size of images for the faster computations, as shown in Fig. 3. The convolutional
neural network learns the patterns within the face image and will recognize it in the
picture.
There are three convolution layers in the network, followed by a pooling layer.
Convolution layer- Parameter tuning
1. Convolution Layer1- 96 filters with size 3 × 7 × 7 are convolved with stride
4 and padding 0, providing an output size of 56 × 56 with 96 kernels. This is
followed with Non-linearity ReLU, maximum pooling which reduces the size
to 28 × 28 with 96 kernels.
2. Convolution Layer2- 256 filters with size 96 × 5 × 5 are convolved with stride
1 and padding 2, providing an output size of 28 × 28 with 256 kernels. This is
also followed with Non-Linearity ReLU, maximum pooling, LRN and reduces
the size to 256 × 14 × 14 as output.
3. Convolution Layer3- 256 filters with size 256 × 3 × 3 are convolved with stride
1 and padding 1, and the layer followed with ReLU and maximum pooling,
provide an output size of 256 × 7 × 7.

Non-Linearity—ReLU Activation Function

After Convolution operation, the output is given to the ReLU layer to allow non-
linearity. In this layer, all pixels in the image having negative value will be replaced
by zero, as shown in Fig. 4.
Pooling
Pooling is used to reduce the size of an input image, by reducing the dimensions of
the image, the network has a lower weight to calculate, so it prevents over fitting.
There are two types of Pooling as Max Pooling and Average Pooling. Average
pooling is that which returns the average of all pixel value in the image portion

Fig. 3 Image pixel after applying convolution

430 T. Agarwal et al.

Fig. 4 Image pixels after applying Non-Linearity ReLU function

Fig. 5 Image pixels after applying pooling

covered by the filter. Maximum pooling returns the maximum value in the image
portion covered by the kernel. Maximum pooling is used in the proposed work to
reduce the dimension of the input image, as shown in Fig. 5.
Fully Connected Layer
A Fully connected layer is used to compute the class score. 512 neurons are fully
connected to the 256 × 7 × 7 output of Convolution layer 3 and followed by a
Non-Linearity ReLU layer and dropout layer. At the end, there is a softmax layer
connected on top of the fully connected layer to provide the loss and final class score
and maximum probability considered as a predicted class, as shown in Fig. 6.

4 Implementation and Results

4.1 Description of the Datasets

We test our CNN design on the Adience benchmark dataset [5] and the Gender
Classification dataset [17]. The Adience benchmark dataset is prominently signified
for the gender and age classification. The automatically uploaded images on Flickr
from smart-phone devices has made the Adience dataset. These images without prior
Age and Gender Classification Based on Deep Learning 431

Fig. 6 Fully connected network

filtering uploaded on Flickr, as is normally the case on social platform viewing situ-
ations of these images are highly unconstrained, considering real-world challenges
of face aspect in Internet images. The entire Adience dataset includes 2284 subjects
with 26,000 images roughly. These images are used in the techniques to achieve the
performance gain rather than better processing attributed to the network architecture.
Another dataset for testing and training is the Gender Classification dataset down-
loaded from Kaggle [17]. The dataset is of cropped images of males and females. It
is split into training and the validation directory. Training contains 2500 images of
each class, and a validation directory contains 100 images of each class.
Below Table 1 shows dataset distribution of Adience faces benchmark into age
and gender classes [5].
Below Table 2 shows the distribution of gender classification dataset into different
gender classes [17].

Table 1 Adience faces benchmark

Age 0–2 4–6 8–12 15–20 25–32 38–43 48–53 60–100 Total
Male 745 928 934 734 2308 1294 392 442 8192
Female 682 1234 1360 919 2589 1056 433 427 9411
Both 1427 2162 2294 1653 4897 2350 825 869 19,487

Table 2 Gender
Gender No. of images
classification dataset
Male 2600
Female 2600
Both 5200
432 T. Agarwal et al.

4.2 Parameter Tuning

The parameters are tuned in the algorithm to control the behaviour of parame-
ters. Normally the algorithm creates a different model for each value of the tuning
parameter. The parameters chosen for tuning in the CNN algorithm are shown in
Table 3 for age and gender classification.

4.3 System Setup

Results are observed by using the system having the specifications, as shown in Table
4.

4.4 Result and Discussion

The methodology used of deep learning based on CNN is for face detection and
recognition, which requires certain limiting factors for the orderly design of the
system. Based on the various parameters, changes like by providing one hidden layer
then two, we have observed the changes in the output and changes in the accuracy.

Table 3 Parameter tuning

Parameter Specification
Classifier Sequential
Input size 3 × 256 × 256
Classes 2, 8
Convolution layer 3
Pooling 1–1 for Conv Layer (Max)
Fully connected layer 2
Neural network 1, 2, 3
Class mode Categorical

Table 4 System specification

Parameter Specification
Computer Lenovo T480
Operating system Windows 10
CPU Intel Core i5-83500U
RAM 8 GB
Hard disk drive 235 GB SSD
Age and Gender Classification Based on Deep Learning 433

We have also observed the changes by providing initially one convolution and
pooling layer and then increasing the convolution and pooling layer. There is a change
in the result after every parameter change, and based on the observations, we have
taken the most effectual parameter for face recognition with average accuracy. As
we do the preprocessing of images, it will suppress unwilling distortions or enhance
some important for further processing that reduces the complexity and increases the
accuracy of CNN.
We train our CNN design on 80% of the Gender classification dataset and the
remaining 20% used for testing of CNN model. We achieved more accuracy on
Gender classification.
Confusion Matrix:
The performance of classification algorithms is summarized in a confusion matrix.
Classification accuracy alone can be misleading if we have an unequal number of
observations in each class or if we have more than two classes in our dataset. As we
have 2 class for gender and 8 age group, so the confusion matrix is 2 × 2 and 8 ×
8. Confusion matrix for the age group on Adience dataset and for Gender class on
Gender Classification dataset as shown in Fig. 7a and b.
We have implemented the CNN model on the dataset as mentioned in Table 1,
and sample results are mentioned in figure from Figs. 8a and 9.
Image was taken of a child and detected as male gender within the age group of
4–6 as positive misclassification.
Image taken of a girl and truly detected as female gender within the age group of
25–32.
Image taken of an old woman and truly detected as Female gender within the age
group of 48–53.
Image was taken of a boy and detected as male gender within the age group of
25–32 as positive misclassification.
Image was taken of a sofa and no face detected.

Fig. 7 a Confusion matrix- adience dataset, b. Confusion matrix- gender classification dataset
434 T. Agarwal et al.

Fig. 8 a Gender Male with Age 4–6, b. Gender Female with Age 25–32, c. Gender Female with
Age 48–53, d. Gender Male with Age 25–32

Fig. 9 Sofa Set—No face detection

Age and Gender Classification Based on Deep Learning 435

Formulas for calculating Accuracy, Precision, Recall, and F1 score:

Trulydetected + Not detected

Accuracy = (1)
(Trulydetected + Not detected + PositiveMisclassification
+ NegativeMisclassification)
Truly detected
Precision = (2)
(Truly detected + Positive Misclassification)
Truly detected
Recall = (3)
(Truly detected + Negative Misclassification)
2xRecallx Precision
F1score = (4)
Recall ± Precision

Implementation result Accuracy, Precision, Recall, F1 score on both dataset

mentioned in Table 5.
It is seen in Fig. 8d the result is partially correct. The gender is properly classified
but the age group is partially correct leading to misclassification because factors
like makeup, lighting, obstructions, and facial expressions would cause difficulty to
accurately detect an exact age group from a single image. The image taken without
any face gives the result as no face detected as shown in Fig. 9. The participant in
Fig. 8d. is one of the author Tejas Agarwal of this paper.

5 Conclusion

This paper presented the age and gender classification using CNN method to process
real world unconstrained database. The test result verified the effectiveness of the
proposed method using a deep convolutional neural network.
Taking an example of the related problems of face recognition, we analyze how
good deep Convolutional Neural Network algorithm performs on these tasks using
unconstrained data. Hereby it can be concluded that pre-processed images datasets

Table 5 Results obtained

Performance parameter Adience set (Value) Gender classification (Value)
Accuracy 0.8 0.95
Precision 0.75 1
Recall 1 0.95
F1 score 0.857 0.97
Time required for testing (sec) 3 3
436 T. Agarwal et al.

give good accuracy. Also, some parameters such as kernel size, number of convolution
layers and hidden layers affect the accuracy.
Considering the smaller size of unconstrained image datasets categorized for age
and gender, CNN provides an improved result for age and gender classification.
Adience dataset and Gender Classification dataset were used for the testing and
training of this CNN model. As a result of this experimental studies, gender classifica-
tion on adience was done with 80% accuracy, and gender classification on the Gender
Classification dataset was done with 95% accuracy. As we increase the number of
layers in the CNN model, it provides improved accuracy.

Acknowledgements We wish to record our deep sense of gratitude and profound thanks to our
project teacher Prof. Prashant Aher for helping us on drafting this paper. He educated us on the
ethics of paper writing and to bring this paper into fruition.

References

1. Liu X, Li S, Kan M, Zhang J, Wu S, Liu W, Han H, Shan S, Chen X (2015) AgeNet: Deeply
learned regressor and classifier for robust apparent age estimation. In: Proceedings of the IEEE
international conference computer vision workshop, pp 258–266
2. Rothe R, Timofte R, Van Gool L (2016) Deep expectation of real and apparent age from a
single image without facial landmarks. Int J Comput Vis
3. Duan M, Li K (2018) An ensemble CNN2ELM for age estimation. IEEE Trans Inf Forensics
Secur 13(3)
4. Deep Kaur K, Rai P (2017) An analysis on gender classification and age estimation approaches.
Int J Comput Appl (0975–8887), 171(10)
5. Kaggle, https://www.kaggle.com/ttungl/adience-benchmark-gender-and-age-classification
6. Fu Y, Guo G, Huang TS (2010) Age synthesis and estimation via faces, A survey. IEEE Trans
Pattern Anal Mach Intell 32(11):1955–1976
7. Kwon YH, da Vitoria Lobo N (1999) Age classification from facial images. In: Proceeding
conference computer vision and image understanding, vol 74. IEEE, pp 1–21
8. Geng X, Zhou ZH, Smith-Miles K (2007) Automatic age estimation based on facial aging
patterns. IEEE Trans Pattern Anal Mach Intell 29(12):2234–2240
9. Eidinger E, Enbar R, Hassner T (2014) Age and gender estimation of unfiltered faces. IEEE
Trans Inf Forensics Secur 9(12):2170–2179
10. Makinen E, Raisamo R (2008) Evaluation of gender classification methods with automatically
detected and aligned faces. IEEE Trans Pattern Anal Mach Intell 30(3):541–547
11. Moghaddam B, Yang MH (2002) Learning gender with support faces. IEEE Trans Pattern Anal
Mach Intell 24(5)
12. Perez C, Tapia J, Estevez P, Held C (2012) Gender classification face images using mutual
information and feature fusion. Int J Optomechatronics 6(1):92–119
13. Sun Y, Wang X, Tang X (2013) Deep convolutional network cascade for facial point detection.
In: IEEE conference on computer vision and pattern recognition. Portland, OR, pp 3476–3483
14. Burgos-Artizzu XP, Perona P, Dollar P (2013) Robust face landmark estimation under occlusion.
In: IEEE international conference on computer vision (ICCV). Sydney, Australia, pp 1513–
1520
15. Trivedi G, Pise NN (2020) Gender classification and age estimation using neural networks: a
survey. Int J Comput Appl (0975–8887) 176(23)
Age and Gender Classification Based on Deep Learning 437

16. Inik O, Uyar K, Ulker E (2018) Gender classification with a novel convolutional neural network
(CNN) model and comparison with other machine learning and deep learning CNN models. J
Ind Eng Res 4(4):57–63. ISSN-2077–4559
17. Kaggle, https://www.kaggle.com/cashutosh/gender-classification-dataset
Handling of Auxiliaries in Kannada
Morphology

Bhuvaneshwari C. Melinamath

Abstract Study of auxiliaries from a morphological perspective is extremely inter-

esting from the semantic and pragmatic points of view, and they still await detailed
and careful study.This paper deals with the role played by auxiliaries in Kannada
Morphology. Two kinds of auxiliaries are indicated in the Kannada language, namely,
aspect auxiliaries and modal auxiliaries. These auxiliaries are useful information of
derivative stems in Kannada verbs. Many of these aspect auxiliaries are used as
verbalizers to derive verbs from nouns and verbs from adjectives. Auxiliaries play
a very important role in Morphology of Dravidian languages like Kannada. They
are useful in the complex verb formation process. Aspect markers have a major role
in verb morphology. This paper explores the distribution of non-finite auxiliaries
and modal auxiliaries as part of the morphological analyser in Standard Kannada.
Auxiliaries form the basis for the formation of multiple stems in Morphology when
they are not used separately like in the English language.

Keywords Natural language processing · Finite-state transducers · Natural

language understanding · Natural language generation · Auxiliaries

1 Introduction

NLP (Natural Language Processing) is a branch of AI (artificial Intelligence) studies

the problems of automated generation and understanding of natural human languages.
The goal of NLP is to design and build software that will analyze, understand,
and generate languages that humans use to communicate with each other. Natural
Language Processing (NLP) is a convenient description for all attempts to use
computers to process natural language. Natural Language Processing (NLP) includes
Natural Language Understanding (NLU) and Natural Language Generation (NLG).
Study of auxiliaries from a morphological perspective is extremely interesting from
the semantic and pragmatic points of view, and they still await detailed and careful

B. C. Melinamath (B)
Department of Computer and Engineering, SVERI College of Engineering, Pandharapur,
Maharastra, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 439
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_44
440 B. C. Melinamath

study. The comparative study would yield rich information regarding the semantic
structure of languages, apart from being useful for translation systems, etc. Auxil-
iaries are the helping verbs. Auxiliary verbs are used in conjunction with main verbs.
Auxiliary verbs usually accompany the main verb. The main verb provides the main
semantic content of the clause. An example is the verb have in the sentence I have
finished my dinner. Here, the main verb is finished, and the auxiliary “have” helps
to express the perfect aspect. To know state of the art, we carried out the literature
survey.

2 Literature Survey

The syntax of auxiliaries has given rise to much discussion in the generative literature
[1] discussed finite-state transducers based system for Hindi, [2] have discussed the
auxiliaries from a generative perspective [3, 4] focus on syntactic structures. With
respect to auxiliary [5], have focussed on serial verb construction using auxiliaries
[6]. Focussed on the syntax of valuation in auxiliary-participle constructions [7] has
focussed on verb phrase use of ellipsis, phases and the syntax of morphology.) [8] has
discussed issues conserving the Clitics, morphological merger, and the mapping to
phonological structure, [9] has discussed the Kannada morphology aspect concerning
syntax and semantic of Kannada Language.

2.1 English Auxiliaries

English auxiliary verbs or helping verbs such as will, shall, may, might, can, could,
must, ought to, should, would, used to, need are used in conjunction with main verbs
to express shades of time and mood. The combination of helping verbs with main
verbs creates what are called verb phrases in English. In English, shall is used to
express simple future. The different forms of the has. have used to express tenses
present perfect and past perfect. There is also a separate section on the Modal auxiliary
such as can, could, may, might, must, ought to, shall, should, will, and would, do not
change the form of different subjects. For example “I can write". in this example, the
modal auxiliary express various meanings of necessity, advice, ability, expectation,
permission, possibility, etc.

3 Proposed Methodology

The auxiliaries occupy more than 75% of corpus file, hence analysis and handling
of Kannada auxiliary is crucial for translation purpose, the methodology is shown
in Fig. 1. The input module reads the input and tokenizes the input sentence in to
Handling of Auxiliaries in Kannada Morphology 441

Fig. 1 Proposed architecture for auxiliary analysis

words and each word is converted into transliteration module using ir.pl program,
each transliterated word is searched in the dictionary, if word is found then tag in
the dictionary is assigned to it, otherwise the word is passed to next module to check
for the morphological inflections. The morph module analysis the input and gives
output with part of speech tags.

3.1 Kannada Auxiliaries

Kannada auxiliaries can be divided in to aspect auxiliaries and modal auxiliaries.

However, the auxiliaries in English occur as free morphemes and easy for analysis.
But in Kannada they occur as bound morpheme suffixed along with verb with which
they occur and thus yield rich information in semantic structure of Kannada language.

3.2 Aspect Auxiliaries

The general occurrence of aspect auxiliary is past participle form of the verb followed
by aspect auxiliary. Aspect auxiliary always preceded by past participle form. (past
verbal participle + Aspect auxiliary verb). Another essential distinction between
Kannada auxiliary and English auxiliary is the verbs which are acting as aspect
markers have used as the main verb. Also, this is not true with English auxiliaries
like has, shall etc. The example illustration is given in Table 1.
Kannada has a set of verbs that may be added to verbal participle to give certain
semantic nuances to the meaning of the sentence. Aspect markers are very similar to
the main verbs in their morphology and syntax. In fact, they are derived from certain
main verbs. But semantically they do not express the lexical meaning like their main
verbs express in their auxiliary aspect usage. The aspectual biDu ‘completive’ does
not mean the same as main verb biDu ‘leave’. Consider an example of verb formation
by adding a set of the auxiliary, as shown in Table 1. In Kannada thousands of
such verb forms can be generated. That is why the study of auxiliaries in Dravidian
languages is a challenging task. In Kannada language, the verbs which are acting as
442 B. C. Melinamath

Table 1 Inventory of aspect auxiliaries verbs in Kannada

aspect auxiliary are actually the main verb. Apart from their usage as the main verb,
these verbs are acting as auxiliaries also. This is not true for the English language.
Consider an example below.
Handling of Auxiliaries in Kannada Morphology 443

Here bareduko Du is a though looks like a simple single verb. But it is the complex
verb, which is formed by the addition of aspect auxiliary ‘koDu’ to the baerdu form,
which is a non-finite (past verbal participle form) of verb ‘bare’ (write). This kind
of formation of verb leads to lakhs of complex verbs. But in reality, there are only
around 2000 basic verb roots. Another thing is the complex verbs formed by this
process follow the TAM (tense, aspect and modality) inflections to the auxiliary verb
and not to the first verb. But however, the meaning of such formation is inferred from
the first verb itself. In the below example in Table 1 “tiMdubiDu”. The meaning here
is eaten itself; it is not like ate and then left. This process of complex verb formation
is productive and regular in Kannada. It is not wise to store such complex verbs in the
dictionary; instead, we have handled the formation of such complex verbs through
our morphology. In currently existing morphological systems these verbs are stored
as basic verbs and kept in the dictionary. Another speciality of Kannada is more than
one aspect marker can be attached; this is not the case with English. Consider an
example below.

In this example, 4 aspect markers are added; one can observe the complexity in
word formation with respect to Kannada, the meaning of this word is inferred from
first verb maaDu ‘do’. TAM inflections, PNG (person, gender, number) inflections
follow last verb biDu (Leave). From the CIIL (central institute of Indian language)
3 million corpus, different auxiliaries are explored and are given in Table 1.
444 B. C. Melinamath

3.3 Modal Auxiliaries

Modal auxiliaries contribute different shades of grammatical meaning. The various

possible modal auxiliaries are shown in Table 2.
Modal auxiliaries are always attached at infinitive ‘al’ form.

Table 2 Modal auxiliaries verbs in Kannada

Handling of Auxiliaries in Kannada Morphology 445

The modal auxiliary in example 7 and 8, aagu and paDu denote passive
constructions.

4 Observations and Conclusion

A Kannada Sample file Account4.aci.out the file from DoE CILL Corpus is selected
for analysis, and it is observed that many words in the corpus are formed with
auxiliaries and also modal auxiliaries. The analysis varies with the type of corpus.
The occurrence of various auxiliaries in the corpus is shown in Fig. 2.
In the above corpus the auxiliary haaku ‘put’ has occurred maximum followed by
nooDu and koDu, auxiliaries biDu and hoogu have not occurred at all. The occur-
rences various with the type of corpus used for testing. Similarly, the occurrences of
modal auxiliaries in the file is also tested and shown in Fig. 3.
Randomly five files F1, F2, F3, F4, F5 of size around 1000 words is taken for
testing, and the result of auxiliary verbs is shown in Fig. 4. It shows that auxiliary
beeku ‘must” have occurred maximum times followed by aagu, paDu andkuuDadu
has not occurred. This variation depends on the corpus.
The morphology of Kannada becomes complex due occurrence of auxiliaries as
bound morpheme with the root or derivative stem. In contrast, they occur as free
morphemes in English. We observe that 5 to 6 levels of aspect marker can be added
to the derivate stem to form a complex verb root in Kannada. Auxiliary verbs help
the main verb to denote the actions of the subject. They help in making compound
words and passive voice statements. Auxiliaries are useful in serial verb construction.

Fig. 2 The occurrence of Aspect Auxiliaries

aspect auxiliaries in file 45
Account4.aci.out
Occurance of Auxiliaries

40
35
30
25
20
15
10
5
0
aaDu biDu hoogu koDu nooDu haaku
Aspect Auxiliaries
446 B. C. Melinamath

Modal Auxilaries
60

No Of Modal Auxilaries
50
40
30
20
10
0

Modal Auxiliary Types

Fig. 3 The occurrence of modal auxiliaries in file Account4.aci.out

Fig. 4 Aspect auxiliaries distribution CILL corpus sample files

Morphology of auxiliary verbs like aspect auxiliaries and modal auxiliaries play
a very important role as far as a Dravidian language like Kannada is considered.
The morphological richness of any language is the presence of auxiliaries as bound
morphemes.

References

1. Singh DM, Seema S (2012) FST based morphological analyzer for Hindi language. J Comput
Sci Issues 9:349–353
2. Akmajian A, Steele SM, Wasow T (1979) The category AUX in universal grammar. Linguis Inq
10:1–64
3. Chomsky N (2001) Derivation by phase. In: Kenstowicz M (ed) Ken hale: a life in language.
MIT Press, Cambridge, MA, pp 1–52
4. Emonds J (1978) The verbal complex V’-V In French Linguistic Inquiry 21
5. Heine K (2001) The serial verb constructions and other complex predicates
Handling of Auxiliaries in Kannada Morphology 447

6. Wurmbrand S (2012) The syntax of valuation in auxiliary-participle constructions. In: Coyote

working papers: proceedings of the 29th west coast conference on formal linguistics. University
of Arizona, Tucson
7. Rouveret A (2012) VP ellipsis: phases and the syntax of morphology J Nat Lang Linguist Theory
30:897–963
8. Marantz A (1988) Clitics, morphological merger, and the mapping to phonological structure.
In: Hammond M, Noonan M (eds) Theoretical morphology. Academic Press, New York, pp
253–270
9. Melinamath BC (2014) Morphological generator for Kannada nouns and verbs. J Adv Res
Comput Sci Softw Eng 4(4). ISSN: 2277 128X
Dysarthria Detection Using
Convolutional Neural Network

Pratibha Dumane, Bilal Hungund, and Satishkumar Chavan

Abstract Patients suffering from dysarthria have trouble controlling their muscles
involved in speaking, thereby leading to spoken speech that is indiscernible. There
have been a number of studies that have addressed speech impairments; however addi-
tional research is required in terms of considering speakers with the same impairment
though with variable condition of the impairment. The type of impairment and the
level of severity will help in assessing the progression of the dysarthria and will also
help in planning the therapy.This paper proposes the use of Convolutional Neural
Network based model for identifying whether a person is suffering from dysarthria.
Early diagnosis is a step towards better management of the impairment. The proposed
model makes use of several speech features viz. zero crossing rates, MFCCs, spectral
centroids, spectral roll off for analysis of the speech signals. TORGO speech signal
database is used for the training and testing of the proposed model. CNN shows
promising results for early diagnosis of dysarthric speech with an accuracy score of
93.87%.

Keywords Dysarthria · Convolutional Neural Network · Speech Disorders ·

Speech Intelligibility · Speech Analysis · Machine Learning

1 Introduction

Dysarthria is a physiological condition caused by weakness, paralysis, flexibility,

spasticity, sensory loss, incoordination of muscles responsible for producing speech.
Dysarthria is a neurogenic speech issue that can influence any of the following subsys-
tems required for producing speech: Respiration, Phonation, Resonance, Articula-
tion, and Prosody. These speech disorders are brought about by a pathology that
affects the nervous system. Most neurogenic disorders fit into one or the other

P. Dumane · S. Chavan
Don Bosco Institute of Technology, Kurla (W), Mumbai 400070, MS, India
B. Hungund (B)
Narsee Monjee Institute of Management Studies, Vile Parle, Mumbai 400056, MS, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 449
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_45
450 P. Dumane et al.

dysarthria types. Signs and indications of dysarthria differ based upon the cause and
the type of dysarthria. These disorders may lead to symptoms like slurred speech, slow
speech, inability to speak louder than a whisper or speaking too loudly, rapid speech
that is difficult to understand, nasal, raspy or strained voice, uneven or abnormal
speech rhythm, difficulty in moving tongue or facial muscles.
Dysarthria can be caused by damage to the nervous system, and because of the
intertwined intricate relationship between the different functioning elements, it is
difficult to consider any of the functions like respiration, phonation, articulation inde-
pendent of each other. Respiration is controlled by abdominal muscles, by thoracic
musculature and by the diaphragm. Phonation is controlled by the laryngeal struc-
tures. Resonance is determined by pharyngeal and oral musculature, by the soft
palate, by the tongue and by the nasal cavities. Articulation is under the control of
tongue, jaws and lips [1].
The differing degrees of defects are an outcome of the problems at one or more of
these control points because of which velocity, range and amplitude of the movement
are altered. Based on the corresponding site of neural impairment, dysarthria can be
classified into the types viz. flaccid dysarthria, spastic dysarthria, ataxic dysarthria,
hypo kinetic dysarthria, hyperkinetic dysarthria, mixed dysarthria [2].
Production of speech sounds involves a series of internal and external events
that occur at different levels in the auditory system. Decision support for planning
the therapy and also improving the automatic recognition of dysarthric speech is
possible with the assessment of the level of severity of dysarthria [3]. Convolutional
Neural Network (CNN) is a promising tool as presented in this paper that can be
used effectively for detection of dysarthria as compared to the traditional methods.
The manuscript is organized as: Sect. 2 discusses the dysarthria related previous
work. Section 3 covers the proposed methodology used for developing the classi-
fication model using CNN. Results are discussed in Sect. 4. Section 5 gives the
conclusion.

2 Literature Review

Literature suggests that the dysarthric speech severity classification according to the
types of dysarthria can be improved if the automatic assessment of dysarthric speech
can be done. The first step towards this is to identify whether the speech data is
that of a dysarthric or a non-dysarthric person. In [3], 31 audio descriptors and 10
feature sets are used for classification of the severity level of dysarthria. Spectral and
harmonic features were estimated using multi-taper based spectral estimation. Balaji
and Sadashivappa [4] have compared the speech waveforms of persons by mapping
the unintelligible speech data with the intelligible data. Genetic algorithms have been
used for selecting speech disorder specific prosody features in [5] and support vector
machines were used for diagnosis and assessment of dysarthric speech. Spangler
et al. [6] used the XGBoost classification algorithm for a fully automatic process for
detection of dysarthria. Kim et al. [7] utilized the KL-HMM framework composed
Dysarthria Detection Using Convolutional Neural Network 451

of DNN acoustic modeling and categorical distribution-based probabilistic lexical

modeling for detecting dysarthric speech. Millet and Zeghidour [8] have proposed
a fully learnable audio frontend that combines the time-domain filter banks and per
channel energy normalization for detection of dysarthria.
CNN started gaining prevalence in audio processing as the spectrogram image is a
visual representation of the audio frequency which can be used to extract more unique
features from a matrix spectrum through which classification can be performed. CNN
outperforms other models in audio classification. Spectrograms generate different
pattern representations which can be easily detected using CNN [9]. The changes in
audio depressions can be easily detected through CNN models, because the important
features are extracted from the Mel Frequency Cepstral Coefficients (MFCC) spec-
trogram [10]. Several studies use CNN models to optimize the model and increase
the accuracy of the model by reducing the errors in the prediction [11].

3 Proposed Methodology

The research was segregated into four sections—data collection and preparation,
data analysis, data preprocessing and data modeling as shown in Fig. 1.
Step 1: Data Collection and Preparation. Universal Access dysarthric speech
corpus and the TORGO database [12, 13] was used for training and testing the model
developed using CNN. This database consists of dysarthric articulation of aligned
acoustics and measured 3D articulatory features from Female and Male speakers. The
data consists of recordings for both dysarthric and non-dysarthric persons mainly to
identify the hidden articulatory parameters using statistical pattern recognition.The
data consists of.wav files of both dysarthric females and males, and non-dysarthric
females and males. The sample data files with their details are given in Fig. 2.
Step 2: Data Analysis. In addition to auditory speech evaluation, some other
techniques of analysis of the speech signals like detecting the changes in the
vowel formants, f0, changes in duration of the speech signal, amplitude variations,
prolonged vowel duration, prolonged voice onset time, variations in tempo, etc.
Reduced values of f1 and f2 are also seen to be associated with patients of dysarthria
as a result of reduction in tongue movements [14]. In this paper, the features used for
assessment are short time Fourier transform, spectral centroids, spectral bandwidth,
spectral roll off, zero crossing rate, Mel Frequency Cepstral Coefficients (MFCC).
The plots of these features for a female with dysarthria are presented in Fig. 3 as a

Fig. 1 Block schematic of methodology

452 P. Dumane et al.

Fig. 2 Sample data file with gender and syndrome details

sample. The features extracted from males and females suffering as well as those not
suffering from dysarthria were studied.
Step 3: Data Pre-processing. As preprocessing of the data, the unstructured data
was converted into numeric features considering 128 sample rates. The audio signals
can be converted into a Mel Spectrogram. Furthermore, to train the model, the target
variable which was whether the person is suffering from dysarthria or not is converted
into categorical variable followed by one-hot encoding. The input with 8216 audios
with 128 sample rate features (columns) and 2-Labelled target variables was prepared
to model the data and train the convolution model. These data are further split into
80% data for training and 20% data for testing the model.
Step 4: Data Modeling. The pre-processed data is provided as an input to the
first CNN layer of the model [9–11]. For each audio file, the data was separated
with the shape (16, 8, 1), thus the first CNN with 64 filters had an input of (64 ×
16 × 8) with the kernel (3 × 3). This layer generated 640 parameters and shrunk
the height and width dimensions and was used to identify the dysarthria or non-
dysarthria patterns from the audio. These layers generate several patterns which will
further pool the features using max-pool method. In max-pool of two-dimensional
method, the important unique features are identified and are further fed to the second
convolution layer with 128 filters of 8 × 4 shrink image. The second convolution
layer and max-pool layer has a 3 × 3 and 2 × 2 kernel respectively to shrink the
height and width of image, the input to this layer was (128 × 8 × 4). At the end
of the convolution layers, a dropout layer was fitted to overcome the problem of
over-fitting and under-fitting, with the frequency rate of 0.1. To complete the model,
a dense layer was fitted to make a classification problem. Before that, a flatten layer
was added which converts the 3D tensor into 1D tensor for dense layer. In the model,
the second convolution and max-pool layer generates an output shape of 128 × 8 ×
4. Hence to flatten the 3D tensor, these shapes generated 1024 parameters as an input
to first dense layer of 1024 neurons. At the end of the model, 2 neurons of dense layer
were added as an output layer of the model with the softmax activation function. The
presented model in Fig. 4 was built using TensorFlow library. The summary of the
model is given in Table 1. To compile the model, various hyper-parameters were
Dysarthria Detection Using Convolutional Neural Network 453

Fig. 3 Speech analysis (Type 1: Female with Dysarthria) a Original speech b Short term fourier
transform (STFT) c Spectral centroids d Spectral bandwidth (p = 1, p = 2, p = 3) e Spectral roll-off
( Zero crossing rate g MFCC spectrogram[Sentence: You wish to knowall about my grandfather]
454 P. Dumane et al.

Fig. 3 (continued)

Fig. 4 The area under the

curve of non-dysarthria
samples
Dysarthria Detection Using Convolutional Neural Network 455

Table 1 Summary of layers

Model: “sequential”
and trainable parameters in
CNN model Layer (type) Output shape Number of Parameters
Convolution 1 16 × 8 × 64 640
Max. Pooling 1 8 × 4 × 64 0
Convolution 2 8 × 4 × 128 73,856
Max. Pooling 2 4 × 2 × 128 0
Dropout 4 × 2 × 128 0
Flatten 1024 0
Dense 1 1024 1,049,600
Dense 2 2 2,050
Total Parameters 1, 126, 146
Trainable 1, 126, 146
Parameters
Non trainable 0
Parameters

Table 2 Parameters of CNN

Parameters Value
model
Optimizer Adam
Learning rate 0.0005
Loss function Categorical cross entropy
Epochs 30
Batch size 50

tuned and Adam optimizer was built, with a learning rate of 0.0005. The parameters
of the model are described in Table 2.

4 Results and Discussion

It can be observed that after 10 epochs, the training loss starts decreasing and holds
a linear loss, though there were some fluctuations observed in between 11 and 20
epochs. Similarly, it can be observed that testing losses are fluctuating for each epoch.
This may be due to over fitting or under fitting of the model. The call-back function
was implemented to stop training the model as the testing set achieves the lowest
loss. Furthermore, it can be observed that after 30 epochs of training of the model,
it gives training accuracy of 96.59% and testing accuracy of 93.87%. Thus it can be
considered that CNN model does identify the unique features among the voice data
to classify the given data into that of dysarthria or non-dysarthria patients.
456 P. Dumane et al.

Table 3 Confusion matrix

Actual/Prediction Dysarthria Non-dysarthria
Dysarthria 756 51
Non-dysarthria 75 1172

Table 4 Performance evaluation of the proposed dysarthria detection system

Parameters Accuracy Recall Precision AUC ROC
Score (%) 93.87 93.98 95.83 98.46

The model which was built to classify the given audio into a dysarthria or non-
dysarthria patient achieved a standard accuracy on a testing set. These sets were
unseen by the model. It can be observed that out of the 2054 random samples of
audios, 1928 audios are correctly predicted and 126 are not correctly predicted,
this can be inferred from the confusion matrix given in Table 3. The errors for
not predicting the labels correctly infers that 75 audios which were actually non-
dysarthria persons are predicted as suffering from dysarthria, similarly for 51
dysarthria patients are predicted as non-dysarthria persons. The reason for the errors
in the model could be because there were certain sentences which were very difficult
to classify, for example some sentences are in the form of only one word—‘sleep’.
The precision of predicting the correct labels is 95.83%. The evaluation metrics for
the model predicted on the test set are presented in Table 4.
The model assures that the area under the curve of non-dysarthria samples guar-
antees 98.46% to correctly identify the actual labels. This can be observed from
the confusion matrix that out of total 1247 non-dysarthria voices 1172 are correctly
predicted signifying the highest ratio. The AUC graph is displayed in Fig. 4.

5 Conclusion

With advancements in the field of artificial intelligence, deep learning models are
useful for decision support systems in the diagnosis of patients suffering from
dysarthria. Disorders like dysarthria not only create trouble for the patient in
conveying their thoughts but also bring in social embarrassment for the patient which
adds to the mental trauma. CNN based model definitely holds promise for helping
in the diagnosis of severity of dysarthria. Analysis helps in planning the therapy and
can also help in recognition of dysarthric speech. In this paper, the accuracy score
of 93.87% has been obtained in the classification of dysarthric speech using CNN
model. Future work involves considering more number of features and correlating
them for classifying the severity of dysarthria. In spite of some clear evidences, there
are some misclassifications which need further investigation.
Dysarthria Detection Using Convolutional Neural Network 457

References

1. Edwards M (2012) Disorders of articulation: aspects of dysarthria and verbal dyspraxia, vol 7.
Springer Science & Business Media
2. Hodge M (2013) Developmental dysarthria (2013). Cambridge Handbooks in Language and
Linguistics. Cambridge University Press. https://doi.org/https://doi.org/10.1017/CBO978113
9108683.004
3. Bhat C, Vachhani B, Kopparapu SK (2017) Automatic assessment of dysarthria severity level
using audio descriptors. In: IEEE International conference on acoustics, speech and signal
processing (ICASSP). IEEE, pp 5070–5074
4. Balaji V, Sadashivappa G (2019) Waveform analysis and feature extraction from speech data of
dysarthric persons. In: 2019 6th international conference on signal processing and integrated
networks (SPIN). Noida, India, pp 955–960. https://doi.org/10.1109/SPIN.2019.8711768.
5. Vyas G, Dutta MK, Prinosil J, Harár P (2016) An automatic diagnosis and assessment of
dysarthric speech using speech disorder specific prosodic features. In: 2016 39th international
conference on telecommunications and signal processing (TSP). Vienna, pp 515–518. https://
doi.org/10.1109/TSP.2016.7760933.
6. Spangler T, Vinodchandran NV, Samal A, Green JR (2017) Fractal features for automatic
detection of dysarthria. In: IEEE EMBS international conference on biomedical & health
informatics (BHI). Orlando, FL, pp 437–440. https://doi.org/10.1109/BHI.2017.7897299.
7. Kim M, Kim Y, Yoo J, Wang J, Kim H (2017) Regularized speaker adaptation of KL-HMM for
dysarthric speech recognition. IEEE Trans Neural Syst Rehabil Eng 25(9):1581–1591. https://
doi.org/10.1109/TNSRE.,.2681691
8. Millet J, Zeghidour N (2019) Learning to detect dysarthria from raw speech. In: 2019 IEEE inter-
national conference on acoustics, speech and signal processing (ICASSP). Brighton, United
Kingdom, pp 5831–5835. https://doi.org/10.1109/ICASSP.2019.8682324A.
9. Khamparia A, Gupta D, Nguyen NG, Khanna A, Pandey B, Tiwari P (2019) Sound classification
using convolutional neural network and tensor deep stacking network. IEEE Access 7:7717–
7727
10. Wang Z, Chen L, Wang L, Diao G (2020) Recognition of audio depression based on
convolutional neural network and generative antagonism network model. IEEE Access
8:101181–101191
11. Sasmaz E, Tek FB (2018) Animal sound classification using a convolutional neural network.
In: 2018 3rd international conference on computer science and engineering (UBMK). IEEE,
pp 625–629
12. The TORGO Database (2020) Acoustic and articulatory speech from speakers with dysarthria,
https://www.cs.toronto.edu/compling-web/data/TORGO/torgo.html
13. Rudzicz F, Namasivayam AK, Wolff T (2012) The TORGO database of acoustic and
articulatory speech from speakers with dysarthria. Lang Resour Eval 46(4):523–541
14. Refai MS, Aziz AA, Osman DM, El-Shafee SF et al (2012) Assessing speech intelligibility in
a group of egyptiandysarthric patients. Egypt J Otolaryngol 28(1):49
AR for Maintenance Training During
COVID-19 Pandemic

Jyoti Pawar and Trupti Bansode

Abstract COVID-19 situations are getting worse day by day. Buying behaviors
and economic priorities for purchasing new goods have changed. In these days,
maintenance of existing equipment has gained more importance. For the purpose of
training the maintenance workers or end users in this pandemic situation, augmented
reality is like a blessing today. The maintenance procedures can be communicated
remotely on demand with the help of web based AR and android apps. This paper
describes advantages of using AR based instruction over paper based instruction to
train maintenance operators. The authors have developed web based AR and android
apps for augmenting 3D models. Many benefits observed by using this technology
are reduced human errors reduced execution time, reduced downtime of equipment,
reduced cost, and increased productivity. Paper mainly focuses on ways of delivering
maintenance instructions remotely.

Keywords Augmented Reality · Virtual reality · Maintenance · COVID-19

1 Introduction

The COVID-19 pandemic has changed the planet forever, and consequently
the impacts from this are far-reaching. Globally, companies are pressured to under-
take strict work from home policies whenever possible, and outright cessation of
operations while social distancing and remote work weren’t possible. Uniquely, this
become real in each marketplace in each nation. Today, a few markets and areas are
returning to normalcy, at the same time as others are persevering with to conform
to the brand new normal of labour. Interestingly, COVID-19 did not present many

J. Pawar (B)
SVERI’s College of Engineering (Poly.), Ranjani Road, Pandharpur, Maharashtra 413304, India
e-mail: [email protected]
T. Bansode
COER, Fabtech Technical Campus, Sangola, Maharashtra 413307, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 459
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_46
460 J. Pawar and T. Bansode

new demanding situations for AR and VR implementations, however it alterna-

tively emphasized capacity and value. When inspecting the modifications necessi-
tated through the pandemic, many overlap considerably with excessive value-brought
answers inside AR/VR which have existed for years [1]. The COVID-19 pandemic
has impacted the AR in preservation enterprise positively. Augmented fact is gaining
importance inside the preservation region in spite of COVID-19 pandemic due
to numerous technological developments. This paper makes a specialty of bringing
up advantages of AR in COVID-19 situations, exploring the utilization instances of
AR and giving a comparative evaluation of AR primarily based totally and paper
primarily based totally practice for preservation operators. Next segment briefs tech-
nical history and use instances of AR. AR based instructions provided to mainte-
nance operator by the system [2]. Then authors have offered effects of comparative
evaluation and explained the advantages of the usage of AR in in addition sections.

2 Related Works

Training maintenance professionals is a crucial part for any industry. Many authors
have worked to study the use of Augmented reality in maintenance training. A review
paper [3] has collected some publications in this regard out of which 18% have
used air for the Aeronautics industry. They have reported that 34% of papers have
described the use of air for assembling the equipment. In paper [4] a pilot study of
Aeronautical engineers mentioning interview based research has explained the use
of wearable Technology. The study reports that some maintenance tasks require less
time if you use AR based applications. In paper [5] the industrial aspect of using
AR is reported with 91% improvement in quality of maintenance work and 32%
improvement it is achieved in required time to complete a manufacturing operation.
These studies have focused on air for maintenance and proved that are significantly
reduces time and improves performance. Some researchers have also proven the
usability of AR in training and education.
For conveying instructions of routine maintenance of Desktop machines or laptops
regarding replacement of Ram chip or hard disc drive AR based solution is much
beneficial. AR has wide acceptance as educational technology in many educa-
tional settings including school going kids for undergraduate students. As a part
of maintenance training this smart Technology has proved its ability.
AR for Maintenance Training During COVID-19 Pandemic 461

3 AR as Immersive Technology

3.1 Augmented Reality

History. The study of augmented reality (AR) can be traced back to Ivan Suther-
land’s invention of the head-mounted display (HMD) in the 1960s. But it gained
a lot of attention after 1990, when Tom Caudell coined the term. Furthermore, for
many years, AR was constricted to academic research and entertainment through
sci-fi films. Early AR systems were mostly experimental and developed focusing on
specific tasks, such as maintenance and repair tasks. More recently, AR has started
to successfully step away from laboratories and appear in a wide range of applica-
tions, which include advertising, education, for advanced driver-assistance systems,
advertising, defense, manufacturing, networking, tourism, medicine, smart cities, for
social media apps that use face detection [6].
According to Gartner research, AR is poised to become an important workplace
tool. Many companies are increasingly applying AR tools to create effective training
programs to increase employee engagement, reduced efforts for actual training and
to achieve success in the highly competitive global marketplace.

3.2 Use Cases of AR During COVID-19

AR for remote assistance: Remote assistance has been the leading use case in AR
for years. Quick and reliable ROI, through both direct travel cost reduction and
increased efficiency through instant knowledge access (and, secondarily, reduced
downtime). Adding value to a process that, at best, was a video call on a mobile
device or, at worst, a traditional telephone call or an asynchronous support ticket,
has been straightforward for AR on both head-worn and mobile devices. Annotation,
spatial awareness, automatic data capture and creation, visualization integration, and
step-by- step instruction all have proven value on both device types.
Remote work is unique in its universal applicability. Internal usage for the enter-
prise sector is relatively straightforward, but customer-facing use cases are growing.
Post- sales support for devices or services has been an increasingly valuable use case,
with AR providing a level of visualization and easy-to-follow guidance not possible
with existing means. Telehealth and telemedicine fit in a similar way.
AR for training: Related to remote work and assistance, training in AR and
VR has been a powerful value driver for companies. Realistic simulation has had
a home in VR, while more real-world training and knowledge transfer fits well
in AR. The capability for in-situ training, where training and workflow accomplish-
ment happen simultaneously, thanks to AR, allows instantaneous employee ramp-up.
With a renewed focus on off-site capabilities, training content can be delivered and
consumed wherever a user is located, with proper visualization and simulation of the
target workflow environment available through the device.
462 J. Pawar and T. Bansode

Data capture: Again tying into remote work and training, data capture in AR
has not seen as much direct activity as training and remote assistance, but it is being
positioned as a promising longer-term approach to content capture, creation, and
distribution. Today, there is an increased emphasis on the quick and easy capture of
data and, more broadly, workflows to maintain efficiency with a reduced workforce
and enable more seamless knowledge capture and transfer. Capturing and logging
data through AR serves as a connection point between data-driven Internet of Things
(IoT) platforms and human workers: data from workers, who without that connection
point would be lost, can now be fed into the system and fully leveraged for analytics
and other uses.
Hands free operation: With head-worn devices, hands-free data access and inter-
action is possible. Input methods like voice, gesture, and gaze can completely elim-
inate hands-on interaction for both the AR/VR device and anything the device can
access. Prior to the pandemic, hands-free was a valuable assurance of an uninter-
rupted workflow and employee safety. Today, there is an added layer of sanitation
potential anywhere head-worn devices are leveraged.

4 AR for Remote Assistance in Maintenance Operation

To design AR based maintenance procedure one needs to consider the suitability

of the maintenance task. According to augmented reality software provider PTC,
not all maintenance tasks are suitable for augmented reality [7]. The procedure-
based maintenance-related tasks are best suitable for augmented reality training. For
example, maintenance tasks such as how to replace a key part on a critical asset
can be accurately and efficiently taught using AR. However, general maintenance
information or data does not require this smart technology.
Comparison of paper based instruction and AR based instruction for mmain-
tenance operator
For a procedure based task where printed manuals are used, the scenario can be imag-
ined similar to this one a worker is performing certain regular checkup of equipment
for the sake of preventive maintenance. If you find some abnormal condition with
the equipment or some suspicious behavior at a process output he has to find the
instruction manual ok all the supervisor and wait for the instruction. To ensure that
the equipment is working properly he has to study the manual. But if an AR based
manual is provided to such situations, he simply needs to carry a smart phone or
display device. With the empowerment of AR, the employee is able to interact with
the industrial equipment.
Since AR technology operates in a real environment and generates relevant health
information runtime to enhance Real world experience, the technology is well suited
to transfer skill sets within the employees. It is much better approach than the old-style
training methodologies which used printed manuals, videos etc.
AR for Maintenance Training During COVID-19 Pandemic 463

With an AR-based solution, this dependency on hard copy instructions or user

manuals is eliminated. The understanding of how to fix issues is visually communi-
cated to the engineer. A step-by-step, three dimensional overlaid instruction manual
over real equipment guides engineers to fix critical machine-related issues.
AR allows custom-designed training and maintenance solutions. These solutions
when developed with AR allow complex procedures to be animated directly on
the equipment. With these augmented instructions and reference materials overlaid
directly on the physical equipment, workers can learn procedures more effectively
and perform them more accurately.

5 Results and Discussion

In AR, the virtual objects are superimposed to the real word so that by moving the
point of view of the camera, the symbol positions with respect to the external reference
system does not change. This is obtained by computing the position of the camera
with respect to the external environment. AR based applications are of two types-
one that used markers and other that use objects or specific locations to superimpose
the virtual content. The marker based AR experiences have the facility to run over
any equipment like laptop or handheld device but they have a limitation that the user
should have the printed marker available for generating augmented content in the
camera view. AR.JS is the library which can provide marker based augmentation in
the web browser. Another type of AR experiences uses Android apps specifically
designed to use camera Input and location information to decide the content for
augmentation. Some recent libraries like AR core from Google identify the plane
surface and place the augmented model on that plane surface. In some cases such
AR experiences need extra hardware facilities. The authors have developed market-
based AR experience for augmentation of 3D model and video for web application
based interface. For handheld device based interface, they have developed an Android
application which can augment 3D model of a measurement scale and maintenance
device called Hand held electric saw.
Fig: 1 shows Snapshot of web based application using model-viewer component.
Using model-viewer allows us to augment virtual content without marker. Fig: 2 is
Snapshot of a 3D model of a cube and polygon over a marker in live webcam feed.
Fig: 3 Snapshot of augmentation of 3D model over a marker provided by Google
Poly recorded using browser in a handheld device.The libraries used for development
of web applications using markers shown in figure are AR.js and Aframe.js.
Fig: 4 Snapshot of three dimensional augmented keyboard in real time using a
mobile app. Fig: 5 is the screenshot for Mobile Augmented reality using android
app. It displays a virtual desktop computer placed near a wall. Fig: 6 snapshots of 3D
model of handheld electric saw augmented in real space with mobile app. This app
can be further developed to augment the series of instructions for using that electric
saw for the purpose of maintenance. These instructions will provide the operator
directions to use and operate the saw in real time. The operator is able to see the
464 J. Pawar and T. Bansode

Fig. 1 Snapshot of web-based AR using model viewer

Fig. 2 Snapshot for

web-based AR using
three.ar.js

3d model and the instructions are provided in the same window where the model
appears.

6 Future Scope and Conclusion

In this paper we have discussed the use cases for AR and their impact during COVID-
19 situations. The system described here will be very much useful and it will surely
prove its capacitance while training maintenance workers.
AR for Maintenance Training During COVID-19 Pandemic 465

Fig. 3 Snapshot of
Marker-based AR using 3d
models available on Google
Poly

Fig. 4 Snapshot of three-dimensional virtual keyboard augmented in real-time using a mobile app

A lot of possibilities are coming into picture if we combine augmented or virtual

reality with the latest technologies like IoT and machine learning. This can help to
facilitate less requirement of labour energy and labour safety and improved training
outputs. The authors are working on combining computer vision methods applying.
Deep learning algorithms to combine with AR for further improvements. Putting
efforts to make it possible for identifying the asset and generating smart instructions
for assets will allow for less human error, increase safety, and also allow technicians
to walk step-by-step on repair processes. Many of the barriers to adoption seen in
466 J. Pawar and T. Bansode

Fig. 5 Snapshot of virtual desktop computer augmented to teach computer maintenance

Fig. 6 Snapshots of 3D
model of handheld electric
saw augmented in real space
with mobile app

the enterprise market come down to device cost and ease of use; these issues are
the same in the consumer market. Many of the first wave of consumer AR devices
have prioritized keeping prices down, while delivering on promised value. Google
Glass was technically launched as a consumer product, but the price and lack of
clarity around value quickly pushed the device to the enterprise market where it
found purchasers.
Looking ahead, things become both more clear-cut and more complicated. Proven
value in work-from-home and remote collaboration extends beyond the COVID-19
impact. A consistent increase in remote work and tele presence usage overall naturally
includes AR/VR, and as companies and customers return to normal, much of this
AR for Maintenance Training During COVID-19 Pandemic 467

adaptation will persist. The pandemic forced pilots for remote assistance and other
AR use cases for many previously hesitant to adopt immersive technology. These
forced programs will encourage future investment, serving as a kick start for the
market that had been occurring slowly but surely over time.

References

1. Gattullo M, Uva AE, Fiorentino M, Gabbard JL (2015) Legibility in industrial AR: Text style,
color coding, and illuminance. IEEE Comput Graph Appl 35(2)
2. Azuma RT (1997) A survey of augmented reality. Presence: Teleoperators Virtual Environ 6:355–
385
3. Palmarini R, Erkoyuncua JA, Roy R, Torabmostaedi H (2018) A systematic review in augmented
reality applications in maintenance. Rob Comput Integr Manuf 49:215–228
4. Robertson T, Bischof J, Geyman M, Ilse E (2017) Reducing maintenance error with wearable
technology. IEEE
5. The Area website (2020) https://thearea.org/augmented-reality-can-increase-productivity/.
[Accessed: June 2020]
6. The PTC website (2020) https://www.ptc.com/en/service-software-blog/augmented-reality-mai
ntenance-and-repair. [Accessed: June 2020]
7. Airbus Company website (2020) https://airbus-xo.com/real-benefits-virtual-reality/ [Accessed:
June 2020]
Face Recognition Based Attendance
System Using Cv2

Vedant Khairnar and C. M. Khairnar

Abstract Face recognition based attendance based system will be used in the near
future in classrooms instead of the traditional system; it may replace even biometric
attendance systems. The purpose of the present work is to devise a novel attendance
system using cv2. Facebook also uses face recognition technology as it tags the names
of faces as soon as you upload photos which have been tagged by you previously. The
algorithm identifies the unique features of the faces in the database and encodes them
into pattern image. Python modules are used to Then the machine learning algorithm
called classifier is used to find the name of the person. Image capture, facial features,
face recognition and attendance system, are the stages of the procedure.

Keywords Deep learning · Face recognition · Attendance

1 Introduction

In this paper, face recognition is done with the help of Python programming. All the
faces are initially taken in image form and folder is prepared. The Names are also
given to each image. Facial features are found so that the database is ready.
Once the database is ready, The camera is used to take attendance, and its output
is used along with a database for face detection and face recognition.
After face detection and face recognition, the program prints the result immedi-
ately after the program is run. Also, another program loop can express the output in
separate excel sheet with names, present, absent, date and, time details.

V. Khairnar
Ramdeobaba College of Engineering, Nagpur, Maharastra, India
C. M. Khairnar (B)
Government College of Engineering, Nagpur, Maharastra, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 469
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_47
470 V. Khairnar and C. M. Khairnar

2 Face Detection and Attendance System Review

Algorithm of Viola and Jones, Principles analysis are used [1]. The first step used is
to detect the face and its recognition, after this the classifier program has compared
the faces with the database of students faces.
Optical flow field [2] method is used here in this study of the face detection; it
uses objects, comparatively a new method for recognition and detection.
The portable device [3] is proposed and developed here for the student attendance
system. The device is to be designed and optimized for camera, server and other
hardware is also to be customized.
A filtering system [4] based on Euclidean distances calculated by three face recog-
nition techniques, namely Eigenfaces, Fisherfaces and Local binary pattern has been
developed for face recognition.
Automated attendance management system [5] along with the techniques used to
handle the threats like spoofing.
The new type of touchless attendance system is studied [6] for the performance,
efficiency and accuracy. The efficiency depends on the hardware as well as the
database quality collected.
Touchless attendance system for institute or classroom is developed and studied
[7].
The present system [8] studied here is about face detection and used for the class-
room attendance using Discrete Wavelet Transforms (DWT) and Discrete Cosine
Transform (DCT), the facial feature extraction is easy by this method.
Deep Learning algorithm helps in detecting the faces in the classroom when
captured by high definition camera and then subjected to the classifier program [9].
The research [10, 11], includes for Face detection Students and the system is
based on CNN perspectives and algorithms.

3 Methodology

In all the biometric attendance process, the biometric features of the person are
compared with the database templates previously prepared at the time of enrollment
of all the persons. The methodology used in our research can broadly be divided into
the following procedural stages.

3.1 Hardware Requirements and Software Installations

Camera for the classroom having an optimum resolution. SSD memory for storage of
image database. Computer with a powerful graphics card. The cost of this attendance
Face Recognition Based Attendance System Using Cv2 471

system is mainly the cost of hardware like computer and camera. Software required
is windows/ Linux for execution. For development Face recognition library.

3.2 Image Capture

“OpenCV” Python library is used to capture the images generated by the camera.

3.3 Deep Learning: Face Detection and Recognition

Nowadays, the camera of the Smartphone also has the face detection feature if noted
carefully, as that feature is used to, focus on the faces and capture.
Another example of face detection is that Facebook also uses this feature. In case
we upload the group photo on Facebook, then it will automatically tag the faces in the
photos, if and only if you have previously tagged those faces with names. Facebook
remembers those tag names using machine learning and can detect them.
Feature base technique is used in this library for face recognition. Facial features
such as the nose, eyebrows, lips are used. The shape of these facial features is different
for different persons. Hence detections can be possible with the help of these deep
learning techniques.
The unique feature in the face, such as long/short nose, which distinguishes it
from other person faces is specially used to recognize and detect the face of a person.
Histogram of an oriented gradient, pattern image is used in this type of machine
learning. The algorithm matches the image captured by the camera with these pattern
images in databases, which highlight the special features of the person, and detects
the face of the person. This image ensures the detection of the face even if it is slightly
tilted/ oriented or in bad light conditions.
Then basic machine classification algorithms can easily find the names of the
person.
In our program “Face recognition” Python library built using deep learning, is
used for face detection and face recognition.

4 Algorithm

import datetime
from datetime import datetime

import cv2
import face_recognition
import numpy as np
472 V. Khairnar and C. M. Khairnar

import openpyxl

def addInExcel(d):
fp = “Attendance.xlsx”
wb = openpyxl.load_workbook(fp)
sheet = wb.get_active_sheet()

max_row = sheet.max_row
max_column = sheet.max_column

now = datetime.now()
# print(“Data Acquired :”, d)
sheet.cell(row=1, column=max_column + 1).value =
now.strftime(“%m/%d/%Y, %H:%M:%S")
for index in range(1, max_row):
sheet.cell(row=index + 1, column=max_column + 1).value =
d[sheet.cell(row=index + 1, column=2).value]
wb.save(fp
video_capture = cv2.VideoCapture(0)

obama_image = face_recognition.load_image_file("images/obama.jpg")
obama_face_encoding = face_recognition.face_encodings(obama_image)[0]
gates_image = face_recognition.load_image_file("images/gates.jpg")
gates_face_encoding = face_recognition.face_encodings(gates_image)[0]

known_face_encodings = [
obama_face_encoding,
gates_face_encoding
]
known_face_names = [
"Barack Obama",
"Bill Gates"
]
attendance_list = {"Barack Obama": "Absent", "Bill Gates": "Absent", " ": " "}
face_locations = []
face_encodings = []
face_names = []
process_this_frame = True

while True:
ret, frame = video_capture.read()
small_frame = cv2.resize(frame, (0, 0), fx=0.25, fy=0.25)
rgb_small_frame = small_frame[:, :, ::-1]

if process_this_frame:
attendance_list[name] = "Present"
process_this_frame = not process_this_frame
Face Recognition Based Attendance System Using Cv2 473

for (top, right, bottom, left), name in zip(face_locations, face_names):

top *= 4
right *= 4
bottom *= 4
left *= 4
cv2.putText(frame, name, (left + 6, bottom - 6), font, 1.0, (255, 255, 255), 1)

cv2.imshow(’Video’, frame)

# Hit ’esc’ on the keyboard to quit!

k = cv2.waitKey(30) & 0xff
if k == 27:
present = 0
absent = 0
print("------------------------------------------")
print("Attendance as per", datetime.now())
print("Name Absent/Present")
for i in attendance_list:
print(i, ":", attendance_list[i])

if attendance_list[i] == "Present":
present += 1
elif attendance_list[i] == "Absent":
absent += 1
print("Total Present :", present)
print("Total Absent :", absent)
print("------------------------------------------")
attendance_list[’Total Present’] = present
attendance_list[’Total Absent’] = absent

# print(attendance_list)
addInExcel(attendance_list)
Break

cv2.destroyAllWindows()
video_capture.release()
"""
------------------------------------------
Attendance as per 2020-05-06 09:48:11.096916
Name Absent/Present
Barack Obama : Present
Bill Gates : Absent
:
Total Present : 1
Total Absent : 1
474 V. Khairnar and C. M. Khairnar

Fig. 1 Input to the camera

5 Result and Discussion

The library used for getting the output in excel sheet is “Openpyxl” It is a Python
library for expressing the data in excel sheet.
The sample attendance for one person is taken. Database for two people is prepared
one with name Barak Obama and another with name Bill Gate, Fig. 1 shows the
screenshot capture for one person in front of the camera.
The output is shown in excel sheet generated using a Python program is shown
in Fig. 2. The excel sheet screenshot shows the present, absent, date and time of
the attendance taken. The efficiency of the program is 90%, hence can be used for
practical purposes.

6 Conclusion

Deep Learning algorithm helps in detecting the faces in the classroom when captured
by the high-definition camera and then subjected to the classifier program.
The future scope for the work can be refining the algorithm so that proxy
attendance can be detected. Also, Cloud Integration can be done.
Face Recognition Based Attendance System Using Cv2 475

Fig. 2 Output as an attendance sheet

References

1. Wagh P, Thakare R, Chaudhari J, Patil S (2015) Attendance system based on face recognition
using eigen face and PCA algorithms. In: International conference on green computing and
internet of things (ICGCIoT). Noida, pp 303–308. https://doi.org/10.1109/ICGCIoT.2015.738
0478
2. Bao W, Li H, Li N, Jiang W (2009) A liveness detection method for face recognition based
on optical flow field. In: International conference on image analysis and signal processing.
Taizhou, pp 233–236. https://doi.org/10.1109/IASP.2009.5054589
3. Bhattacharya S, Nainala GS, Das P, Routray A (2018) Smart attendance monitoring system
(SAMS): a face recognition based attendance system for classroom environment. In: IEEE 18th
international conference on advanced learning technologies (ICALT). Mumbai, pp 358–360.
https://doi.org/10.1109/ICALT.2018.00090
4. Samet R, Tanriverdi M (2017) Face recognition-based mobile automatic classroom attendance
management system. In: 2017 international conference on cyberworlds (CW). Chester, pp
253–256. https://doi.org/10.1109/CW.2017.34
5. Chintalapati S, Raghunadh MV (2013) Automated attendance management system based on
face recognition algorithms. In: IEEE international conference on computational intelligence
and computing research. Enathi, pp 1–5. https://doi.org/10.1109/ICCIC.2013.6724266
6. Surekha B, Nazare KJ, Viswanadha RS, Dey N (2017) Attendance recording system using
partial face recognition algorithm. In: Dey N, Santhi V (eds) Intelligent techniques in signal
processing for multimedia security. Studies in computational intelligence, vol 660. Springer,
Cham
7. Kar N, Debbarma MK, Saha A, Pal D (2012) Study of implementing automated attendance
system using face recognition technique. Int J Comput Commun Eng 1(2)
8. Lukas SM, Desanti AR, Krisnadi DRI (2016) Student attendance system in classroom using
face recognition technique. In: International conference on information and communication
technology convergence (ICTC). Jeju, pp 1032–1035. https://doi.org/10.1109/ICTC.2016.776
3360
476 V. Khairnar and C. M. Khairnar

9. Nandhini R, Duraimurugan N, Chokkalingam SP (2019) Face recognition based attendance

system. Int J Eng Adv Technol 8(3S)
10. Sanagala RK, Rathnam S, Babu RRV (2019) Face recognition based attendance system for
CMR college of engineering and technology. Int J Innovative Technol Exploring Eng 8(4S2)
11. Senthamil Selvi K, Chitrakala P, Jenitha AA (2014) Face recognition based attendance marking
system. Int J Comput Sci Mobile Comput 3:337–342
IoT Enabled Secured Card Less Ration
Distribution System

Shilpa K. Rudrawar, Kuldeepak Phad, and Prajwal Durugkar

Abstract Proposed paper put light on the automation in distribution of goods by

using IOT based ration distribution system which uses the biometric verification
and the cloud storage technology. Proposed system looks like an Automated Teller
machine (ATM). We can simplify the process by using an interactive approach.
Aadhar cards contain details like contact number, residential details, details of bank
account and available scheme. Details of customers are stored and maintained as the
database in the cloud storage by the government. Here we are storing the customer’s
data in the cloud using storage technology such as Google Firebase. This database
contains the necessary information such as Aadhar details, allotted goods, bank
account information, and ration card type. To carry out the transaction and the with-
drawal of goods one needs the One time password (OTP) which is sent through
the SMS or E-mail to notify and alert the user during the process. After entering
the OTP required amount regarding grains purchased will get deducted from the
linked bank account. Proposed system will help to minimize the issues like lack of
rationing material which was caused due to the smuggling goods by the ration shop
owner, workers and the dealers in order to gain profit. This system will be transparent
and will help customers by giving smooth and automized experience of purchasing
rationing. This proposed system will be the solution to One Ration One Nation.

Keywords Ration system · OTP · IoT · Cardless system · Biometric verification ·

Google firebase · Aadhar card

1 Introduction

It is not easy to distribute rationing in most populated countries like India. Second
highest populated countries like India there are some reserved castes and the people
under poverty line have been benefited with the livelihood commodities by the Public

S. K. Rudrawar (B) · K. Phad · P. Durugkar

Department of Electronics and Telecommunication Engineering, MIT Academy of Engineering,
Alandi (D), Pune, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 477
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_48
478 S. K. Rudrawar et al.

distribution system. Ration card holders of various types like BPL,APL and Antyod-
hana Anna Yojana have been benefited with the low price purchase of goods and the
food items like wheat, rice, sugar, cereal, Kerosene, edible oils etc. with the viable
amount by government schemes [1, 2].
Aadhar card is considered as the symbolic representation of citizenship of Indians.
There are lots of schemes that have been provided with this as it is linked at different
places such as banks, hospitals, rationing shops which contain all the details of
customers like identification marks, photo, name, economic status, religious status
[3, 4].
This current ration cards system can be replaced by card less system proposed in
the paper by proper implementation of hardware structure and creating the database
of all customers containing their personal/family details, ration card details, biometric
details and by linking it to Aadhar card which will be stored in cloud for maintaining
transparency [5]. The automated ration distribution system can be installed at the
ration shop which is having two layers of security. First is fingerprint verification
and the other is OTP.
Hardware structure has the fingerprint scanner; Load cell, Motors, speaker
and touch screen display as a user interface are interfaced to Raspberry Pi. The
System will be linked with the mobile phone and the government database. If some
unauthorized user tries to access an account, the authorized person will get alerted.

2 Literature Survey

In the present Public Distribution System (PDS), paper based ration cards are used
by the customers which are categories into three parts based on the economic status
of the family. This ration card needs to be renewed every year, which is a very time
consuming process. Nowadays it is replaced with the rationing machine known as
EPoS machine which is used to authenticate the user and generate a bill based on
the total material user has been allotted. Then the ration shop owner gives him the
utilities [6].
Dr. M. Pallikonda Rajesekan in his research paper “Automatic Smart Ration Distri-
bution System for Prevention of Civil Supplies Hoarding in India” proposed the
system used for distribution of ration with the automated system. Which takes the
manual input from the user for registration of ration cards and then for distribution
of material. He proposed the system which uses the GSM based communication and
AUTORAT system [7].
Padmashree S. in his research paper “Automatic Ration Material Dispensing
System” also worked on the similar kind of system. Here he has proposed a system
which works on the RFID technology for the authentication of users. RFID cards use
the radio frequency to detect the card of a particular user; each card has its own unique
identification number [8]. Madhu Kumar N. and his team used the combination of
both RFID and the GSM based technology for the authentication and communica-
tion with users [9]. Brendon Desouza states in his research work “Smart Ration Card
IoT Enabled Secured Card Less Ration Distribution System 479

Automation System” that the GSM based technology can be used to achieve the high
level of security in Public Distribution services. He developed the system by using
the real time embedded system by using ARM based microcontroller [5].
The system we are proposing here is totally card less system and Google firebase
is used to store the customer’s information as well as online updating the material
allotted to the customer once taken which makes the system more transparent and
avoids malfunctioning. Customers from any region can collect their allotted grains
from any other region location where this hardware is placed or fixed. So here we
addressed and provided a solution for One Nation One Ration scheme.

3 Block Diagram

The Fig. 1 shows the block diagram of the system and components required to run
the entire system. Each block work in hand with others to perform tasks.
The microcontroller unit incorporates the weight measurement system which
controls the motion of valves by providing signals to servo motors. Raspberry Pi
is the most important part of this system which acts like the heart of a human and
is used to perform all computational activities. By providing particular delay to the
opening and closing of the valve we can withdraw the required amount of material.

Fig. 1 System block diagram

480 S. K. Rudrawar et al.

• Raspberry Pi3 is a 32-bit processor that has two serial communication ports. One
of these ports is used for FINGERPRINT MODULE and GPIO pins are used to
connect different peripherals like Load cell, speaker and servo motors. 3.5 inch
capacitive touch screen interface is provided which shows the information on the
screen also used to take actions.
• A DC servo motor is used to control the outlet valve to dispense the rice and
wheat.
• After successful verification and entering OTP with the help of the keypad shown
on the screen, the transaction will be done.
Proposed system operates on 230 V input and 5 V, 1 A output to power the
raspberry pi board. The Raspberry Pi board is having support for the Linux operating
system on which we have different modules such as eSpeak (Google text to speech
for Linux), finger print module and python 3.7 to write the source code for the
application. Load cell not coming alone it requires HX711 is an SOC having 24- bit
ADC which is designed for weighing measurement and also used in industries for
control applications to the load cell which is used for weight measurement.
Load cells convert the forces applied into an electrical signal which is then
measured and converted into standard form. It will work on the principle of change in
resistance results in change in voltage [10]. Upon application of external stress it will
produce changing output voltage. Available in different shapes, sizes and numbers of
wires (either 4 or 6 depending upon distance between load cell and microcontroller).
The load cell is nothing but the four strain gauge which is configured in Wheatstone
bridge configuration with four different resistors. An excitation voltage up to 10 V
is applied to one end and the voltage is measured between the other two ends. It
is at equilibrium if no force is applied then the output is zero volts or near to zero
when all the resistors are having the same resistance and also known as balanced
bridge circuit. By the application of strain there is the change in resistance resulting
in the variable output voltage. Then this is noted and digitized then amplified the
small millivolts level signals into high amplitude signals. Load cell has four wires
(Two are excitation wires and other two are for output).positive or negative sign in
the output voltage of load cell is dependent upon location of strain applied. Whereas
HX711 has also four wires as VCC, SCK, DT and GND. When output is not fetching
any data then the digital output pin DT is at high. Serial input clock SCK must be
low.
Whenever DT is at 0, data sent by the application 25–27 positive pulses at the
SCK pin, data is shifted out from DT output pin. Each SCK pulse shifts from one
bit, starting with the MSB bit first, until each 24 bits is shifted out. After the 25th
pulse DT pin will pull back to high [11]. To avoid communication error, SCK pulse
should not be less than 25 and not more than 27 within one conversion period.
Fingerprint scanners are a light sensitivity transducer to produce digital images.
And they are available in different variations, many of them have external four pin
connection interface By way of serial interface, fingerprint sensor can communicate
with microcontroller runs on either 5 V or 3.3 V power supply. TX/RD pin connects
IoT Enabled Secured Card Less Ration Distribution System 481

with TXD (TX-OUT pin of the microcontroller). We have used optical fingerprint
scanners, capture high contrast images.
Details are captured with the high number of diodes per inch. They work by
reflecting light over their fingers and taking photos. The ridges of fingerprint help to
produce the digital image.
Figure 2 shows the verification of the user with the help of Fingerprint sensor.
Then this sensor verifies the user by looking into Aadhar Data stored.
We have used the Raspberry Pi 3 module which is supporting the operating system
like Linux. To make our system more user friendly GUI (Graphic User Interface) has
been developed. It can be made with Python programming. We can create GUI with
third party modules like tkinter. For that we have to import all supporting modules.
We can also import the Google text to speech for Linux named as eSpeak by which
we will read out all the instructions.
Source code is the heart of all the hardware which is interfaced to the micro-
controller board. In Python IDE we can write, compile and run the source code.
Google Firebase is used to create and store a real-time database of all the customers.
The data like name, address, Aadhar number, allotted ration, bank account number,
mobile number is stored in the tabular format like we store in excel sheet. Database
is managed and stored on the cloud services provided by Google cloud and Amazon.

Fig. 2 Fingerprint
verification
482 S. K. Rudrawar et al.

4 Methodology

This system proposes the advanced Ration Distribution system, named as “IoT
enabled secured card less Ration Distribution System”. The proposed system is
similar to the ATM and proposed to work for 24/7. It must have a biometric input
to get started. It incorporates an embedded and IOT based automatic ration shop.
Figure 3 shows the front and back end of the proposed system. Steps involved in
implementation of systems are
• Data Collection: We are collecting the data of the user from the Aadhar database
as well as by linking the ration cards of the user with the Aadhar card ID. It will
hold all the data such as Aadhar number, ration card number, bank account details
and allocated goods.
• Network Configuration: Now from the collected data in the system we will
configure the settings which are mainly required for allocating goods to the user
and also to verify the user.
• Storage: We require the storage area to store the grains like rice and wheat as well
as some commodities like sugar.
• Provision: Providing the allocated goods and the information after verifying the
authorized user.
• Warning: If someone tries to use your account then the system will give alert to
both administration as well as particular users. And also it conveys the message
of deposition of material into their account and withdrawal of money from their
bank account.

Fig. 3 Overview of the proposed system

IoT Enabled Secured Card Less Ration Distribution System 483

• Alert: The customer should be alerted whenever there is use of his identity with
the system. This is done by sending an OTP and the emails with the associated
account.

5 Working

At the beginning we need to scan fingerprints, to check if the person is registered

or not. If not then they need to register to the system, for that he must initialize
the process by selecting register. After scanning the fingerprint will get SMS from
the administrator conveying his account has been created. He needs to again scan
the fingerprint. Now he has an account hence the details are displayed after the
fingerprint matches with the database. If he fails then Warning is displayed on the
display stating unauthorized user. After that he must select the action to be taken such
as show account balance, withdraw goods. Depending upon its action system will
try to interact. If he chooses to display account balance then allotted ration balance
will be displayed, if he chooses to withdraw then the system will ask which good he
wants to withdraw and the quantity of the same among the options. System will send
an OTP to the customer stating a transaction has been started. Please provide the OTP
to begin the process. OTP is used because it will provide security and try to protect
the system from thefts. If OTP is wrong three times it will alert the administration
department and send a message to the corresponding customer regarding the same.
If OTP matches then the particular amount will be deduced from the customer bank
account and SMS will be sent to him from the bank. After that the system will begin
to dispense the goods by opening the valves and measuring the same. After some
interval of time it will try to match the dispensed material is the same or near to the
selected amount. If the same process will stop and it will log out from the system.
If it doesn’t match with the selected amount then it will provide the feedback to the
system and again open the valve until it matches the selected amount.

6 Prototype Testing

Figure 4 shows the hardware present inside the system, Where two conveyor belts
platforms are used to deliver the goods after successful verification, validation and
measurement. The Hoffer present above is used to deliver the grains by opening and
closing the valve into the container placed on the platform. It takes almost a couple of
minutes and then it measures the delivered grains with the help of a load cell placed
below the conveyor belt. Then to deliver the goods to the customers those belts are
run until it reaches the customers. And the next transaction restarts only after placing
the container. GUI makes the process smooth and user friendly is shown below in
the Fig. 5 which uses touch screen to interact with the user.
484 S. K. Rudrawar et al.

Fig. 4 Hardware of proposed system

Fig. 5 GUI of system

We have used the Raspberry Pi 3 module which is supporting the operating system
like Linux. To make our system more user friendly GUI can be more user friendly.
GUI is an abbreviation for Graphic User Interface. It can be made with Python
programming. We can create GUI with third party modules like tinker CAD. For that
we have to import all supporting modules. We can also import the Google text to
speech for Linux named as eSpeak which will read out all the instructions. Source
code is the heart of all the hardware which interfaced to the microcontroller board.
In Python IDE we can write, compile and run the source code.
Figure 6 shows customers database. Google firebase is used to create and store.
Google firebase is used to create and store real-time database of all the customers.
The data like name, address, Aadhar number, allotted ration, bank account number,
IoT Enabled Secured Card Less Ration Distribution System 485

Fig. 6 SMS alert given to

customer

mobile number is stored in the tabular format like we store in the Excel sheet. After
successful verification of biometrics, the user receives OTP on the registered mobile
number as shown in Fig. 7. The user has to enter the OTP. The user has three chances
to enter the 4 digit OTP if the password does not match an alarm is triggered. After
successful verification, the user has full access to the system where he can select
the ration material and the quantity of material required. Once transaction has been
completed the database will get updated with amount of grains allocated as per recent
transaction. As shown in Fig. 8. materials will get dispatched through the conveyor
belt inside the system. Hence, the experimental results show that the proposed system
is easy to access and prevent the ration from theft activity.

Fig. 7 Database management

486 S. K. Rudrawar et al.

Fig. 8 Front view of ration

distribution system

7 Conclusion and Future Scope

In a populated country like India it is very difficult to manage the records of the public
distribution system. The proposed system not only helped to manage the records but
also it made the distribution of goods among all in a very systematic and transparent
way. It minimized the chances of malpractices done by the shop owners in order to
gain the profit. Also reduced the human intervention which may result in incorrect
measurement and the reduced the chances of infection to others in the situation like
COVID-19.
Proposed system has achieved accuracy and transparency by verification of the
correct user and maintains the record of the goods. It also saved the time consuming
process of distribution of goods at ration shops. With the help of biometric scanning
it reduced the chances of malpractices and the adulteration as there is no human inter-
vention while distribution of goods. This can be a part of an E-commerce platform
IoT Enabled Secured Card Less Ration Distribution System 487

which promotes the cashless transaction. The proposed system will not only aid the
government agencies but will also help to digitize the system and in term help to
deploy resources efficiently to the citizens.
One limitation of proposed work is to get the solution for high volume of goods
storage.
In this year we have faced the rapid spread of COVID-19 which spread with
contacts of people. It’s our future work to make contactless with the help of face
scanning and with the voice commands which can be achieved with Image processing
as well as some signal processing techniques.

References

1. https://instrumentationtools.com
2. https://mahafood.gov.in/website/english/PDS.aspx
3. https://www.indiaenvironmentportal.org.in
4. https://www.tnpd.gov.in
5. Bagul G, Desouza B, Gaikwad T, Panghanti A (2017) Smart ration card automation system.
Int Res J Eng Technol (IRJET) 04(05)
6. Padmavathi R, Mohammed Azeezulla KM, Venkatesh P, Mahato KK, Nitin G (2017) Digi-
talized aadhar enabled ration distribution using smart card. In: International conference on
recent trends in electronics information and communication technology (RTEICT), vol 2(2)
May 19–20
7. Pallikonda Rajesekara M, Balaji D, Arthi R, Daniel P (2017) Automatic smart ration distribution
system for prevention of civil supplies hoarding in India. In: International Conference on
Advanced Computing and Communication Systems (ICACCS -2017), Coimbatore, India. 978-
1-5090-4559-4. 06–07 Jan 2017
8. Aishwarya M, Nayaka AK, Chandana BS, Divyashree N, Padmashree S (2017) Automatic
ration material dispensing system. In: International conference on trends in electronics and
informatics ICEI 2017. 978-1-5090-4257-9
9. Balasubramani A, Sunil K, Madhu KN (2018) Cashless automatic rationing system by using
GSM and RFID technology. In: Proceedings of the Second International conference on I-
SMAC. ISBN: 978-1-5386-1442-6.
10. Fitzgerald DW, Murphy FE, Wright WMD, Whelan PM, Popovici EM (2015) Design and
development of a smart weighing scale for beehive monitoring. In: 2015 26th irish signals and
systems conference (ISSC). Carlow, pp. 1–6. https://doi.org/10.1109/ISSC.2015.7163763
11. B. Buvansari, G. Ramya, R. Shivapriyaa (2015) GSM based home security system. Int J Eng
Tech Res (IJETR). 3(2), ISSN: 2321–0869
Voice Assisted Bots for Automobile
Applications

Shilpa K. Rudrawar, Nikhil Choudhar, and Ankit Meshram

Abstract The Controlled Infotainment system is based on a single board computer

Raspberry Pi 3 Model B+ . This system is inspired by popular products in the market
\ALEXA”, \MBUX-Mercedes Benz” and \Hyundai-TUCSON". As Infotainment
system is the combination of ‘Information and ‘Entertainment, this includes voice-
controlled multimedia such as online music player. In a hands-busy and eyes-busy
activity such as driving, spoken language technology is an important component of
the multimodal human–machine interface. Adding speech to the HMI introduces two
distinct challenges: (1) accurately acquiring the user’s speech in a noisy car envi-
ronment (2). Creating a spoken dialog system that does not require the driver’s full
attention In order to provide security spy camera is used to capture the image of a
person entering inside the car and accordingly the email will be sent to the owner of
the car.

Keywords Raspbian · Python · Pi cam · Semi-autonomous · XLDE

1 Introduction

In order to perform task by giving voice commands, A system is developed by using

Google voice and API on SBC RPI 3B.The voice command from user is converted to
text format by using Google voice API. The next is then compared with the predefined
commands available in the database of the program If it matches with any of them
then the bash command associate with it will be executed.This is achieve by using
Google speech [1]. The status of the results can be obtained in the form of audio by
using library E-speak, in order to listen users favorite songs online music player and
there stream music using a API. If user wants to listen from its own collection then
the system has option i.e. online music player and the select one of the song from

S. K. Rudrawar (B) · N. Choudhar · A. Meshram

Department of Electronics and Telecommunication Engineering, MIT Academy of Engineering,
Alandi (D), Pune, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 489
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_49
490 S. K. Rudrawar et al.

list. The camera is used as distance sensor to view obstacle. User will able to see a
view of the left-side and the right-side of your car using camera capture the image
during accident in order to identify who made mistakes.

2 Literature Survey

Any field of research or surveillance close examination or security of their concerned

object or subject. In order to overcome conundrum, we need an automatic approach.
An automatic approach best suited for these kind of application is alternatively or
proxy surveillance, a device that can observe a spot through its own eyes and get it
observed on the other end by master controller [2]. Literature presents different tech-
niques through which inspection can be performed through a designed bot. One such
technique proposed in this paper is a voice controlled bots which can be controlled
through voice and which can record surrounding. In this research a detailed survey
is conducted to identify the research challenges and achievements and progress in
this field [3].The major benefit of using voice system is that it can use it for handicap
people also. A camera serves the purpose of capturing images of the surrounding
or recording the visuals to display at the other end of the controller or master [4].
The Wi-Fi feature provides a wireless interface between the server and the bot. The
paper reviewed expressed different ways to control and asses the surrounding using
different techniques in which one of the simplest and costless was using an Arduino
also. The conventional, reliable way of a bots movement control through a controller
input is preferred here.

3 Block Diagram

As shown in Fig. 1, the Voice Controlled Infotainment System consists input devices
such as microphone, OBD II and output devices such as 7” HDMI TFT LCD display,
camera, speaker. The microprocessor works on car battery only through the charging
point for mobile.

3.1 Hardware Part

Microphone
It takes input from user and convert it into electrical signal.i.e. it acts as transducer.
Signal coming from Mic is analog in nature. The speech input from microphone is
given to the raspberry pi and there input speech is compared with command configure
Voice Assisted Bots for Automobile Applications 491

Fig. 1 Block diagram

line in raspberry pi. This process is done by Google. Only one command is effective
at a time.
Raspberry Pi 3
The Raspberry Pi 3 board model B + has a processor of 1.2 GHz 64-bit quad-core
ARMv8 CPU and 1 GB RAM which almost acts like a mini computer. Raspberry
pi 3 Board has 802.11n wireless LAN and Bluetooth 4.1 for this install Raspbian
Jessey in the memory card used for the board. Raspberry Pi 3 has a LINUX based
operating system call Raspbian. There are also 40 GPIO pins which can be used
as both digital input, digital output and to control and interface with various other
devices in the real world, four USB ports, one HDMI port, one Ethernet port, one
3.5 mm Audio jack, micro USB power supply. This board also has serial connections
for connecting a camera (CSI) and a display (DSI).The Camera Module is a great
accessory for the Raspberry Pi, allowing users to take still pictures and record video
in full HD. In our system we use Pi camera to capture image if human appearance
is detected. This captured image is sent by email to owner of the car. Before using
mic, first check whether microphone records properly. First, check if microphone is
listed using the command “lsusb”. Check if mic comes up on the list. Next, in order
to set the mic recording volume to high. To do this, enter the command \alsamixer”
in the terminal. On the graphical interface that shows up, press the up/down arrow
keys to set the volume. Press F6 (all) and then select the Mic from the list. Again,
use the up arrow key to set the recording volume to high.
Steps to connect Headphones with raspberry pi:
• Connect headphones to Rpi
• Type command in terminal raspi-con g
• Go to advance setting Audio
• Select headphone and click ok
492 S. K. Rudrawar et al.

On Board Diagnostic II Scan tool

Society of Automotive Engineers (SAE) and International Standardization Organi-
zation (ISO) issued a set of standards which described the interchange of digital
information between ECUs and a diagnostic scan tool [5]. All OBD-II compliant
vehicles were required to use a standard diagnostic connector (SAE J1962), and
communicate via one of the standard OBD-II communication protocols.
Pi Camera:The Camera Module is a great accessory for the Raspberry Pi, allowing
users to take still pictures and record video in full HD. In our system we use Pi camera
to capture image if human appearance is detected. This captured image is sent by
email to owner of the car.

3.2 Software Part

Raspbian OS
Raspbian OS is a Debian-based operating system built for Raspberry Pi boards. It is
the official operating system of the Raspberry Pi single board computers. It provides
a modified XLDE desktop environment. It includes thousands of packages. Raspbian
OS has been used in this project to provide a user-friendly desktop environment for
running the applications on the Raspberry pi. It provides a python command terminal
which is used to execute any python statement. The python code for the sensor,
camera and motors were done in the python IDLE window by default available in
the Raspbian OS package [6].
Python command shell
It is the window where we write the python commands for running and testing the
sensor or camera or the working of the motors. We initialize the.py files here and run
them to get the desired output. The image captured by the pi camera is processed
and worked upon by executing the python commands for the same. To connect the
server to the raspberry pi, we have to enter the command and URL of the domain to
stream the camera [7].
Linux Commands:
• Man: It is used to show the manual of the inputted command. Just like a lm on
the nature of lm, the man command is the meta command of the Linux CLI.
• ls: It show all of the major directories led under a given le system show all the
folders stored in the overall applications folder. The ls command is used for
viewing folders and directories.
• cd: It will allow the user to change between le directories.
• mv: It allows a user to move a le to another folder or directory.
Voice Assisted Bots for Automobile Applications 493

4 Methodology

In order to perform task by giving voice commands, a system is developed by using

Google Voice and Speech APIs on SBC-Rpi 3 B [8]. The voice command from user
is captured by the microphone. This is then converted to text format by using Google
Voice API. The text is then compared with the pre designed commands available in
the database of the program. If the voice command matches with any of them, then
the bash command associated with it will be executed. This is achieved by using
the Google speech API, which converts the text into speech [9]. The status of the
results can be obtained in form of audio by using library E-speak. In order to listen
user’s favorite songs online our system has option i.e. select online music player and
them Stream music using music API. If user wants to listen from its own collection
then the system has option i.e. online music player and the select one of the song
from list. The camera is used to capture image during absenteeism of owner if any
unauthorized person come. Camera is used as distance sensors to view obstacle. User
will able to see a view of the left-side and/or the right-side of your car using camera
Capture the image during accident in order to identify who made mistake. ELM 327
USB interface used to obtain diagnostic information from vehicle that have OBD
[10].

5 Results and Discussion

5.1 Testing and Troubleshooting

Following are some problems during implementation of project.

Problem1: Interlinking of pages using tkinter.

Solution: Use PyQt module for GUI. Qt is more than a GUI toolkit. It includes
abstractions of network sockets, threads, Unicode, regular expressions, SQL
databases,SVG, OpenGL, XML, a fully functional web browser, a help system, a
multimedia framework, as well as a rich collection of GUI widgets.

Problem2: Error in handing device id:

Solution: To find serial number, type cat = proc = cpuinf o at the command line.

Problem3: How to find values associated with Flac:

Solution: Check the audio folder in this repository for example:
Channels: 2
Sample Rate: 44,100.
Precision: 32-bit.
Sample Encoding: 32-bit Float.
494 S. K. Rudrawar et al.

Fig. 2 Pi camera output

5.2 Camera Output

The Pi camera placed on the bot, takes image recognized by bot is shown in Fig. 2.

5.3 Bot Movement

The bot moves according to voice command received, bot hardware is shown Fig. 3.
The voice commands given to the system were actuated, actions like move left,
right, forward, backward and also the infotainment system, just by asking “Jarvis
Play music” or Jarvis tell me something about … The system greets the user and
asks how could I help! Thus the system is human interactive.
The concept is programmed by playing the wav files through a speaker if any
voice command is received. The wav files from the SD card module and given as
input to the inverting terminal of the amplifier of LM358 whenever a constant supply
from voltage driver is given to non-inverting terminal. Table 1 is showing different
recorded sounds of voice recording of the robot for different voice commands.
Here an application has been built through which we could monitor and access
the motions of the bots. We need to press the voice button icon, just as we do for OK
GOOFLE!. and once the Google-API is on we need to give commands such as turn
forward, right etc.
Voice Assisted Bots for Automobile Applications 495

Fig. 3 Voice operated bot

Table 1 Input for different logic

Voice command Input 1 Input 2 Input 3 Input 4 Direction
Stop 0 0 0 0 Stop
Forward 1 0 1 0 Forward
Backward 0 1 0 1 Backward
Left 0 1 0 0 Left
Right 0 0 0 1 Right

Figures 4 and 5 shows that once voice command like forward is given to bot,
the Google API application displays that command on screen, once it’s clear that

Fig. 4 Voice based

Google-API
496 S. K. Rudrawar et al.

Fig. 5 Command display by

Google API

the same command which was made is displayed on screen we can proceed ahead
(Fig. 6).
The desired outputs were built successfully using Python and its libraries. We
imported various modules such as speech recognition, pyttx—Python text to speech
and many other required for converting speech to text and vice versa. We used to give
commands which can be seen in the above image which states that it first initializes
Jarvis and then it greets the user as per the time and then listens to users command
and recognize, if the audio command given by user is recognized then the following
action is executed, you can see I said open Youtube in the above program and it
executes the same. In this way the smart assistant works over your voice.

Fig. 6 Software output

Voice Assisted Bots for Automobile Applications 497

6 Conclusion

The proposed system is able to provide information and entertainment to persons

sitting inside car. The commands given through voice are able to perform operations
such as music player, information etc. To provide security, Pi-camera is provided
which is able to capture image of unauthorized persons and send email to owner of car.
This system can be used in a wheel chair for handicap people. We can give command
through voice and it will automatically run according to the given commands, making
person’s life easy and safe.

References

1. Gaoar, Kouchak SM (2017) \Minimalist design, 2017, “An optimized solution for intelligent
interactive infotainment systems”. Intell Syst Conf 553–557
2. Mi-JinKim, Wook Jang J (2010) A study on in-vehicle diagnosis system using OBD-II with
nevigation. Int J Comput Sci Netw Secur 10(9):136–140
3. Isabella, Retna E (2012) \Study paper on test case generation for GUI based testing. Int J Softw
Eng Appl 3(1):139–147
4. Jarvis R, Ho N, Byrne J (2007) Autonomous robot navigation in cyber and real worlds. In:
2007 international conference on cyber worlds (CW’07). Hannover, 2007, pp 66–73. https://
doi.org/10.1109/CW.2007.31
5. Obispo SL (2017) M3 Pi: Raspberry Pi OBD-II touch screen car computer, 1–30
6. Sonnenberg J (2012) Service and user interface transfer from nomadic devices to car info-
tainment systems. In: International conference on automotive user inter-face and interactive
vehicular applications, pp 162–165
7. Prabha SS, Antony AJP, Meena MJ, Pandian SR (2014) Smart cloud robot using raspberry
Pi. In: International conference on recent trends in information technology, Chennai, pp 1–5.
https://doi.org/10.1109/ICRTIT.2014.6996193
8. Menciassi JH, Park S, Lee S, Dario GP, Jong-Oh P (2002) Robotic solutions and mechanisms
for a semi-autonomous endoscope. In: IEEE/RSJ international conference on intelligent robots
and systems, vol 2. Lausanne, Switzerland, pp 1379–1384. https://doi.org/10.1109/IRDS.2002.
1043947
9. Denker, İşeri MC (2017) Design and implementation of a semi-autonomous mobile search
and rescue robot: SALVOR. In: 2017 international artificial intelligence and data processing
symposium (IDAP), Malatya, pp 1–6. https://doi.org/10.1109/IDAP.2017.8090184
10. Bensalem S, Gallien M, Ingrand F, Kahloul I, Thanh-Hung N (2009) Designing autonomous
robots. IEEE Robot Autom Mag 16(1):67–77. https://doi.org/10.1109/MRA.2008.931631
Content-Based Image Retrieval Using
Color Histogram and Bit Pattern
Features
Nandkumar S. Admile, Akshay A. Jadhav, Swagat M. Karve,
and Anil A. Kasture

Abstract To Identify the Particular image from a huge dataset is a key problem in
Image Processing. For image retrieval, a block truncation coding based method is
used. The image features derived from colour quantizer as well as a Bitmap image.
Using colour quantizer Color Histogram Features (CHF) is obtained, and using
Bitmap image Bit Pattern Feature (BPF) is obtained. The various distance measures
are employed to match the similarity between images. Simulated result shows better
in term of Average Recall Rate (ARR) and Average Precision rate (APR).

Keywords Average recall rate · Average precision rate · Color histogram

features · Bit pattern feature · Dot diffused block truncation coding

1 Introduction

In recent years image retrieval is a key innovative topic. In earlier days annotated,
the based method is used for image retrieval, but in this method, we have to manually
provide the description of each image. Hence in order to solve this issue instead of
this method content-based image retrieval invented. It is the method to identify the
same image available in the dataset—a number of techniques employed for image
retrieval. In 1979 Delp and Mitchell developed the Block Truncation Coding (BTC)
which is normally used for image compression [1]. Block truncation coding method
split the image into noncontiguous image block. From each block, high and low mean
values are calculated by performing Thresholding low mean values bitmap image is
derived. In [5–7] author obtains better accuracy using block truncation coding based
image retrieval. In this paper, the author uses RGB colour space to extract several
of image features. In [8] instead of RGB colour space, the author uses a different
colour space to extract several features of the image. The author obtains the improved
results. In [3, 4] and [9] BPH and BCCM derived to obtain the sameness of contents.
In [10], the author uses greyscale colour, which proves it achieves better image quality

N. S. Admile (B) · A. A. Jadhav · S. M. Karve · A. A. Kasture

SVERIs College of Engineering, Pandharpur, Maharashtra, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 499
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_50
500 N. S. Admile et al.

and efficiency as compared to ODBTC [11] and EDBTC [12]. This paper describes
Dot Diffusion Block Truncation Coding (DDBTC) [13, 14] for retrieval of images.
Colour Histogram Features and Bit Pattern Features are extracted.

2 SYSTEM DESIGN: Dot Diffused Block Truncation

Coding

In the DDBTC technique (Fig. 1), a colour image is divided into a number of non-
overlapping image blocks. The DDBTC encoder generates two quantizers (minimum
and maximum) from the RGB colour space (12). Consider f (i, j) be the image block
at position (i, j). Where i = 1, 2 … M/m and j = 1, 2 … N/n. Let fR (x, y), fG (x, y)
and fB (x, y) are the RGB pixel values in the image block (i, j) where x = 1, 2 … m
and y = 1, 2 … n.

Qmin = {min fR (x, y), min fG (x, y), min fB (x, y)} (1)

Qmax = {maxfR (x, y), maxfG (x, y) and max fB (x, y)} (2)

In the Eqs. (1) and (2) Red, Green, Blue colour is represented by R And B.
The gray scale image is obtained using Eq. (3) as:

f∧ (x, y) = 1/3[fR(x, y) + fG(x, y) + fB(x, y)] (3)

From the greyscale image, the Bitmap image is obtained as:

bm = 1; for f∧ (x, y) ≥ f(i, j).

0; for f∧ (x, y) < f(i, j). (4)

3 Feature Descriptors of Image

The feature descriptor of the image is extracted using DDBTC colour quantizers as
well as Bitmap Image. Using minimum and maximum quantizers, Color Histogram
Feature is removed and using Bitmap Image Bit Pattern Features is extracted.
1. Color Histogram Feature (CHF)

Using minimum and maximum quantizers shown in Eqs. (5) and (6) Color Histogram
Feature is extracted.
Content-Based Image Retrieval Using Color Histogram and Bit … 501

Fig. 1 Block diagram [13]

502 N. S. Admile et al.

Let C = {C1, C2 …. CNc} is colour codebook. Minimum and Maximum quantize

of DDBTC represented i and j.

imin (i, j) = arg min qmin (i, j), cmin

k 2
2
(5)
k=1,2..Nc

imax (i, j) = arg max qmax (i, j), cmax

k 2
2
(6)
k=1,2..Nc

i = 1, 2 … M/m; j = 1, 2 …. N/n.
CHFmin and CHFmax are calculated using Eqs. (7) and (8) respectively:
 
M N
CHFmin (k) = Pr imin (i, j) = k/i = 1, 2.... ; j = 1, 2.... (7)
m n

where k = 1, 2, Nmin.
 
M N
CHFmax (k) = Pr imax (i, j) = k/i = 1, 2.... ; j = 1, 2.... (8)
m n

k = 1, 2, Nmax.

4 Bitmap Pattern Feature (BPF)

The Bitmap Pattern Feature is derived from the Bitmap Image.

A. Image Retrieval

For performing image retrieval first, dataset is created, which is called as training
image dataset. Forth similarity measurement CHF and BPF features are used. The
L1 distance represented in Eq. (9), L2 distance represented in Eq. (9), and Modified
Canberra distance are used to calculate the distance between images.
The Various distance metrics are formulated as:
1. L1 distance
N
min
query target
δ(query, target) = α1 ./CHFmin − CHFmin (k)/
k=1
N
max
query target
+ α2 ./CHFmin (k) ∗ CHFmin (k)/
k=1


Nb
target
+ α3 ./BPF.query (k) − BPFmin (k)/ (9)
k=1
Content-Based Image Retrieval Using Color Histogram and Bit … 503

2. L2 distance
N
min
query target
δ(query, target) = α1 ./CHFmin − CHFmin (k)/ ∧ 2
k=1
N
max
query target
+ α2 ./CHFmin (k) ∗ CHFmin (k)/ ∧ 2
k=1


Nb
target
+ α3 ./BPF.query (k) − BPFmin (k)/1 /2 (10)
k=1

3. Modified Canberra Distance

N
min query target
/CHFmin (k) − CHFmin (k)/
δ(query, target) = α1 . query target
k=1
/CHFmin (k) + CHFmin (k)/ + ε

Nmax query
/CHFmin (k) − CHFmin (k)/
target
+ α2 . query target
k=1
/CHFmax (k) + CHFmax (k)/ +ε

Nb
/BPF.query (k) − BPFmin (k)/
target
+ α3 . target (11)
k=1
/BPF.query (k) + BPFmin (k)/ +ε

Performance Measures:
The APR and ARR are measured by the following Eqs. (12) and (13)

1 
Nt
APR = Nq(L) (12)
Nt L q=1

1 
Nt
ARR = Nq(Nr) (13)
Nt Nr q=1

A number of retrieved images is represented by L, a total number of images in

the database are represented by Nt, and a number of similar images are represented
by Nr. The input image and correctly retrieved images are represented by q and Nq,
respectively.
504 N. S. Admile et al.

Fig. 2 Images from corel dataset [13]

5 Results and Discussion

5.1 Experimental Setup

The Corel Dataset is employed [2] to perform the image retrieval task. A sample
image is given in “Fig. 2”.Image dataset contains only natural images.
For Experimental result, one image is randomly selected mentioned in “Fig. 3”.
Minimum and Maximum quantized image obtained by performing the quantization
process mentioned in “Fig. 4”. From Grey colour space Bitmap image is derived
mentioned in “Fig. 5”.
By performing the experiment distance between the query image and target image
is measured shown in “Fig. 6”.

Fig. 3 Query image [13]

Content-Based Image Retrieval Using Color Histogram and Bit … 505

Fig. 4 Quantized image (Minimum and Maximum) [13]

Fig. 5 Bitmap image [13]

Fig. 6 Retrieved images [13]

Calculated results show modified Canberra distance achieves better accuracy as

compared to distance L1 and L2. The results in terms of APR and ARR for distance
L1, L2 and changed Canberra distance, as shown in Tables 1 and 2.
The Fig. 7 shows a graphical representation of image retrieval. In this case,
Average Precision Rate (APR) 0.9375 and Average Recall Rate (ARR) 95.19% is
achieved.
506 N. S. Admile et al.

Table 1 Calculations of L1 and L2 distance [13]

Class L1 distance L1 distance L2 distance L2 distance
APR ARR APR ARR
African 0.6875 60.19 0.6875 60.19
Beach 0.6875 65.19 0.6875 65.19
Building 0.6875 70.19 0.6875 70.19
Bus 0.8125 80.19 0.8125 80.19
Dinosaur 0.8750 90.19 0.8750 90.19
Elephant 0.5625 60.19 0.5625 60.19
Flower 0.8125 70.19 0.8125 70.19
Food 0.8750 65.19 0.8750 65.19
Horse 0.6875 65.19 0.6875 60.19
Mountain 0.9575 95.19 0.9375 95.19

Table 2 Calculations of
Class Modified Can. Dist Modified Can. Dist.
modified of Canberra
Distance [13] APR ARR
African 0.8750 90.19
Beach 0.8750 90.19
Building 0.8750 90.19
Bus 0.9375 95.19
Dinosaur 0.9375 95.19
Elephant 0.9375 90.19
Flower 0.9375 95.19
Food 0.9375 95.19
Horse 0.8750 90.19
Mountain 0.9375 95.19

Fig. 7 Graphical representation of APR and ARR [13]

Content-Based Image Retrieval Using Color Histogram and Bit … 507

Fig. 8 Graphical representation (Precision vs. Recall) [13]

The “Fig. 8” shows the graphical representation of Image Retrieval.

6 Conclusion

Dot diffusion block truncation coding used for performing the image retrieval. The
image brightness and colour distribution represent by colour histogram feature (CHF)
accurately. The image contents are characterized by bit pattern feature (BPF). The
DDBTC provide the highest average precision rate and average retrieval rate. In
future work, the same technique can also be applied for video retrieval. In which
an image sequence DDBTC technique can be applied directly. Use another colour
space can also be possible by this technique.

References

1. Delp EJ, Mitchell OR (1979) Image coding using block truncation coding. IEEE Trans Comm
COM-27 9:1335–1342
2. Corel Photo Collection Color Image Database. [Online]. Available: https://wang.ist.psu.edu/
docs/realted/. Accessed 2001
508 N. S. Admile et al.

3. Qiu G (2003) Color image indexing using BTC, (2003). IEEE Trans Image Process 12(1):93–
101
4. Gahroudi MR, Sarshar MR (2007) Image retrieval based on texture and colour method in
BTC-VQ compressed domain. In: Proceeding of the international symposium signal process
Its applications, pp 1–4
5. Silakari S, Motwani M, Maheshwari M (2009) Color image clustering using block truncation
algorithm. Int J Comput Sci Issues 4(2):31–35
6. Lai CC, Chen Y-C (2011) A user-oriented image retrieval system based on interactive genetic
algorithm. IEEE Trans Instrum Meas 60(10):3318–3325
7. Yu FX, Luo H, Lu ZM (2011) Colour image retrieval using pattern co-occurrence matrices
based on BTC and VQ. Electron Lett 47(2)
8. Yu F-X, Luo H, Lu Z-M (2011) Colour image retrieval using pattern co-occurrence matrices
based on BTC and VQ. Electron Lett 47(2):100–101
9. Xingyuan W, Zongyu W (2013) A novel method for image retrieval based on structure element’s
descriptor. J Vis Commun Image Represent 24(1):63–74
10. Guo JM, Liu YF (2014) Improved block truncation coding using optimized dot diffusion. IEEE
Trans Image Proc 23(3):1269–1275
11. Guo JM, Prasetyo H (2014) Content-based image retrieval using feature extracted from
halftoning based BTC. IEEE Trans 24(3):1010-1024
12. Guo JM, Prasetyo H (2014) Effective image retrieval using feature extracted using dot diffused
BTC. IEEE Trans 24(3):1010–1024
13. Admile NS (2018) Image retrieval based on block truncation coding. In: IEEE international
conference on communication and electronics systems (ICCES-2018). Coimbatore, India
14. Admile NS, Kanase A, Khiste R (2019) User-oriented image retrieval using feature extracted
by dot diffusion block truncation coding. In: IEEE international conference for convergence in
technology (I2CT-2019). Pune, India, pp 29–31
KNN and Linear SVM Based Object
Classification Using Global Feature
of Image

Madhura M. Bhosale, Tanuja S. Dhope, and Akshay P. Velapure

Abstract Machine learning plays a vital role in Object classification due to its
various applications viz autonomous vehicle, driverless cars. In our research work we
have considered machine learning algorithm, linear support vector machine (SVM)
and K-Nearest Neighborhood (KNN) for classification of object like car and truck
which are essential for Autonomous vehicle applications. We have performed RGB
to gray conversion followed by histogram of gradient (HOG) for feature extraction
before applying to KNN and SVM for classification. The dataset required for the
experimentations for training and testing are utilized from kaggle website and the
performance of SVM and KNN have been evaluated on these testing data. Results
show that SVM outperforms the KNN providing accuracy of 71.3%.

Keywords HOG feature · Linear SVM · KNN · Machine learning · Classifier

1 Introduction

Machine learning is very important topic now days. According to global survey
growth in machine learning is 48.3%. We can implement so many applications with
machine learning algorithms such as in medical field, robotics, and weather fore-
casting. Driverless car is one of the famous applications of machine learning. With the
help of machine learning algorithms we can design system with minimum program.
Computer program and its development is prime work of machine learning algorithm
and that can access data and use it learns from them. Feature selection is very impor-
tant task in image processing and computer vision applications [1, 2]. Processing
on huge amount of data makes your algorithm inefficient so feature selection is
important [3–5]. Features are nothing but numerical values extracted from images
[6]. Feature or descriptors selection algorithms [7–9] broadly categorize into local

M. M. Bhosale (B) · T. S. Dhope

Department of Electronics and Telecommunication Engineering, JSPM’s Rajarshi Shahu College
of Engineering, Pune 411033, India
A. P. Velapure
Independent Researcher, Pune, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 509
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_51
510 M. M. Bhosales et al.

feature and global. The entire image describe by Global feature and region of interest
is describe by local features like speeded up robust features (SURF) [10], scale-
invariant feature transform (SIFT) [11], Local binary patterns (LBP) [12]. Global
descriptors are generally used in image classification, object detection. Texture in
an image patch is represented by local descriptor. Invariant moment, histogram of
oriented gradient (HOG), Co-HOG, shape matrices, hue, haralic texture are some
examples of global descriptors. Nowadays, deep learning algorithms used for high
level semantic features. Convolution neural network(CNN) based approaches are
also useful for feature selection of large number of dataset [13] So overall this area
of feature improvement with traditional algorithm is still remains an open research
area [14, 15]. This work is preliminary step of real time object detection. This work
mainly focused on comparison of different machine learning algorithms, based on
performance of the trained model selection of best classifier is possible. Here we
have used HOG feature that is global feature of images. We used SVM and KNN
algorithms and compared the performance of these classifiers and then we have
recognized that SVM works better for this particular set of parameters application.
We have applied this on database in future we can apply it on real time with more
features and with some different algorithms. The paper is organized in 5 sections.
Section 2 elaborates Literature Review. Methodology discussed in Sect. 3, in which
we have explained block diagram of the system then about HOG feature and different
machine learning algorithms like KNN and SVM. Results have been discussed in
Sect. 4 wherein we have compared SVM and KNN with different parameters which
is followed by conclusions in Sect. 5 and future scope in Sect. 6.

2 Literature Review

There is a huge research has been going on object detection as specified in [16–
18]. Solved the problem of grouping and extraction of image features from complex
images using hierarchical based approach like voting and clustering studied in [19].
Fused local codebook less model feature and local binary pattern features for high
resolution scene classification elaborated in [20]. In [21], authors work on two prob-
lems; the methods defining directly texture in color space but more emphasis on
color. Combining different features with their good and bad parameters may worsen
the performance of system so they proposed new algorithm which will combined
features by using Color Intensity Based Local Difference Pattern (CILDP) and bag
of visual words (BOW) and the average precision improved by 12.53% and 6.68%.
Work focused on object detection on [22] tracking system based on SIFT and SURF
and their comparative analysis. SURF gives more accuracy as compared to SIFT. In
[23] paper Combined global feature with local feature for texture classification, with
the combination global matching and Local Binary Pattern Variance (LBPV) giving
drastic improvement in the more than 10% accuracy in sometimes.
KNN and Linear SVM Based Object Classification Using Global … 511

3 Methodology

The proposed methodology, overall algorithm has been splinted into two parts one
is training and other is testing.

(1) Training Model: Overall training model has been splinted into 3 steps. As
shown in Fig. 1a. First step is preprocessing then feature extraction then train
data using different machine learning algorithms.

(a) Gray scale conversion is utilized for preprocessing: Let v(x) = input
image as and converter output image = z(x) as specified in [24].

z(x) = log v(x) (1)

v(x) Is uniformly increasing step wedges and z(x) will convert that
uniformly increasing step wedges into exponentially increasing wedges.
(b) Feature Extraction using HOG: This gives information about appearance
about local object and shape, using edge direction distribution or inten-
sity gradient distribution [25]. Image is segmented into small connected
regions and then calculates histogram of those regions. Overall final
descriptor is addition of all these histogram [26]. For enhancing the

Fig. 1 a Block diagram of training phase of system, b Block diagram testing phase of system
512 M. M. Bhosales et al.

accuracy of histogram we have calculated contrast histogram, by calcu-

lating histogram of large region. HOG does not change with geometric
and photometric transformation. Following are the steps of histogram of
gradient:
(i) Gradient Computation: in this 1-D in centered derivative mask
is applied in both vertical and horizontal directions. Filtering of
color and intensity data can be possible using following kernel
mask.

[−1, 0, 1] and [−1, 0, 1]T (2)

(ii) Orientation Binning is useful for calculation of cell histograms. In

a cell each pixel has been weighted value by either rectangular or
radial in shape and histogram of channels are evenly distributed
over 0°–180° or 0°–360° based on is signed or unsigned gradient.
(iii) Descriptor Block is essential to knowing the changes in contrast,
illumination and grouping the cell together into larger spatially
connected blocks. Rectangular HOG (R- HOG) and Circular
HOG(C-HOG) are two geometries exist in descriptor block. R-
HOG block contains square grid. Number of cells per block, the
number of pixels per cell and the number of channels per cell
histogram are three parameter presents in R- HOG. And C- HOG
have single cell and central cell those with vary with central cell.
(iv) Block normalization: Assume y be non-normalization vector
containing histograms ||Y ||k. k Norms, where k = 1, 2 …. n
and f = small onstant. Normalization factor is shown in Eqs. (1
and 2):
y
x1 = √ (3)
|y2| ∧ 2 + f ∧ 2
y
x2 = √ (4)
||y1|| + f

(v) Object Recognition is done with linear SVM and KNN.

(vi) Training model formation using machine learning algorithm: in
our algorithm for training purpose we can use support vector
machine and KNN algorithms.
A. Linear Support Vector Machine (SVM): In linear classifier, n dimensional
points are separated with (n-1) dimensional hyperplane [14, 21]. The best hyper-
plane is one which separate two classes with maximum marginal value. Suppose
−→ −→ yn). Here y(j) is 1 or −1 repre-
we have n number of points (q1, y1) … (qn,
senting the class of point q(j). Now we are interested in exploring maximum
margin hyperplane separating class of point q(j) having y(j) = 1 from class
KNN and Linear SVM Based Object Classification Using Global … 513

of y(j) = −1. Distance between hyperplanes and nearest q(j) point should be
maximum. Equation of the hyperplane is,

s.
q −d = 0 (5)

−−→
where s = normal vector to the hyperplane, d/||q|| = offset of hyperplanes from
origin along s. Figure 2 shows graph of SVM.
B. K-Nearest Neighborhood (KNN): it is useful for both classification and
regression problem. Figure 3. shows flowchart of KNN.
(2) Second part of algorithm is testing. In this we do same operation on testing
data which we already performed on training data. And provide that data as an
input to trained model and predict the output. As shown in Fig. 1b.

Fig. 2 Graph of SVM – d= 1

– d= 0

– d= -1

Fig. 3 Flowchart of KNN algorithm

514 M. M. Bhosales et al.

Fig. 4 Flowchart of K-Fold

cross validation algorithm

C. K- Fold Cross Validation: it is stastical procedure to estimating skill of

machine learning algorithm due to simplicity and optimistic approach. The
procedure has single parameter k. If we chose k = 10 then it term tenfold cross
validation Fig. 4 shows flowchart of K-fold cross validation algorithm.

4 Results

For experimentation work we have used open source database of kaggle [27] and
python 3.6 for simulation. Database having two classes, truck and other is car. First
we have divided dataset into two parts, 70% for training and 30% for testing using k-
fold algorithm then resized trained images into 128*128 images after that converted
into gray scale then extracted total 16,384 features of 782 images using HOG as
shown in sample image of truck and car in Figs. 5 and 6 respectively.

Fig. 5 Image preprocessing and feature extraction of sample truck image

KNN and Linear SVM Based Object Classification Using Global … 515

Fig. 6 Image preprocessing and feature extraction of sample car image

Fig. 7 KNN training process output

After feature extraction we have trained our model using KNN algorithm as shown
in Fig. 7 with K (nearest data point) = 5, k- fold algorithm for cross validation with
k (cross validation parameter) = 10. This provides model training accuracy 65.45%
and model testing mean accuracy 59.84% and standard accuracy 58.60%.
Further we have trained our model with SVM as shown in Fig. 8. For this we have
selected linear kernel and with random state parameter is 9 with k-fold algorithm for
cross validation with k = 10. This gives model training accuracy 63.63% and model
testing mean accuracy as 71.31% and standard accuracy is 74.03.comparative study
between SVM and KNN as shown in Table.1.

5 Conclusion

In this paper, we have discussed various object detection papers in the literature
review. We presented comparative study between KNN and SVM algorithm for
object classification for two class truck and car. From our work we have observed
that for SVM with linear kernel and random state = 9, k (cross validation = 10) our
model accuracy is 71.31% and for KNN with K (nearest data point) = 5, k (cross
validation) = 10 our model accuracy is 59.84%. So SVM outperforms KNN.
516 M. M. Bhosales et al.

Fig. 8 SVM training process output

Table. 1 Comparative study between SVM and KNN

Classifier Parameters Model accuracyusing
K-fold
KNN K(nearest data point) = 5, k = (cross validation = 10) Mean Accuracy =
59.84%
Standard Accuracy =
58.60%
SVM Linear kernel, random state = 9, k(cross validation) = 10 Mean Accuracy =
71.31%
Standard Accuracy =
74.03%

6 Future Scope

In future scope focus need to be given on local features or combination of local and
global descriptors for better results for same database which are considered in this
paper and for improvement in results you can improve database count and you can
modify SVM and KNN model parameters such as kernel, C, gamma, random state
etc. Also you can compare with different other classifiers such as logistic regression,
random forest, decision tree etc. you can also add more classes for same applications
such as motorcycles, humans, traffic signals etc. Based on our current results and
future experimentations, results will be helpful in the area of driverless.
KNN and Linear SVM Based Object Classification Using Global … 517

Acknowledgements We would like to show our gratitude to JSPM’s Rajarshi Shahu College of
Engineering, for sharing their pearls of wisdom with us during the course of the proposed work.

References

1. Chen J, Lei B, Song Q, Ying H, Chen DZ, Wu J (2020 Proceedings of the IEEE/CVF. In:
Conference on computer vision and pattern recognition (CVPR), pp 392–401
2. Chen S, Tan X, Wang B, Lu Hu X, Fu Y (2020) Reverse attention based residual network for
salient object detection. IEEE TIP 29:3763–3776
3. Fan DP, Lin Z, Ji Z, Fu H, Cheng M (2020) Taking a deeper look at co-salient object detection.
In: CVPR, pp 2919–2929
4. Banerji S, Verma A, Liu C (2012) Cross disciplinary biometric systems. LBP and color
descriptors for image classification. Springer, Berlin, pp 205–225
5. Yu J, Qin Z, Wan T, Zhang X (2013) Feature integration analysis of bag of features model for
image retrieval. Neurocomputing 120:355–364
6. Miao J, Niu L (2016) A survey on feature selection. Procedia Comput Sci 91:919–926
7. Lowe DG (2004) Distinctive image features from scale-invariant keypoints. Int J Comput Vision
60(2):91–110
8. Yongsheng D (2017) Multi-scale counting and difference representation for texture classifica-
tion. Vis Comput. https://doi.org/10.1007/s00371-017-1415-4.
9. Ledoux A, Losson O, Macaire L (2016) Color local binary patterns:compact descriptors for
texture classification. J Electron Imaging 25(6):061404
10. Bay H (2008) Speeded-up robust features (SURF). Comput Vis Image Underst 110(3):346–359
11. Ojala T, Pietikäinen M, Harwood D (1996) A comparative study of texturemeasures with
classification based on featured distributions. Pattern Recognit 29(1):51–59
12. Shuang B, Zhaohong L, Jianjun H (2017) Learning two-pathway convolutional neural networks
for categorizing scene images. Multimed Tools Appl 76(15):16145–16162
13. Felzenszwalb P, Girshick R, McAllester D, Ramanan D (2010) Object detection with
disciminatively trained part-based models. IEEE Trans Pattern Anal Mach Intell 32:1627–1645
14. Lee H, Grosse R, Ranganath R, Ng A (2009) Convolutional deep belief networks for scalable
unsupervised learning of hierarchical representations. In: Proceedings annual international
conference on machine learning, pp 609–616
15. Mohan A, Papageorgiou C, Poggio T (2001) Example-based object detection in images by
components. PAMI 23(4):349–361
16. Papageorgiou C, Poggio T (2000) A trainable system for object detection. IJCV 38(1):15–33
17. Viola P, Jones MJ, Snow D (2003) Detecting pedestrians using patterns of motion and
appearance. In: The 9th ICCV, Nice, France, vol 1, pp 734–741
18. Schneiderman H, Kanade T (2004) Object detection using the statistics of parts. IJCV
56(3):151–177.
19. Gavrila DM (1999) The visual analysis of human movement: A survey. CVIU 73(1):82–98
20. Foresti GL, Regazzoni C (2000) A hierarchical approach to feature extraction and grouping.
IEEE Trans Image Proc 9(6):1056–1074
21. Bian X, Chen C, Tian L, Du Q (2017) Fusing local and global features for high-resolution
scene classification. IEEE J Sel Top Appl Earth Observations Remote Sens 10(6):2889–2901
22. Li L, Feng L, Wu J, Sun M, Liu S (2018) Exploiting global and local features for image retrieval.
J Central South Univ 25(2):259–276
23. Sakai Y, Oda T, Ikeda M, Barolli L (2015) An object tracking system based on SIFT and SURF
feature extraction methods. In: 18th International conference on network-based information
systems
518 M. M. Bhosales et al.

24. Guo Z, Zhang L, Zhang D (2010) Rotation invariant texture classification using LBP variance
(LBPV) with global matching. Pattern Recognition 43(3):706–719
25. Gavrila DM, Giebel J, Munder S (2004) Vision-based pedestrian detection: the protector+
system. Proceeding of the IEEE Intelligent Vehicles Symposium, Parma, Italy
26. Dalal N, Triggs B (2005) Histograms of oriented gradients for human detection. In: Proceedings
of the 2005 IEEE computer society conference on computer vision and pattern recognition
(CVPR’05), vol 05, pp. 1063–6919
27. https://www.kaggle.com/enesumcu/car-and-truck
MURA: Bone Fracture Segmentation
Using a U-net Deep Learning in X-ray
Images

Komal Ghoti, Ujjwal Baid, and Sanjay Talbar

Abstract Developing a robust bone fracture segmentation technique using deep

learning is an important step in the medical imaging system. Bone fracture segmenta-
tion is the technique to separate out the various fracture and Non-fracture tissues. The
fracture can occur in upper extremity parts of the human body like elbow, shoulder,
finger, wrist, hand, humerus and forearm etc. X-ray is one of the widely used imaging
modality for visualizing and assessing bone anatomy of the upper extremity. X-ray is
used in the diagnosis and planning of the treatment for the bone fracture. The problem
of computational bone fracture segmentation has gained researchers attention over a
decade because of high variation in fracture size, shape, location, variation in inten-
sities and variation textures. Many semi-automatic and fully automatic methods have
been proposed and they are becoming more and more mature. A recent technique that
is CNN based deep learning gives the promising result of the segmentation. In this
Method, MURA (Musculoskeletal Radiographs) database is used. The CNN based
U-Net model is trained using the MURA Database. After the training, the Model
is tested on the test images. The Evaluation parameters Like Dice Coefficient and
Validation Dice coefficient are found out to check the robustness of the technique.
The CNN based U-Net architecture gives the training dice coefficient of 95.95% and
validation dice coefficient of 90.29% for whole bone fracture segmentation.

Keywords U-net · Bone fracture segmentation · Deep learning · X-ray images ·

MURA dataset

1 Introduction

Medical image processing is the technique for visualization of interior parts and
posterior parts of the body for clinical analysis. Medical image processing helps to
represent interior structures of the bone, muscle and is useful for diagnosis.

K. Ghoti (B) · U. Baid · S. Talbar

Center of Excellence in Signal and Image Processing, SGGS Institute of Engineering and
Technology, Nanded, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 519
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_52
520 K. Ghoti et al.

Musculoskeletal disorders are commonly due to an injury to the bones, joints,

muscles, tendons, ligaments or nerves. There are different reasons behind this cause
like by jerk in muscles or bone, car accidents, falls, fractures, sprains, dislocations,
overuse and direct damage to muscle [1]. The locomotor or human musculoskeletal
system provides the ability for movement of muscles, form, stability, and support.
The purpose of a skeletal X-ray is to identify and evaluate bone density and structure.
Radiography shows the internal form of an object; it uses imaging techniques
such as X-ray, gamma rays or similar ionizing radiation and non-ionizing radiation.
As the human body made up of different bone densities, therefore ionizing and non-
ionizing radiations help for viewing the internal structure of the bone. Radiography
is also used in the field of security, such as airport and industrial area [2]. Wilhelm
Roentgen discovered the X-ray in 1895, which is the main diagnostic tool used in
the medical field. There is approximately 650 medical and dental examination per
thousand patient per year.
X-ray is the high energy electromagnetic radiation which is ranging between 0.01
and 10 nm. It uses the small amount of radiation which is passed through our body
to capture the injury or crack in the bone, diseases like infections, degeneration of
bones and tumors.
X-ray image shows the different shades for the various part of our body because
various tissues absorb the different amount of radiation. Bones look white in the
image because they absorb more radiation whereas fat, tissues absorb less due to its
which it looks gray. Air absorbs the least radiation. These are the four densities of
the X-ray [3].
Bone is a rigid organ which comprises part of the vertebral skeletal. Bone is to
Protects and supports the different organs of the body. There are various types of
bones present in the human body like long, short, fat, irregular, and sesamoid bones.
There are about 206 bones that are present in our body with different structure,
size, and shape. It is composed of around 300 bones at birth [4]. Fracture in the bone
is a very common problem in the humans and is classified as follows,
• Traumatic fracture.
• Pathologic fracture.
• Periprosthetic fracture.
Fracture (Crack) can also be classified based on soft tissue involvement as:
• Closed fracture.
• Open fracture/compound fracture.
Other than these fractures or crack can also be classified using other categories,
i.e. displacement, fracture pattern, fragments, etc. Today, there are numerous types
of tools such as CT (computed tomography), MRI (magnetic resonance imaging),
Ultrasound, etc. are available for detecting the abnormalities present in the X-ray
body is most commonly used in the detection of the fracture because it has some
advantages like it is faster and easier for the doctors to studying about the bones
and joints. Doctors generally prefer the X-ray technique to detect whether the crack
exists or not and for the exact location of the crack.
MURA: Bone Fracture Segmentation Using a U-net Deep Learning … 521

This study gives details about the different methods for fracture or crack detection
based on X-ray images. For that purpose, we have been studied frequently used
techniques of previous papers. For this paper, we used the MURA (musculoskeletal
radiographs) dataset, which contains seven different classes such as hand, wrist,
finger, shoulder, hummus, elbow, forearm. These images are classified as normal
and abnormal in the classification task [5].
The final goal of the paper is to segment out the lesion from the remaining bone
tissue that the temporal changes can be examined. The use of a deep learning approach
based on the U-Net model in if to avoid the human interventions and it can be seen as
automatic and accurate combine bone and fracture segmentation. This could permit
doctors to find a similar lesion in the bone X-ray images.

2 Anatomy of Upper Extremity

In this paper, MURA dataset is used only upper extremity part of the human body in
shown Fig. 1.
In MURA dataset, seven different upper extremity part of the human body are
present such as elbow, humerus, shoulder, finger, forearm, hand and wrist showed in
Fig. 2. The shoulder is an upper part of the back and arms. Long bone in the arm that
runs from the shoulder and elbows is known as the humerus. Elbow is the visible
joint between the upper and lower part of the arm. The forearm is part of the arm
between the elbow and wrist. The wrist is the joint connecting the hand with the
forearm. The finger is a basic part of the hand. In the above study, different views
are present in Fig. 3.

Fig. 1 Basic diagram upper extremity part of human body [6]

522 K. Ghoti et al.

Fig. 2 Upper extremity part included shoulder, wrist, elbow, forearm, humerus, hand [7]

The shoulder (radiography) different views are present such as shoulder Rx

Anteroposterior (AP), shoulder internal rotation, shoulder external rotation, shoulder
Axial, shoulder Outlet, scapula Lateral, scapula Axial Acromioclavicular joint (AC)
comparative Clavicle. Anteroposterior (AP), Lateral (LAT) internal rotation, Lateral
(LAT) external rotation, Oblique (OBL) this is the view of the humerus (radiog-
raphy). Anteroposterior (AP), Lateral (LAT), Oblique (OBL), Greenspan projection
is the view of the elbow (radiography).
Only two views are present in the forearm (radiography) such as Anteroposterior
(AP), Lateral (LAT). Wrist (radiography) six different view is present Anteroposterior
(AP), Lateral (LAT), Internal rotation, External rotation, Carpal tunnel. Three views
are present in hand (radiography) Antero- posterior (AP), Lateral (LAT), Oblique
(OBL).
MURA: Bone Fracture Segmentation Using a U-net Deep Learning … 523

Fig. 3 Different view of shoulder, elbow, wrist, hand and finger [7]

3 Mura

MURA means Musculoskeletal radiographs. Mura is a big database [8] of bone X-

rays are shown in Fig. 4. The X-ray study is normal or abnormal, algorithms are tasked
determining the 1.7 billion people worldwide affect Musculoskeletal conditions, and
they are the most common cause of long-term pain and disability, the 30 million
emergency department visits annually. Our dataset provides accurate localization of
the fracture in the bone due to which experts can diagnose at an early stage.
MURA is a big dataset of the upper extremity part of bone X-rays. This dataset
is publicly available for research. The total number of patient and images in mura
dataset is 14,656 and 40,561. Dataset consists of 36,808 train images and 3753
validation images. Each consists of upper extremity study training and validation
types such as Elbow, Finger, Forearm, Hand, Humerus, Shoulder and Wrist. The
labels are manually created as fractures or non-fracture by the experts. The total size
of the MURA dataset is 3.5 GB. Images format in the MURA dataset is’.png’. But
we have 2000 images use for bone fracture segmentation. 2000 images for training
images and 2000 for the mask. Training and masks images are splinted in training
and testing images. 25% images for testing.
524 K. Ghoti et al.

Fig. 4 Sample images from dataset [7]

4 Related Work

Chokkalingam et al. [9] developed a scheme to diagnose rheumatoid arthritis by

a sequence of image processing techniques. The system can be enhanced by the
improvement of the edge detection and find better segmentation technique. Mean,
Median, Energy, Correlation, Bone Mineral Density (BMD) etc. are Gray level co-
occurrence matrix (GLCM) features. After getting each and every feature, it can be
stored in the dataset. The training dataset is trained with non-inflamed and inflamed
values and with the help of classifier (neural network).
Swathika et al. [10] developed an algorithm for bone fracture detection in X-ray
images. First, they applied morphological gradient technique to remove the noise
and, i.e. enhances the image details fracture region is highlighted. After that, they
detected the edges using Canny edge detection technique. As compared to other edge
detection methods, the proposed technique shows give efficient results for fracture
detection.
Aishwariya et al. [11] discussed the technique to detect the boundaries of objects
of X-ray images in which fracture is detected. They have used a canny edge detector
to locate edges in X-ray images and to use boundary detection, the system which
detects the fracture automatically. Active Contour Model, Geodesic Active Contour
Model uses the boundary detection technique.
Lindseya et al. [12] proposed the fracture detection of bone using the Deep Convo-
lution Neural Network (DCNN). In this paper, used wrist X-ray images to detect
MURA: Bone Fracture Segmentation Using a U-net Deep Learning … 525

crack and input radiograph are first preprocessed by rotating, cropping, and applied
an aspect ratio to yield a fixed resolution of 1024 × 512. DCNN is an extension of
the U-Net architecture and achieved the fracture probability of 0.98.
Dong et al. [13] developed a brain tumor segmentation using U-Net based deep
convolutional networks. Proposed Model is trained using Multimodal Brain Tumor
Image Segmentation (BRATS 2015) dataset, which consist of 220 high-grade brain
tumor and 54 low-grade tumor cases. Cross-validation has shown that this method
can obtain promising segmentation efficiently.
Ng et al. [14], proposed a method for the classification using DenseNet169 archi-
tecture based deep convolutional network. MURA dataset is used Hight classification
accuracy of 0.93% including seven different classes.

5 Deep Learning

Deep structured learning; hierarchical learning which is part of machine learning

is called deep learning is shown in Fig. 5. It is based on the data representation. It
requires the learned features rather than manually designed because they are incom-
plete, over-specified, and it takes a long time to validate. DL is a flexible, universal
and learnable framework for the representation of the world, visual information.
DL supports both supervised and unsupervised learning. High levels feature from
the data can also be learned by using the DL. It requires a large amount of data,
and performance can be reduced if data is small. Performance is increased with an
increase in data. DL acts like a brain, learns from the ANN and allows a machine for
analyzing the data like a human.

Fig. 5 Representation of deep learning [15]

526 K. Ghoti et al.

6 Methodology

The while images which are required for this project are collected from Stanford
radiologists the images are in the size of 512 × 512 after resizing. we have used
2000 images in the MURA database including elbow, forearm, hand, finger, humerus,
wrist, shoulder.
A. Input Image: Input image is taken from the Database. The size of the image is
512 × 512 in ‘.png’ format (Fig. 6).
B. Conversion: In this stage, the input image is converted into the Gray in shown
in Fig. 7.
C. Creating Mask: For the segmentation stage, the mask of the input image is
needed. Mask is nothing but the region of interest. We can create the region of
interest of various shapes like circle, polygon, rectangle, and ellipses, but here
we used freehand shape because bones do not have any fixed shape.
Mask for bone fracture segmentation:
In the deep learning approach, we need ground truth data to compare network
generated masks with the annotated masks. Some datasets are provided with
their ground truth masks. But this dataset does not have annotated masks with
them. So, the masks are manually created, and they binarized to get the masks
as shown in the Fig. 6. The ground truth marked on MURE data is verified by
the expert radiologist.
D. Data Augmentation: Augmenting the data is nothing but increasing the number
of images in the database to get better generalization in the Model used where
more amount of the data, as well as variation, is needed so, for getting enough
variation, we must generate more data from the dataset with limited images.
This is what data augmentation is. The augmentation process increases the
training images based on the operations which are used when a neural network
is developed, which improves the overall network performance whenever the

Fig. 6 a Input from the

database. b Mask of the
input image
MURA: Bone Fracture Segmentation Using a U-net Deep Learning … 527

Fig. 7 Overview of the

proposed system Input Image RGB Gray

Model Creating
Mask

Result

training images in the database are relatively smaller in number. Augmentation

includes horizontal, vertical flipping and adding random Noise, zooming and
blurring.
E. Model: U-Net architecture is proposed by Ranneberger et al. [16]. U-Net model
is developed for the Biomedical image segmentation, which is basically a
CNN (convolutional neural network) is shown in Fig. 8. The architecture of
this fully convolutional network is modified and extended for less training
images to gives more precise segmentation. It follows the encoder and decoder
approach. Its architecture consists of three parts 1. contracting or downsampling
part; 2. Bottleneck layer; 3. Expansive or upsampling part this is the u-shaped
architecture.

a. Downsampling Path: It is also called a contracting path. It has two 3 ×

3 convolutions followed by the ReLU and 2 × 2 max-pooling for the
Downsampling. At each stage of the downsampling and the Number of
features, the channel is doubled.

Fig. 8 Architecture of U-net

528 K. Ghoti et al.

b. Upsampling Path: It is also known as expansive path. It performs the

Upsampling of the feature map by using the 2 × 2 convolutions, concate-
nation between feature map from the downsampling and 3 × 3 convolution
followed by the rectified linear unit.
c. Skip Connection: Skip connections from the down- sampling path are
concatenated with a feature map during the upsampling path. The
skip connection provides local information to global information while
upsampling.
d. Final Layer: Finally, the 1 × 1 convolution is used at the final layer of
mapping the feature vector to the number of classes. each feature vector
to the desired no. of classes.
e. Advantages of U-net: This U-net follows the encoder and the decoder
approach. The three parts of this architecture give it U shape; therefore, it
is known as the U-net model.

7 Results

The proposed bone fracture segmentation method is implemented on Python version

3.6. We used the test and train images for the U-Net segmentation, test images are
the input image, and train images are the mask images.
The images are of size 512 × 512; however, they are converted into grayscale
images. The system has been tested on more than 2000 images. The proposed method
shows the efficient and successful results for the segmentation of the Fracture regions
in an affected bone X-ray image.
Consider, S and G are segmentation results and the ground truth. Then, the Dice
coefficient is defined as for Eq. 1,

2(G ∩ S)
DICE = (1)
|G| + |S|

The total number parameter: 31,030,593 in U-Net Model.

The training and validation Dice coefficient of the dataset is given in the Table 1:
Thus, Fig. 9 shows the result for bone fracture segmentation where the similarity
between the segmentation performance by network and the segmentation masks
which are manually create is shown. The first image in each row represents input
images, the second image is of the manually create masks, and the third image in
each row represents the masks generated by our network.

Table 1 Accuracy of test and Train images bone fracture segmentation

Dataset Training dice coeff Validation dice coeff
MURA 0.9595 0.9029
MURA: Bone Fracture Segmentation Using a U-net Deep Learning … 529

(a-1) Input image (b-1) Mask Image (c-1) Predicted

(a-2) Input image (b-2) Mask Image (c-2) Predicted

(a-3) Input image (b-3) Mask Image (c-3) Predicted

Fig. 9 Result for bone fracture segmentation

530 K. Ghoti et al.

8 Dependencies

For the deep learning approach, there are some libraries, and high computation
systems are required. Following some dependencies used for the training model,
Keras with a TensorFlow backend Next to this We used scikit-learn, simpleitk,
beautiful soup, OpenCV.

9 Conclusion

The paper presented bone fracture Segmentation Using Deep learning. This method
required to test and train images that contains the original input image and mask
images, respectively. The system has been tested on images obtained from Miles
Research. The proposed method gives approximately 95.95% for training accuracy
and 90.29% for testing accuracy, including bone fracture segmentation. The proposed
system is very effective and can be implemented in real-time applications. We can
also classify the bone X-ray images, whether it is either normal or abnormal. The main
advantage of the approach, which is presented in this paper is its uniform nature and
can be applied to different medical image segmentation tasks. The proposed Model
achieved an acceptable performance with the MURA dataset images.

References

1. URL https://www.boneandjointburden.org/2014-report (2017)

2. Khatik I (2017) A study of various bone fracture detection techniques. Int J Eng Comput Sci
6(5)
3. Johari N, Goyal S, Singh N, Pal A (2016) Bone fracture detection algorithms based on image
processing- a survey. Int J Eng Manag Res 36, 2
4. Kaur T, Garg A et al. (2016) Bone fraction detection using image segmentation. In: Int J Eng
Trends Technology (IJETT), 36(2)
5. Rajpurkar P, Bagul A, Ding D, Ball RL, Langlotz C, Ng A et al (2017) MURA ataset: towards
radiologist-level abnormality detection in musculoskeletal radiographs, arXiv preprint arXiv:
1712.06957v2
6. https://images.app.goo.gl/yX76GigDgLgkph6k6
7. https://www.stanford.edu/site/terms/
8. Artificial intelligence in medicine and imagin (2017) Available labeled medical datasets. https://
aimi.stanford.edu/available-labeled-medical-datasets, [Online; Accessed -2 Dec 2017]
9. Chokkalingam SP, Komathy K et al (2014) Intelligent assistive methods for diagnosis of
rheumatoid arthritis using histogram smoothing and feature extraction of bone images. World
Acad Sci Eng Technol Int J Comput Inf Syst Control Eng 8(5):834–843
10. Swastika B, Anandhanarayanan K, Baskaran B, Govindaraj R (2015) Radius bone fracture
detection using morphological gradient based image segmentation technique. Int J Comput Sci
Inf Technol 6(2):1616–1619
11. Aishwarya R, Kalaiselvi GM, Archana M et al (2013) Computer- aided fracture detection of
X-ray images. IOSR J Comput Eng 44–51
MURA: Bone Fracture Segmentation Using a U-net Deep Learning … 531

12. Lindsey R, Daluiski A, Chopra S, Chapelle ALA et al (2018) Deep neural network improves
fracture detection by clinicians. In: PNAS
13. Dong H et al (2017) Automatic brain tumor detection and segmentation using U-net based
fully convolutional networks. In: MIUA 2017, 069, v3: ’Automatic Brain Tumor Detection
and Segmentation Using U-Net
14. Rajpurkar P, Irvin J, Bagul A, Ding D, Duan T, Mehta H, Yang B, Zhu K, Laird D, Ball RL,
Langlotz C, Shpanskaya K, Lungren MP, Ng A et al (2018) MURA, 22 May 2018, “Large
dataset for abnormality detection in musculoskeletal radiograph.” arXiv preprint arXiv:1712.
06957v4.
15. https://images.app.goo.gl/RdQGf9XyrJrcwXh66
16. Ronneberger O, Fischer P, Brox T (2015) 18 May 2015, U-Net: Convolutional networks for
biomedical image segmentation. arXiv:1505.04597v1
Effective Usage of Oversampling
with SMOTE for Performance
Improvement in Classification Over
Twitter Data

Deepak Patil, Poonam Katyare, Parag Bhalchandra, and Aniket Muley

Abstract This paper highlights an attempt for addressing the issue of imbal-
anced classification resulted due to deployment of machine learning algorithms
over an imbalanced dataset. It has used Synthetic Minority Oversampling Tech-
nique (SMOTE). This type of augmentation of the dataset is extremely necessary
as it leads to poor performance in the minority class. Four machine learning algo-
rithms were deployed on the Twitter dataset using the Python platform. Standard
data preprocessing including data cleaning, data integration, data transformations,
and data reduction was carried out first as the most necessary arrangement before
experimentations.

Keywords Social network analysis · SMOTE · Minority class · Imbalanced

classification

1 Introduction

The performance issue of classification-based machine learning algorithms has been

a topic of interest. Since classification algorithms get training to predict a particular
group for a data point, an imbalanced dataset will make it biased to wrongly classify
data points [1, 2]. A possible solution could be class-based deployment of the samples
[2, 3]. This is done by using the concept of minority class which is particularly the
correct place to get classified and a majority class where biased classification will
get intended. Research in past has tackled this question using the under sampling
of the majority class and oversampling of the minority classes. The combination of

D. Patil (B)
Department of Computer Science, Smt. K.R.P. Kanya Mahavidyalaya, Islampur, Sangli, MS, India
P. Katyare
MCA Department, PCCOE, Pune, India
P. Bhalchandra
School of Computational Sciences S.R.T.M. University, Nanded, MS, India
A. Muley
School of Mathematical Sciences, S.R.T.M. University, Nanded, MS, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 533
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_53
534 D. Patil et al.

under sampling and oversampling can also work depending upon the situation and
distribution of the sample. The SMOTE (Synthetic Minority Over-sampling Tech-
nique) [2, 4] is an example of oversampling approach where we do not duplicate data
points or examples in the minority class rather we synthesize new examples from the
existing examples so as to add new information in the model. This is done by selecting
minority groups and making changes in the existing minority instances. This is also
known as amplification. To demonstrate the working of SMOTE for performance
improvement of selected classification algorithms, we have used a Twitter dataset
[5] that has 973 networks with 81,306 nodes and 1,768,149 edges. The dataset has
been anonymized to preserve confidentiality. An Ego network is drawn to represent
this database using positive and negative edges. To demonstrate the actual working,
we have identified two classes for our study and this created class imbalance problem.
Due to imbalance distribution of class in the considered dataset we witnessed down-
graded performance of the selected machine learning algorithms. This is addressed
by re-sampling of dataset [3, 4]. Accordingly SMOTE method was implemented.
Numbers of experiments are done to find best accurate classification model for our
study according to the objectives.

2 Experimental Set Up

Standard data preprocessing including data cleaning, data integration, data trans-
formations, and data reduction [1, 6] was carried out first as the most necessary
arrangement before experimentations. Data reconstruction [6, 7] was done later
to precisely realize under sampling as well as over sampling strategies. We have
assumed that the imbalance is there by default. Hence one class was labeled as
majority and other as minority [7]. Imbalance distribution always hampers training
mode and perhaps make entire training biased by totally ignoring minority class.
Hence random sampling was done and the dataset was randomly re-sampled. This is
done either of the way like, delete examples from majority class (under sampling) or
duplicate examples if the minority class(over sampling). We work on the assumption
that the predictions on minority class are important for improved performance. For
these issues, we have used SMOTE (Synthetic Minority Oversampling Technique)
for oversampling of dataset [5]. Its standard algorithm [2, 3] is presented in Fig. 1.
Table shows an example of calculation of random synthetic samples. The amount
of over-sampling is a parameter of the system, and a series of ROC curves can be
generated for different populations and ROC analysis performed.
During the course of experimentations, we have found that our Twitter dataset
ego network graph shows approximately 21,476 number of positive sample edges in
the majority class and 10,738 number of negative sample edges in the minority class.
This was, however, an observation for the training set used in cross-validation. The
SMOTE algorithm in Fig. 1 selected minority class instance first at a random and
then found its k nearest minority class neighbors. Creation of synthetic instances then
started by choosing randomly, one of the k nearest neighbors and then connecting
Effective Usage of Oversampling with SMOTE for Performance … 535

Fig. 1 SMOTE algorithm

536 D. Patil et al.

Fig. 2 Positive and negative Positive & Negative Edges

edges of the ego network for
Twitter dataset 25000 21476
20000
15000 10738
10000
5000
0
Number of Positive Sample Number of Negative
edges Sample edges

these both to form a line segment in the feature space. The synthetic instances
are generated as a convex combination of the two chosen instances. This mecha-
nism of creating synthetic creation can be repeated as many times as needed until a
targeted balanced percentage is reached. However, this is great deal of disadvantage
to SMOTE approach to take into consideration new synthesized instances without
taking into consideration of the majority class [2, 8]. Figure 2 shows positive and
negative edges of the ego network during the course of observations.
Now, we have applied selected machine learning algorithms for classification
including K-Nearest Neighbors Algorithm (KNN), Logistic Regression Classifier
Algorithm (LRC), Artificial Neural Network Algorithm (ANN) and Support Vector
Machine Algorithm (SVM) for comparison of the performance improvement.
The KNN [1, 9] is a widely used classification algorithms which makes prediction
based on the neighborhood distance with an assumption that the similar things exist
in proximity [10].
The LRC [1, 9] is typically recommended for categorical based classification. It
is typically used for binary classification problems. Decisions in LRC are made with
the help of a logistic curve limited to values between 0 and 1. A sigmoid function is
used for testing of positive and negative values [11].
The ANN [1] works mimicking of human brains. We have used a multilayer
perceptron as a common artificial neural network model. It has an input layer, one
or more hidden layers, and an output layer. Neurons have an activation function that
operates upon the value received from the input layer. The outputs from the first
hidden layer neurons are multiplied with the weights of the second hidden layer; the
results are summed together and passed to the neurons of the proceeding layers. This
process continues until the outer layer is reached.
The SVM [9, 10] is extremely recommended for two-class learning classification
problems. It uses a hyper plane for separation of these two classes.
All these considered algorithms are implemented on the dataset and their perfor-
mance is evaluated using a confusion matrix. This matrix has rows and columns
where rows are the actual class and the columns show predicted class. In order to
tabulate results, we assumed following terminology,
Effective Usage of Oversampling with SMOTE for Performance … 537

1. Use of Confusion matrix for describing performance of the model. It has some
attributes like False Positive (FP), False Negative (FN), True Positives (TP) and
True Negatives (TN).
2. Total sample count is sum of all FP, FN, TP and TN.
3. Use of Accuracy parameter to record correctness of the model, which is TP +
TN/Total sample count.
4. Precision is taken as TP + FP/Total sample count.
5. Recall is taken as TP + FN/Total sample count.
6. F1 Score is harmonic mean of Precision and Recall as they both deal with
relevance. It is best value at 1 and worst at value 0.
7. Area Under the Curve (AUC) shows ability of the model to distinguish between
FN and FP classes and used as a synopsis of receiver operator characteristic
(ROC) curve that narrates about performance of binary classification algorithm
[11].
Above considerations used to measure the performance of the model are shown
in the Table 1. Graphical representations of the results are shown in Figs. 3 and 4.

Table 1 Experimental results

Algorithm Accuracy % Precision Recall F1-score ROC-AUC score
KNN 90.31 90.01 90.73 90.37 90.30
LR 75.89 80.54 68.44 74.00 75.91
ANN 87.77 81.17 98.43 88.97 87.74
SVM 77.70 80.81 72.80 76.60 77.71

Fig. 3 Accuracy of the Algorihm Accuracy

algorithms 95
90.31
90 87.77

85

80 77.7
75.89
75

70

65
KNN LR ANN SVM
538 D. Patil et al.

Fig. 4 Precision, recall and Algorithm Performance

F1 scores of the algorithms

100

50 F1-score
Recall
0 Precision
KNN LR ANN SVM
Precision Recall F1-score

3 Conclusion

This paper demonstrates data augmentation for the imbalanced dataset related to
minority class using SMOTE approach. This was mainly done to demonstrate the
performance improvement of selected classification algorithms over imbalanced
Twitter dataset. Experiments were done on the python platform. It is witnessed that
the KNN showed maximum accuracy whereas Logistic regression as the lowest
accuracy for the SMOTE based performance enhancements.

References

1. Jan S, Ruby R, Najeeb PT, Muttoo MA (2017) Social network analysis and data mining, by ©
2017, IJCSMC Saima Jan et al. Int J Comput Sci Mob Comput 6(6):401–404
2. Chawla NV, Bowyer KW, Hall LO, Kegelmeyer WP (2002) SMOTE: synthetic minority over-
sampling technique. J. Artif Intell Res 16:321–357
3. Mohammed AJ, Hassan MM, Kadir DH (2020) Improving classification performance for a
novel imbalanced medical dataset using SMOTE method. Int J Adv Trends Comput Sci Eng.
9(3). ISSN 2278–3091
4. Papagelis M, Das G (2010) Sampling online social networks. IEEE Trans Knowl Data Eng
X(X):Month 20xx
5. Web resource at https://www.kaggle.com
6. Boyd D, Ellison N (2009) Social network sites. J Comput-Mediated Commun from: https://
jcmc.indiana.edu/vol2013/issue2001/boyd.htl
7. Wasserman S, Galaskiewicz J (1994) Advances in social network analysis: research in the
social and behavioral sciences. Sage Publications, Thousand Oaks
8. Zheng Z, Cai Y, Li Ye (2015) Oversampling method for imbalanced classification. Comput Inf
34:1017–1037
9. Jiawei H, Kamber M (2001) Data mining: concepts and techniques San Francisco. California,
Morgan Kauffmann
10. Muthuselvi et al (2012) Information retrieval from social network, IJCA. In: Proceedings on
E-governance and cloud computing services vol 4
11. Erdogan SZ et al (2006) A data mining application. A student database. J Aeronaut Space
Technol 2(2)
Multi-Classification of Breast
Histopathological Image Using Xception:
Deep Learning with Depthwise Separable
Convolutions Model

Suvarna D. Pujari, Meenakshi M. Pawar, and Madhuri Wadekar

Abstract One of the best methods for Breast cancer diagnosis is histopathological
images from the visual analysis of histology, but pathologist requires lots of expe-
rience and training to an accurate diagnosis. Therefore, computer-aided diagnosis
(CAD) is an automated and more precise method. Recent developments in computer
vision and deep learning (DL), DL based models are popular in analyzing the hema-
toxylin–eosin (H&E) stained breast cancer digital slides. This paper proposed a
deep learning-based framework, called multi-classification of breast histopatholog-
ical image using Xception: Deep Learning with Depth wise Separable Convolutions
model (MCBHIX). Xception based on depthwise separable convolution layer. We
trained this network from scratch for binary classes and for multi-classes BreakHis
dataset. The accuracy achieved by MCBHIX- 99.01% for binary type and 96.57%
for Multiclass.

Keywords Histopathological image · BreakHis · Deep learning · Xception

model · Depth wise convolution

1 Introduction

Breast cancer (BC) is the most common type of cancer, especially in women, 14%
of cancers in Indian women. For every four minutes, an Indian woman is diagnosed
with breast cancer. A 2018 report of Breast Cancer statistics recorded 1, 62,468 new
registered cases and 87,090 reported deaths. The survival rate Cancer is 60% for the

S. D. Pujari (B) · M. M. Pawar · M. Wadekar

SVERI’s College of Engineering, Pandharpur.Gopalpur- Ranjani Road, Gopalpur, Pandharpur,
Dist-Solapur 413304, India
e-mail: [email protected]
M. M. Pawar
e-mail: [email protected]
M. Wadekar
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 539
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_54
540 S. D. Pujari et al.

Indian women as compared to 80% in the U.S [1]. In 2020 BC goes as high as two
million [2].
More accurate and early detection and diagnosis of BC reduce the mortality
rate, there are many tests available including physical examination, Mammography,
tomosynthesis, magnetic resonance imaging(MARI), ultrasound and biopsy [3]. Out
of them, biopsy serves as the gold standard for BC detection. Pathologists need profes-
sional background, rich experience for manual classification of breast hematoxylin
and eosin stained slides is costly. Recent improvements in biomedical image anal-
ysis using computers, to improve clinicians accuracy the computer-Aided Diagnosis
(CAD) systems have developed that can help pathologists to be more productive,
objective and consistent in diagnosis.
For classification BC, many handcrafted feature-based algorithms like geometric
means of symmetric positive definite matrices (mSPD) [4, 5], morphological [6]. But
it is hard to classify H & E stained images using handcrafted features, due to complex
structure, huge variation in inter-class and intra-class, the similarity in micro-
scopic images these are challenges. Supervised Machine learning algorithms-support
vector machine(SVM) [2, 7], K-Nearest Neighbor, Decision Tree [8] Recently,
Deep learning is continuously fueling medical image processing, bilinear convo-
lutional neural network (BCNN) [9], Deep layer CNN(DCNN) [10], VGG16 [11],
ResNet(MuDeRN) [12] these classified in only two classes benign and malignant.
Malignancy tumour detection is still a challenging problem. In this work, we use
the openly available BreaskHis Dataset, which contains eight types. We designed a
model which does a binary classification as well as multi-classification images into
eight classes’. There are treatment options available for BC patients, and the BC
subtype could help to predict the patient’s response to therapy; for example, invasive
lobular cancer gains a clear benefit from systemic treatment when compared to inva-
sive ductal cancer. The correct recognition of benign lesion type is also important
because the patient’s risk of developing subsequent BC varies among different types
of benign lesions [12]. Xception Model which is entirely dependent on depth wise
convolution layers [13] which reduces the computational cost of the training model.

2 Methods and Materials

2.1 Dataset Used

BREAKHIS DATABASE contains microscopic biopsy breast tumor collected from

82 patients using four magnification levels −40X, 100X, 200X, and 400X of
700 X 460 pixels resolution, three-channel RGB color images shown in Fig. 1. This
database has built-in P&D Laboratory Pathological Anatomy and Cytopathology,
Parana, Brazil.
Multi-Classification of Breast Histopathological Image Using Xception: Deep … 541

Fig. 1 Some samples of different categories

Table 1 The BreakHis database distributed into four magnification levels and main classes, their
subcategories
Class Sub-class Magnification level Total
40X 100X 200X 400X
Benign Adenosis 114 113 111 106 444
Fibroadenoma 253 260 264 237 1014
Tubular adenoma 109 121 108 115 453
Phyllodes tumor 149 140 150 130 569
Malignant Ductal carcinoma 864 903 896 788 3451
Lobular carcinoma 156 170 163 137 626
Mucinous carcinoma 205 222 196 169 792
Papillary carcinoma 145 142 135 138 560
Total 1995 2081 2013 1820 7909

The BreakHis dataset mainly consists of two main classes: Benign and malignant,
from 24 patients have 2480 benign images, and from 58 patients have 5429 malignant
images [14, 15].
The benign and Malignant Histopathological images again divided into
four subclasses each. Benign classes-Adenosis(A), Fibroadenoma(F), Tubular
Adenoma(TA), Phyllodes Tumor(PT) and Malignant Classes-Ductal Carci-
noma(DC), Lobular Carcinoma(LC), Mucinous Carcinoma(MC), Papillary Carci-
noma(PC).The statistics of all data given in Table 1.

2.2 Methodology

Extracting mitosis from H & E stained images using segmentation techniques is

not easy as Deep learning-based algorithm to classify these images due to misclas-
sification with non-mitosis [5, 6, 16]. As shown in Fig. 2 proposed MCBHIX
model for microscopic image classification consists of 1. Input image 2.data
augmentation.3.classifier. This frame gives the whole slide image as input (WSI).
542 S. D. Pujari et al.

Fig. 2 Flow diagram of proposed MCCBHIX model

The input image is from a database with 229 × 229 pixel resolution having
morphological features took, which is then augmented using Keras ImageDataGen-
erator function. Augmentation means creating more images adding some operation
like rotation, zooming, shift, and flip with slight modification without changing the
original structure.
Classifier: The classifier for classification we used the DCNN based xception
model is a low-cost, low computation model, as shown in Fig. 4. It is entirely based
on the Depth wise separable convolution layers. It has 36 convolution layers of
kernel size 3 × 3 to extract features which structured into 14 modules, all have a
linear residual connection between them except the first one [13].
Residual called identity/skip links solve vanishing gradient problems. A convolu-
tion layer used to extract the features, learn features from frames in CNN. The depth
wise Convolution shown in Fig. 3. Reduce computational cost shown in (Eqs. 2, 3)
which reduce the over fitting and computational cost.

Fig. 3 Depth wise convolution

Multi-Classification of Breast Histopathological Image Using Xception: Deep … 543

Fig. 4 Xception: Deep Learning with Depth wise Separable Convolutions model
(arXiv:1610.02357v3)

Depth wise Convolution

Suppose the input image size of Df X Df X M, Df X Df is the image size, and M
is the number of channels and N number of kernels of Dk x Dk dimension shown
Fig. 3.
Then the convolution parameters are calculated in (Eq. 1).

N × D f 2 × Dk 2 × M (1)

The depth-wise convolution parameters are in (Eq. 2). It follows depth-wise

Convolution and then point-wise Convolution.
Depth wise convolution–M × DK2 × Dp2 .
Point-wise convolution- M × Dp2 × M

M × Dk2 × Dp2 + M × Dp2 × N = M × Dp2(Dk2 + N) (2)

Complexity of Depthwise Convolution

Ratio =
Complexity of standard Convolution
M × Dk2 × Dp2 + M × Dp2 × N 1 1
= = + (3)
N × Df2 × Dk2 × M N Dk2
544 S. D. Pujari et al.

(1/N + 1/Dk2) times lesser multiplications for the depth-wise Convolution than
the standard Convolution (Eq. 3).
The proposed model uses Adam optimizer, categorical cross-entropy and relu as
an activation function.
Adam Optimizer: Adam is an adaptive moment based algorithm used to update
the network weights during the training of data. Adam works. It is memory efficient,
computationally efficient and requires less hyperparameter tuning [17].
Loss: Categorical cross-entropy is log loss and used for multi-class classification
shown in (Eq. 4) [18].
  
Categorical cross − entropy = − [( y log ŷ) + (1 − y)log 1 − ŷ ] (4)

ReLu: Rectified Linear Unit is the activation function, gives the output if the input
is positive otherwise zero. Many networks use it as a default activation function.

3 Result Discussion

The experimental results of the proposed MCBHIX model have compared with
existing VGG16. MCBHIX has trained on GPU Tesla P100, for 100 epochs, in
Google Colab Notebook is a free cloud service and supports free GPU. MCCHIX
performance has better than VGG16 in terms of accuracy, Loss and computation
shown in Table 2. We achieved accuracy- 96.57%, which is 23.08% higher, Loss −
8.6%, which is 11.88% lesser than VGG16s accuracy and Loss has shown in Table
2.

Table 2 Accuracy (%), Loss (%) and many other parameters of MCBHIX and VGG16
Classification CNN Image Accuracy Loss CNN Computational
architecture size (%) (%) layers Parameters (million)
Multi-class VGG16 224 × 73.49 20.48 16 14.9
24x3
MCCBHIX 229 × 96.57 8.6 36 22.9
229 x 3
Binary VGG16 224 × 92.58 19.56 16 14.7
224x 3
MCCBHIX 229 × 99.01 1.68 36 22.5
229x3
Multi-Classification of Breast Histopathological Image Using Xception: Deep … 545

4 Conclusion

In this paper, we proposed the MCBHIX model to classify H&E stained breast
histopathological microscopy images into (i) Binary classification and (ii) Multi-
class Benign and Malignant classes into eight categories A, F, TA, PT, DC, LC, MC,
PC. This model has trained on the BreakHis dataset. We achieved an accuracy of
99.01% for binary class and 96.57% for Multi-Class classification.

References

1. Rangarajan B et al (2016) Breast cancer: an overview of published Indian data. South Asian J
Cancer 5(3):86
2. Reddy A, Soni B, S Reddy (2020) Breast cancer detection by leveraging machine learning.
ICT Express
3. Guo Y, Shang X, Li Z (2019) Identification of cancer subtypes by integrating multiple types
of transcriptomics data with deep learning in breast cancer. Neurocomputing 324:20–30
4. Khan AM, Sirinukunwattana K, Rajpoot N (2014) Geodesic geometric mean of regional covari-
ance descriptors as an image-level descriptor for nuclear atypia grading in breast histology
images. In: International workshop on machine learning in medical imaging. Springer, Berlin.
5. Janowczyk A, Chandran S, Madabhushi A (2013) Quantifying local heterogeneity via
morphologic scale: distinguishing tumoral from stromal regions. J pathol informa. 4(Suppl)
6. Das A, Nair MS, Peter DS (2020) Batch mode active learning on the riemannian manifold for
automated scoring of nuclear pleomorphism in breast cancer. Artif Intell Med 103:101805
7. Chekkoury A et al (2012) Automated malignancy detection in breast histopathological images.
In: Medical imaging: computer-aided diagnosis. 2012. International society for optics and
photonics
8. Gupta P, Garg S (2020) Breast cancer prediction using varying parameters of machine learning
models. Procedia Comput Sci 171:593–601
9. Wang C et al (2017) Histopathological image classification with bilinear convolutional neural
networks. In 2017 39th annual international conference of the ieee engineering in medicine
and biology society (EMBC). IEEE
10. Zainudin Z, Shamsuddin SM, Hasan S (2019) Deep layer CNN architecture for breast cancer
histopathology image detection. In: International conference on advanced machine learning
technologies and applications. Springer, Berlin
11. Toğaçar M et al (2020) BreastNet: a novel convolutional neural network model through
histopathological images for the diagnosis of breast cancer. Phys A 545:123592
12. Gandomkar Z, Brennan PC, Mello-Thoms C (2018) MuDeRN: multi-category classification
of breast histopathological image using deep residual networks. Artif Intell Med 88:14–24
13. Chollet F ( 2017) Xception: deep learning with depthwise separable convolutions. In:
Proceedings of the IEEE conference on computer vision and pattern recognition
14. Yosinski J et al (2014) How transferable are features in deep neural networks? In: Advances in
neural information processing systems
15. Benhammou Y et al (2020) BreakHis based breast cancer automatic diagnosis using deep
learning: taxonomy, survey and insights. Neurocomputing 375:9–24
546 S. D. Pujari et al.

16. Cireşan DC et al (2013) Mitosis detection in breast cancer histology images with deep neural
networks. In: International conference on medical image computing and computer-assisted
intervention. Springer, Berlin
17. Kingma DP, Ba J (2014) Adam: a method for stochastic optimization. arXiv preprint arXiv:
1412.6980
18. Ho Y, Wookey S (2019) The real-world-weight cross-entropy loss function: modeling the costs
of mislabeling. IEEE Access 8:4806–4813
Dense Haze Removal Using Convolution
Neural Network

Mayuri Dongare and Jyoti Kendule

Abstract Pictures caught in murky climate show up low conversely. Debasement

in the picture contrast is because of lessening in the light energy reflected from
the scene object. In this paper, we propose a picture de-right of passage network
which upgrades the perceivability of pictures caught in murky climate. The proposed
network comprises of multi-scale convolution channels consolidated by commence-
ment module to extricate the multi-scale highlights. Alongside the multi-scale high-
light extraction, we propose a utilization of thick associations with engender learned
highlights inside the origin modules. Combinely, the proposed network is planned by
joining the standards of both initiation and thick module, along these lines, named
as beginning thick organization. To prepare the proposed network for picture de-
inception, we utilize primary similitude list metric alongside the L1 misfortune.
Existing benchmark information bases are used to assess the favorable to presented
network for picture de-right of passage. Exploratory examination shows that the
proposed network beats the current methodologies for picture de-preliminaries.

Keywords Hazy image · Train · ID-net · Generative adversarial network

1 Introduction

Visibility of the outdoor images decreases due to the presence of fog within the
atmosphere. Thus, within the presence of the haze or fog particles, a computer vision
algorithm faces difficulty to realize the specified output as generally they anticipate
an information picture without a quality debasement. In this manner, the presence
of the cloudiness or haze particles within the atmosphere degrades the performance
of computer vision algorithms like object detection [1], moving object segmentation
[2] etc. Therefore, to enhance the performance of vision algorithms within the hazy
environment, image de-hazing may be a required pre-processing task. Research in
the field of image de-hazing is roughly divided into learning based methods [3–5]

M. Dongare (B) · J. Kendule

SVERI’s COE, Pandharpur, Maharashtra, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 547
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_55
548 M. Dongare and J. Kendule

Fig. 1 Haze-free image recovered by the proposed method. Left column: Input hazy image Right
column: Haze-free image recovered using the proposed method

and prior based methods [6–13]. Among these, prior based methods believe the haze
relevant prior and extract haze relevant features.
Figure 1 shows the outdoor hazy image and the haze-free image recovered by the
proposed net- work. The proposed network is built using basic principles of inception
and dense modules. Thus, named as inception-dense network (ID-Net). The proposed
ID-Net bypasses the estimation of intermediate feature maps and directly recovers
the haze-free image.
The key features of this work are listed below:
1. End-to-give up conditional generative adverse community named as ID-Net is
proposed for picture de-hazing.
2. A novel generator community is proposed that’s de- signed the usage of a
aggregate of each inception and dense module.
3. Experimental evaluation has been done on current benchmark datasets.

2 Literature Survey

Haze impact is without delay proportional to the intensity of an item from the
digital digicam device. To apprehend this non- linearity, numerous methods had
been proposed consisting of polarized filters [14, 15], use of more than one pix of
equal scenery [16, 17], earlier primarily based totally hazy models [6–13] etc. At the
beginning, with inside the vicinity of photograph de-hazing, Schechner et al. [14,
15] proposed the polarized filters. Their method works with more than one pix of the
equal scene however differs in polarization angle. This method fails because of its
multi-photograph dependency. Nayer et al. [17] overcame the hardware complexity.
By utilizing 3D geometrical model Cozman et al. [16] resolved multi-image depen-
dency this technique is based upon the depth information of the hazy images. In the
last decade because of the convincing assumptions regarding the haze spread or haze
Dense Haze Removal Using Convolution Neural Network 549

density, image de-hazing has made remarkable progress. Tan et al. [11] Proposed
assessment enhancement of the hazy scene. By maximizing the nearby assessment
of the hazy photo they eliminated haze. However, whilst there may be a intensity
discontinuity within side the hazy photo this approach fails and create blocking off
artifacts. He et al. [8, 18] proposed darkish channel prior (DChP) to repair the visi-
bility within side the hazy scene. It accommodates of darkish pixels i.e. pixels which
can be having very low depth amongst one of the colour channels for a given hazy-
unfastened scene. DChP fails in complex edgy systems and additionally undergoes
the halo effect [3]. To estimate the sturdy trans- undertaking map of the hazy scene,
researchers observe post- processing strategies which include guided filtering [19],
median filtering [9, 20] etc. Lai et al. [21] proposed priors to estimate the most
excellent transmission map. Wang et al. [22] applied multi- scale retinex set of rules
to estimate the brightness components. Further, with the help of a bodily model,
they recovered the haze-unfastened photo. Zhu et al. [13] proposed a color atten-
uation prior (CAP) which having a HSV color space to extract the haze-relelated
features. To avail the benefits of multiple haze priors, Tang et al. [12] proposed
regression framework for image de- hazing. They have proposed the extraction of
varied haze relevant features using existing haze relevant priors and learned the
integrated features to estimate the robust scene transmission map. However, it will
increase the mistakes upstretched to the hired priors. Thus, to reduce the cascading
error, researchers employ convolutional neural networks (CNN). Existing learning-
primarily based totally approaches [3–5, 23, 24] estimate the scene transmission
map the usage of CNN. Further, a international airlight estimation accompanied via
way of means of atmospheric scattering version restores the haze-loose scene. All
above strategies offers the identical perception that so one can get better a haze-loose
photo, estimation of a correct scene transmission map is essential. The atmospheric
mild is calculated one by one and the easy photo is recovered the usage of the atmo-
spheric scattering version. Still any such manner does now no longer at once degree
or reduces the reconstruction distortions. As a end result it’ll supply upward thrust
to the sub-ideal photo recovery quality. The mistakes in every separate estimation
step will amplify the general error. In this aspect, Li et al. [25] proposed an end-to-
end architecture called as AOD-Net. They analyzed the internal relationship between
traditional atmospheric model and the end-to-end de-hazing network. Further, Swami
et al. [26] discussed an end-to-end network based on conditional GAN for image clar-
ification. Recently, researchers [27–29] make use of unpaired training approach for
haze removal [27, 28, 30] utilized unpaired training approach for image de-hazing
whereas [29] found its use for moving object segmentation. In the next Section, we
have discussed the proposed method for single image haze removal.
550 M. Dongare and J. Kendule

3 Proposed Method for Image Dehazing

3.1 Proposed Generator Network

The proposed generator network architecture is divided into three parts namely:
(1) Encoder block (2) Inception block, and (3) Decoder block. Encoder/Decoder
block consists of simple convolution/ deconvolution layer followed by non-linear
activation function (ReLU). We use instance normalization [31] to normalize the
network feature maps. Figure 2 show the encoder blocks. We have designed four
encoder blocks among which initial two blocks down-samples the input feature maps.
Purpose of the down-sampling operation is to increase the receptive field in the
network. Another advantage of the down-sampling the feature maps are to reduce
the computations. It is well known fact that the number of computations in CNN
are directly proportional to the spatial size of the feature maps those which are
processed through the network. Down-sampling operation is purposefully designed
so as to reduce the number of computations and to increase the receptive field. We
design four encoder and two decoder blocks with filters having a spatial size of 3 × 3.

3.1.1 Network Loss Function

It is a prime requirement of any image restoration technique to recover the structural

details. Thus, it is required to acquaint the network learning about structural loss
along with the L1 loss and adversarial loss [32]. Thus, we utilized the structural
similarity index metric (SSIM) as a loss function along with traditional L1 loss. Also,
to generate the true edge information we considered the edge loss while training the
proposed Inception-Dense Network. Therefore, overall loss function is, [33], peak
signal to noise ratio (PSNR) and color difference measure (CIEDE 2000) [34] for

Fig. 2 Generator network of the proposed inception-dense network

Dense Haze Removal Using Convolution Neural Network 551

Table 1 Quantitative
SOTS SSIM PSNR CIEDE2000
analysis of image De-hazing
on SOTS Database [37] TPAMI-11 [8] 0.8179 16.6215 9.9419
TIP-14 [13] 0.8364 19.0524 8.3102
TIP-16 [3] 0.8472 21.1412 6.2645
ECCV-16 [5] 0.8102 17.5731 10.7991
ICCV-17 [25] 0.8512 19.0868 8.2716
NIPS-18 [35] 0.8299 18.3767 11.5527
WACV-18 [4] 0.8152 20.1186 8.3060
CVPR-18 [40] 0.8800 22.3113 –
CVPR-18 [41] 0.8378 20.8100 –
ECCV-18 [42] 0.8640 20.3125 7.1152
WACV-19 [27] 0.8919 22.9207 8.9262
TIP-19 [43] 0.8261 21.3400 –
IJCV-19 [44] 0.8600 21.5600 –
TIP-19 [45, 46] 0.8716 21.4375 –
CVPRW-19 [46] 0.8500 19.8280 8.2993
CVPRW-19 [47] 0.9221 24.0200 –
Proposed method 0.9383 27.5070 5.2428

quantitative evaluation. We categorize the experiments into two parts: performance

of the proposed Inception-Dense Network on synthetic and real-world hazy image.

3.1.2 Quantitative Analysis

D-Hazy [23] is a standard dataset used to evaluate the performance of various algo-
rithms for image dehazing. It comprises of pair of 1449 indoor hazy and respective
haze-free scenes. We utilized the entire database i.e. 1449 images for quantitative
analysis of the proposed network for image de-hazing. The performance of the pro-
posed network is compared with the existing state-of-the- art methods on D-Hazy
[23] database as shown in Table 1.

4 Experimental Results

In this Section, we convey both quantitative and subjective assessment to approve

the proposed network for picture de-right of passage. We consider primary closeness
list (SSIM) increments by 5% when contrasted with start to finish profound figuring
out how to the earlier based profound learning approaches [3–5] and techniques
[25, 28, 35, 36] which shows the heartiness of RI-GAN to recuperate the fog free
552 M. Dongare and J. Kendule

scene. Additionally, there is a huge improvement in the PSNR and CIEDE2000 of the
proposed RI-GAN as contrasted and the current cutting edge strategies. Proposed
network builds SSIM by practically 9%. SOTS information base [37] is created
from set of 50 pictures and their individual profundity maps from NYU-profundity
information base [38, 39]. From each fog free picture and its profundity map, 10
foggy pictures are created with various estimations of β and airlight utilizing climatic
dispersing model. Hence, despite the fact that there is a cover of some scene in D-
Hazy and SOTS information base, distinctive airlight and thickness of murkiness
have a huge effect between them. In this manner, we assessed the exhibition of
proposed RI-GAN on SOTS information base. We thought about each of the 500
murky pictures for the examination. Table 1 depicts the result of the proposed and
existing methods on SOTS database. We can observe that proposed Dense- Inception
network outperforms the other existing state-of- the-art methods in terms of PSNR
and appears very close to AODNet [25].

4.1 Real World Hazy Images

Because of the inaccessibility of pair of this present reality murky and murkiness
free scenes, it is hard to convey quantitative examination of picture de-preliminaries
calculations for true cloudy scenes. In this way, we convey just subjective examina-
tion for this present reality dim scenes. Five as often as possible utilized true foggy
scenes are used here for investigation. Result correlation of proposed and existing
methodologies on these pictures is appeared in Figs. 3 and 4. We can plainly see
that the proposed RI-GAN produces the proper scene data simultaneously protects
the primary subtleties in recuperated dimness free scene. We think about the conse-
quences of existing earlier based hand-created and learning approaches [3, 4, 8] and
start to finish dehazing approach [25]. Qualitative examination shows that proposed

Fig. 3 Visual after effect of proposed and existing strategies on genuine foggy pictures. a Input
murky picture. After effects of b [8], c [5], d [4], e [25], f [40], g [42], h [48]
Dense Haze Removal Using Convolution Neural Network 553

Fig. 4 Visual consequence of proposed and existing strategies on true murky pictures. a Input
foggy picture. Consequences of b [8], c [5], d [4], e [25], f [40], g [42], h [48]

Network beats the other existing methodologies and produces an outwardly charming
dimness free scene.

5 Conclusion

In this work, we propose a start to finish generative antagonistic de-initiation network

for picture de-inception. A tale generator named as Inception-Dense Network which
is de-marked utilizing Inception block and dense associations is favorable to modeled
for picture de-initiation. To safeguard the primary data in the recuperated fog
free scene, we have used a blend of SSIM and edge misfortune while preparing
the supportive of presented network for picture de-inception. Execution of the
proposed network has been assessed on two benchmark datasets in particular: D-Hazy
[23], SOTS [37] and genuine foggy pictures. The subjective examination has been
completed by dissecting and contrasting the consequences of proposed network and
existing cutting edge techniques for picture de-right of passage. Exploratory investi-
gation shows that the proposed technique out-plays out the other existing strategies
for picture de-right of passage. Later on, this work can be reached out to examine the
impact of fog on the exhibition of various calculations for significant level PC vision
errand, for example, object location, human activity acknowledgment, and individual
re-recognizable proof. Likewise, the design of the proposed Inception-Dense module
can be reached out for other PC vision applications such single picture profundity
assessment, semantic.
554 M. Dongare and J. Kendule

References

1. Dudhane A, Murala S (2019) Cardinal color fusion network for single image haze removal.
Mach Vis Appl 30(2):231–242
2. Patil PW, Murala S (2018) Msfgnet: a novel compact end-to- end deep network for moving
object detection. IEEE Trans actions Intell Transp Syst
3. Cai B, Xu X, Jia K, Qing C, Tao D (2016) Dehazenet: an end-to-end system for single image
haze removal. IEEE Trans Image Proces 25(11):5187–5198
4. Dudhane A, Murala S (2018) IEEE, 2018, Cˆ 2msnet: A novel approach for single image haze
removal. In: 2018 IEEE winter conference on applications of computer vision (WACV), pp
1397–1404
5. Ren W, Liu S, Zhang H, Pan J, Cao X, Yang M-H (2016) Single image dehazing via multi-scale
convolutional neural networks. In: European conference on computer vision. Springer, Berlin,
pp 154–169
6. Fattal R (2008) Single image dehazing. ACM Trans Graphics (TOG), 27(3):72
7. Gibson KB, Vo DT, Nguyen TQ (2012) An investigation of dehazing effects on image and
video coding. IEEE Trans Image Proces 21(2):662–673
8. He K, Sun J, Tang X (2011) Single image haze removal using dark channel prior. IEEE Trans
Pattern Anal Mach Intell 33(12):2341–2353
9. Huang S-C, Chen B-H, Wang W-J (2014) Visibility restoration of single hazy images captured
in real-world weather conditions. IEEE Trans Circuits Syst Video Technol 24(10):1814–1824
10. Ancuti CO, Ancuti C, Hermans C, Bekaert P (2010) A fast semi-inverse approach to detect
and remove the haze from a single image. In: Asian conference on computer vision. Springer,
Berlin, pp 501–514
11. Tan RT (2008) Visibility in bad weather from a single image. In: IEEE conference on computer
vision and pattern recognition, 2008. CVPR 2008. IEEE pp 1–8
12. Tang K, Yang J, Wang J (2014) Investigating haze-relevant features in a learning framework
for image dehazing. In: Proceedings of the IEEE conference on computer vision and pattern
recognition, pp 2995–3000
13. Zhu Q, Mai J, Shao L (2014) Single image dehazing using color attenuation prior. In: 25th
British machine vision conference, BMVC
14. Schechner YY, Narasimhan SG, Nayar SK (2001) Instant dehazing of images using polariza-
tion. In: Proceedings of the 2001 IEEE computer society conference on computer vision and
pattern recognition, CVPR 2001. vol 1, pp I–I
15. Shwartz S, Namer E, Schechner YY (2006) Blind haze separation. In: 2006 IEEE computer
society conference on computer vision and pattern recognition (CVPR’06), vol 2, pp 1984–1991
16. Cozman F, Krotkov E (1997) Depth from scattering, Jun 1997. In: Proceedings of IEEE
computer society conference on computer vision and pattern recognition, pp 801–806
17. Nayar SK, Narasimhan SG (1999) Vision in bad weather. In: The proceedings of the Seventh
IEEE international conference on computer vision, vol 2. IEEE, pp 820–827
18. He K, Zhang X, Ren S, Sun J (2016) Deep residual learning for image recognition. In:
Proceedings of the IEEE conference on computer vision and pattern recognition, pp 770–778
19. He K, Sun J, Tang X (2010) Guided image filtering. In: European conference on computer
vision, Springer, Berlin, pp 1–14
20. Yu J, Xiao C, Li D (2010) Physics-based fast single image fog removal. In: 2010 IEEE 10th
international conference on signal processing (ICSP), IEEE, pp 1048–1052
21. Lai Y, Chen Y, Chiou C, Hsu C (2015) Single-image dehazing via optimal transmission map
under scene priors. IEEE Trans Circuits Syst Video Technol 25(1):1–14
22. Wang J, Lu K, Xue J, He N, Shao L (2018) Single image dehazing based on the physical model
and msrcr algorithm. IEEE Trans Circ Syst Video Technol, pp 1–1
23. Ancuti C, Ancuti CO, De Vleeschouwer C (2016) IEEE 2016, D-hazy: A dataset to eval-
uate quantitatively dehazing algorithms. In: 2016 IEEE international conference on image
processing (ICIP), pp 2226–2230
Dense Haze Removal Using Convolution Neural Network 555

24. He K, Sun J (2015) Convolutional neural networks at con- strained time cost. In: Proceedings
of the IEEE conference on computer vision and pattern recognition, pp 5353–5360
25. Li B, Peng X, Wang Z, Xu J, Feng D (2017) Aod-net: All- in-one dehazing network. In:
Proceedings of the IEEE international conference on computer vision. pp 4770–4778
26. Swami K, Das SK (2018) Candy: conditional adversarial networks based end-to-end system
for single image haze removal. In: 2018 24th international conference on pattern recognition
(ICPR). IEEE, pp 3061–3067
27. Dudhane A, Murala S (2019) IEEE, 2019, Cdnet: Single image de-hazing using unpaired
adversarial training. In: 2019 IEEE Winter Conference on Applications of Computer Vision
(WACV), pp 1147–1155
28. Engin D, Genc A, Kemal Ekenel H (2018) Cycle-dehaze: Enhanced cyclegan for single image
dehazing. In: Proceedings of the IEEE conference on computer vision and pattern recognition
workshops, pp 825–833
29. Patil P, Murala S (2019) Fggan: A cascaded unpaired learning for background estimation and
foreground segmentation. In: 2019 IEEE winter conference on applications of computer vision
(WACV), IEEE, pp 1770–1778
30. Szegedy C, Liu W, Jia Y, Sermanet P, Reed S, Anguelov D, Erhan D, Vanhoucke V, Rabinovich
A (2015) Going deeper with convolutions. In: Proceedings of the IEEE conference on computer
vision and pattern recognition, pp 1–9
31. Ulyanov D, Vedaldi A, Lempitsky V (2016) Instance normalization: the missing ingredient for
fast stylization. arXiv preprint arXiv:1607.08022
32. Isola P, Zhu J-Y, Zhou T, Efros AA (2017) Image-to-image translation with conditional adver-
sarial networks. In: 2017 IEEE conference on computer vision and pattern recognition (CVPR).
IEEE, pp 5967–5976
33. Wang Z, Bovik AC, Sheikh HR, Simoncelli EP (2004) Image quality assessment: from error
visibility to struc tural similarity. IEEE Trans Image Proces 13(4):600–612
34. Sharma G, Wu W, Dalal EN (2005) The ciede2000 color- difference formula: Implementation
notes, supplementary test data, and mathematical observations. In: Color research & appli-
cation: endorsed by inter-society color council, the colour group (Great Britain), Canadian
society for color, color science association of Japan, Dutch society for the study of color, The
Swedish colour centre foundation, colour society of Australia, Centre Franc¸ais de la Couleur
30(1):21–30
35. Yang X, Xu Z, Luo J (2018) Towards perceptual image dehazing by physics-based disentan-
glement and adversarial training. In: In Thirty third-second AAAI conference on Artificial
Intelligence (AAAI-18)
36. Zhu J-Y, Park T, Isola P, Efros AA (2017) Unpaired image- to-image translation using cycle-
consistent adversarial net- works. In: 2017 IEEE international conference on computer vision
(ICCV). IEEE, pp 2242–2251
37. Li B, Ren W, Fu D, Tao D, Feng D, Zeng W, Wang Z (2019) Benchmarking single-image
dehazing and be- yond. IEEE Trans Image Proces 28(1):492–505
38. Silberman N, Hoiem D, Kohli P, Fergus R (2012) Indoor segmentation and support inference
from rgbd images. In: European conference on computer vision. Springer, pp 746–760
39. Srivastava RK, Greff K, Schmidhuber J (2015) Highway networks. arXiv preprint arXiv:1505.
00387
40. Ren W, Ma L, Zhang J, Pan J, Cao X, Liu W, Yang M-H (2018) Gated fusion network for
single image dehazing. In: The IEEE conference on computer vision and pattern recognition
(CVPR)
41. Zhang H, Patel VM (2018) Densely connected pyramid de-hazing network. In: Proceedings of
the IEEE conference on computer vision and pattern recognition, pp 3194–3203
42. Yang D, Sun J (2018) Proximal dehaze-net: a prior learning- based deep network for single
image dehazing. In: Proceedings of the European conference on computer vision (ECCV), pp
702–717
43. Hu H-M, Guo Q, Zheng J, Wang H, Li B (2019) Single image defogging based on illumination
decomposition for visual maritime surveillance. IEEE Trans Image Process
556 M. Dongare and J. Kendule

44. Ren W, Pan J, Zhang H, Cao X, Yang M-H (2019) Single image dehazing via multi-scale
convolutional neural net- works with holistic edges. Int J Comput Vision, pp 1–20
45. Dudhane A, Murala S (2019) Ryf-net: Deep fusion network for single image haze removal.
IEEE Trans Image Proces 29:628–640
46. Dudhane A, Singh Aulakh H, Murala S (2019) Ri-gan: An end-to-end network for single
image haze removal. In: Proceedings of the IEEE conference on computer vision and pattern
recognition workshops, pp 0–0
47. Chen S, Chen Y, Qu Y, Huang J, Hong M (2019) Multi-scale adaptive dehazing network. In:
Proceedings of the IEEE conference on computer vision and pattern recognition work- shops,
pp 0–0
48. Guo T, Li X, Cherukuri V, Monga V (2019) Dense scene information estimation network for
dehazing. In: Proceedings of the IEEE Conference on computer vision and pattern recognition
workshops, pp 0–0
ICT Based Societal Technologies
Diabetic Retinopathy Detection
with Optimal Feature Selection:
An Algorithmic Analysis

S. Shafiulla Basha and Syed Jahangir Badashah

Abstract This work aims to establish a new automated Diabetic Retinopathy (DR)
recognition scheme, which involves phases such as “Preprocessing, Blood Vessel
Segmentation, Feature Extraction, and Classification”. Initially, Contrast Limited
Adaptive Histogram Equalization (CLAHE) and median filter aids in pre-processing
the image. For blood vessels segmentation, Fuzzy C Mean (FCM) thresholding is
deployed that offers improved threshold values. As the next process, feature extrac-
tion is performed, where local, morphological transformation oriented features and
Gray-Level Run-Length Matrix (GLRM) is based on extracted features. Further, the
optimal features are selected using a new FireFly Migration Operator-based Monarch
Butterfly Optimization (FM-MBO) model. Finally, Convolutional Neural Network
(CNN) is deployed for classification purposes. Moreover, to attain better accuracy,
the count of convolutional neurons of CNN is optimally elected using the proposed
FM-MBO algorithm.

Keywords Diabetic retinopathy detection · Segmentation · Feature selection ·

CNN classifier · Algorithmic analysis

1 Introduction

DR exists as a main problem of DM and it leads to blindness and visual impair-

ment, thus causing vision loss amongst the adults in working-age [1, 2]. DR can
be detected earlier by observing certain aspects such as blood vessels irregulari-
ties, leaks and so on [3, 4]. Usually, DR occurs when blood vessels are affected by
diabetes. This paves the way for blindness in patients affected by diabetes. Further-
more, enhanced screening leads to a premature diagnosis that reduces the risk of
blindness [5]. Nowadays, Computer-Aided Diagnosis (CAD) systems are introduced

S. S. Basha
Y.S.R. Engineering College of Yogi Vemana University, Proddatur 516360, India
S. J. Badashah (B)
Sreenidhi Institute of Science and Technology (Autonomous), Hyderabad, Telangana State
501301, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 559
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_56
560 S. S. Basha and S. J. Badashah

that offers automated diagnosis in a precise way. It diagnoses DR by mining the optic
disc by concerning the above-said issues in the existing systems. Also, it aids the
experts in taking appropriate decisions, thus leading to quicker and more accurate
diagnostic decisions with high reliability. The major contribution is to introduce a DR
detection system with optimization assisted CNN that intakes the optimally selected
features as input. Moreover, an algorithmic analysis that proves the betterment of
proposed work with respect to varied parameters. The paper is arranged as follows.
The reviews are presented in Sect. 2. Section 3 describes the proposed DR detec-
tion: pre-processing, segmentation and feature extraction. Section 4 explains the
optimal feature selection and optimization-based classification. Section 5 discusses
the outcomes and the paper is concluded by Sect. 6.

2 Literature Review

In 2020, Kumar et al. [6] have established a better method for detecting hemorrhages
and MA that contributed on the overall enhancement in the earlier identification of
DR. In 2020, Zago et al. [7] have sestablished a novel methodology for recognizing
DR by means of a deep patch-based model. In 2019, Liu et al. [8] have implemented a
novel technique called WP-CNN that made use of numerous weighted paths of CNN.
In 2018, Wan et al. [9] have implemented an automated technique for classifying a
specified set of fundus images.

3 Proposed DR Detection: Preprocessing, Segmentation

and Feature Extraction

3.1 Preprocessing Stage

Initially, the input image B is subjected to pre-processing which is accomplished by

CLAHE and median filter.
CLAHE [10]: The partition of B is accomplished by CLAHE via 5 phases.
Initially, the input image is split into minute blocks of a similar size. In each block,
the higher value of histogram is reduced by evaluating the clip point.

Pxi  σ 
β= 1+ Mss (1)
drr 100

Equation (1) portrays the clip point estimation, in which β represents the clip
point, Pxi indicates pixel count in every section,drr represents the dynamic range in
particular block, Mss denotes maximum slope, σ refers to clip factor. Thus, Bclahe is
achieved from the pre-processing stage by means of CLAHE.
Diabetic Retinopathy Detection with Optimal Feature Selection … 561

Median Filter [11]: The Bclahe processing in median filtering is given by Eq. (2),
in which, gr(xe, ye) and Bclahe (xe, ye) indicates the output image and original image
in order and Hm denotes the 2-D mask. Thus, Bmedian is achieved from the median
filter.

gr(xe, ye) = med {Bclahe (xe − je, ye − ke)je, ke ∈ Hm} (2)

3.2 FCM-Based Segmentation of Blood Vessel

Now, the image Bmedian is subjected under blood vessel segmentation. Usually, the
FCM [12] model partitions Me components into Xe classes. The entire pixels in
an image indicated as Me defined as Me = Mea × Meb and Xe = 3 that denotes
3-class FCM clustering. Here the optimized objective function is given by Eq. (3),
in which hele ∈ He, He = {he1 , he2 , he3 , . . . , heXe } indicates the cluster centre and
wene ∈ W e, W e = {we1 , we2 , we3 , . . . , weMe } specifies the measured data. The
criterion in  ∗ indicates similarityamongstthe cluster center and measured data
ge ∈ [1, ∞] denotes a real integer.


Xe 
Me
Ob = (eln )ge wene − hele 2 (3)
le=1 ne=1

Thus, the final segmented image Bfcm is obtained from FCM.

3.3 Feature Extraction

The segmented blood vessel Bfcm is then exposed to the feature extraction process,
anywhere local, morphological transformation and GLRM √ features are
 extracted.
√
Local Features [13]: Here, the 4 scales denoted by 2, 2, 2 2, 4 are exploited
for image filtering. In addition, a 2-D Gaussian filter is deployed for highlighting the
vessel area. The 2-D Gaussian kernel is formulated by Eq. (4).
 
1 xe2 − ye2
Geke (xe, ye) = exp − (4)
2π σ 2 2σ 2

Thus, the extracted features were indicated by Feloc .

Morphological Transformation [13]: As per this theory, vessel region is un-
illuminated and they are computed by means of “bottom-hat transformation” denoted
by Seie as given in Eqs. (5) and (6) respectively. Here ‘·’ specifies the closing function,
ie indicates the length of structuring item, Keie refers to linearized structuring item
562 S. S. Basha and S. J. Badashah

 ∈ {3, 7, 11, 15, 19, 23} and θ represents angular revolution θ ∈ MeB , MeB =
ie
xe0 ≤ xe ≤ π, xe mod π 12 = 0 .

Seieθ = Bfcm · Keieθ − Bfcm (5)


Seie = Seieθ (6)
θ∈MeB

The features extracted at this phase are denoted by Femor .

GLRM Features [14, 15]: The FeGLRM features are attained from GLRM. The
presented feature is a combination of local, morphological transformation-based and
GLRM features. Thus, the final combined feature set Fefinal is taken as Fefinal =
Feloc + Femor + FeGLRM . The framework of the adopted scheme is exposed by Fig. 1.

Retinal
Fundus image

Pre-Processing
Blood Vessel
Segmentation
CLAHE

FCM Thresholding
Median Filtering

Proposed Feature Extraction

Classification FM-MBO

Local features

Convolution
al Neurons Optimal Morphological
feature transformation
selection based features

GLRM features
CNN Proposed
FM-MBO

Affected Healthy

Fig. 1 Overall model of the proposed framework

Diabetic Retinopathy Detection with Optimal Feature Selection … 563

4 Optimal Feature Selection and Optimization Based

Classification

∗
From the extracted features, the precise or optimal features Fefinal are selected via
the optimization algorithm, and they are subjected tothe classification process. This
precise selection enhances the detection accuracy, thereby increases the system
performance.

4.1 Optimized CNN for Classification

∗
The optimal feature set Fefinal is then subjected to classification using CNN frame-
work. Usually, CNN [16] includes input and output layer and convolution layer i.e.
multiple hidden layers. Multiple layers are exploited to accomplish the overall feature
map as given by Eq. (7).
le
Ca,b,c = wicleTi ua,b
le
+ bicle (7)

le
In Eq. (17), Ca,b,c refers to convolutional feature, wicleTi specifies weight vector,
le
and bic indicates bias term of cth filter in le layer, ua,b
le
specifies the input patch at
the position (a, b) in le layer. The nonlinearities in CNN are determined using an
th
le
activation function which is given in Eq. (8), where Bia,b,c indicates activation value.

le
Bia,b,c = Bi Ca,b,c
le
(8)

The resolution of feature maps is reduced by the pooling layer denoted by pool(·)
and it is defined in Eq. (9).
 
le
Pia,b,c = pool Bii.j,c
le
, ∀(i, j) ∈ Ria,b (9)

1 
Nn
Li = li θ ; vi(ne) , tl (ne) (10)
Nn ne=1

In the above equation, Ria,b denotes local neighborhood at the position (a, b) and
each variable in CNN is denoted by θ . The optimal parameters are portrayed by
minimizing the specific loss function as revealed  in Eq. (10), where, Li is the loss of
CNN, Nn is the desired input–output relations ui(ne) , vi(ne) ; ne ∈ [1, 2, . . . , Nn] ,
ui(ne) is the neth input feature, vi(ne) specifies target label, and tl (ne) indicates output.
Two convolutional layers are exploited in CNN, where O1 and O2 denotes the count
of hidden neurons that needsto be optimized using the proposed FM-MBO algorithm.
The minimum and maximum bounding limits of O1 and O2 are 1 and 50.
564 S. S. Basha and S. J. Badashah

Fig. 2 Solution encoding

Fe1 Fe2 .... Fe128 O1 O2 J

4.2 Solution Encoding and Objective Function

A major objective of this work concerns on maximizing the accuracy of DR recog-

nition as given in Eq. (11), where, C = TP+TN TP+TN
+FP+FN
. Here “TP indicates true
positive, TN specifies true negative, FP denotes false positive, and FN specifies
false negative”.

OB = max(C) (11)

The solution encoding of this work is illustrated in Fig. 2. This includes a selection
∗
of optimal features Fefinal from the extracted features Fefinal . Among the 244 extracted
features, only 128 feature subsets are optimally selected. Further, the count of hidden
neurons O1 and O2 are optimally tuned using the algorithm.

4.3 Proposed FM-MBO Algorithm

Even though various optimization issues can be resolved by the existing MBO,
it suffers from poor performance and pre-mature convergence [17]. Therefore, to
improve its efficiency, FireFly (FF) algorithm is incorporated with the conventional
MBO model. In general, monarch butterflies are found at certain places that are
considered as land1 and land2. The migration operator is deployed for producing
new ones and after giving birth to new ones, the parent butterfly leaves the pack.
Migration Operator: Consider MPe as the total population, ce represents the
proportion of butterflies in land1 (subpopn1) and land2 (subpopn2) correspondingly.
me+1
The migration operation is specified by Eq. (12), in which, Jie,ye refers to yeth
componentof Jie and Jse1 represents the novel location of the monarch butterfly se1 .
me+1
Jie,ye = Jse1,ye
me
(12)

The current generation is denoted by me.By exploiting subpopn1, Monarch

butterfly se1 is elected randomly. In Eq. (13), the value of se is computed, where
time(t) represents migration time and rnm refers to an arbitrary value taken from a
uniform distribution.

se = rnm ∗ time(t) (13)

The component ye for the new monarch butterfly is instigated depending on the FF
updateas given in Eq. (14) [18], where, initial parameter signifies the present position
Diabetic Retinopathy Detection with Optimal Feature Selection … 565

of firefly; next parameter indicates its attractiveness to light intensity perceived by

neighboring fireflies and final parameter indicates the random motion of fireflies with
the absence of brighter ones.
   
me+1
Jie,ye = Jie,ye
me
+ η0 ∗ exp −γ gn2ieje ∗ Jie,ye
me
− Jje + α ∗ (εi ) (14)

The coefficient α indicates a randomization variable, εi denotes a random integer

taken from Gaussian distribution, γ specifies an absorption coefficient and rand
denotesthe random number that lies between [0, 1].
Butterfly Balancing Operator: In balancing operation, when rnm ≤ ce for all
me+1
components, the memory is updated as specified in Eq. (15), here, Jxe,ye specifies
ye component of Jxe for generation me + 1. Likely, Jbest,ye specify ye component
th me th

of Jbest that denotes the best monarch butterfly in land1 and land2.
me+1
Jxe,ye = Jbest,ye
me
(15)

In addition, when ce > rnm, the memory gets updated as specified in Eq. (16), in
which, Jseme3 ,ye refers to yeth component of Jse3 that is chosenfrom land2 in a random
manner. At this point, se3 ∈ {1, 2, . . . , MPe2 }
me+1
Jxe,ye = Jseme3 ,ye (16)

me+1
Jxe,ye = Jxe,ye
me+1
+ λ × fnsye − 0.5 (17)

Equation (17) reveals the memory update depending upon the constraint,rnm >
arbmv, here, arbmv refers to balancing the value of butterfly, and fns denotes monarch
butterfly xe walk steps. It is evaluated by deploying control flight as indicated in
Eqs. (18) and (19).

fns = Levy Jxeme (18)

λ = Hedim me2 (19)

The weighting element λ is computed as per Eq. (19), where Hedim denotes the
value of a single move.
566 S. S. Basha and S. J. Badashah

5 Results and Discussion

5.1 Simulation Procedure

The presentedDR diagnosis modelusing FM-MBO algorithm was implemented in

MATLAB. The dataset for analysis was downloaded from “([https://www5.cs.fau.de/
research/data/fundus-images/])”. Now,algorithmic analysis is carried out by varying
the parameters such as α and γ of FF based evaluation [18]. Moreover, the analysis
was carried out for various measures namely, “accuracy, sensitivity, precision, and
specificity, Net Present Value (NPV), Matthews Correlation Coefficient (MCC), F1-
Score, False Positive Rate (FPR), False Negative Rate (FNR), and False Discovery
Rate (FDR)”.

5.2 Performance Analysis

The performance analysis of the applied DR detection model is exposed in Figs. 3 and
4 for optimistic measures and undesirable measures respectively. From the study, the
FM-MBO system has achieved optimal performance for all the trials when matched
to the other recent schemes. Precisely, from Fig. 3a, the values of accuracy attained
by the presented approach when α = 0.2 and α = 0.5 is 3.53% and 2.35% superior
than the values achieved when α = 0.1 and α = 0.3. On considering precision, the
value of α = 0.3 is found to be higher, which is 46.77%, 3.23%, 20.97%, and 6.45%
better than the values of α at 0.1, 0.2, 0.4 and 0.5 respectively. Thus, the improvement
of the presented methodology has been confirmed via algorithmic analysis.

Fig. 3 Algorithmic analysis of the proposed model by varying the parameters α and γ of FF with
respect to measures like a Accuracy b Sensitivity c Specificity d Precision e NPV f F1-score g MCC
Diabetic Retinopathy Detection with Optimal Feature Selection … 567

Fig. 4 Algorithmic analysis of the proposed model by varying the parameters α and γ of FF with
respect to measures like a FDR b FPR c FNR

6 Conclusion

This paper introduced a advanced DR diagnostic model that includes phases like
“preprocessing, blood vessel segmentation, feature extraction, and classification”.
Initially, pre-processing is done for which CLAHE and median filter were deployed.
Next to pre-processing, blood vessel segmentation was performed using FCM thresh-
olding model. Subsequently, certain features were extracted from the segmented
images. In addition, optimal feature selection was carried out by means of suggested
FM-MBO algorithm, and more over, classification was done by optimizing the convo-
lutional neurons in CNN using the same algorithm.Lastly, the algorithmic study was
held for verifying the superiority of the adopted model. Particularly, on observing the
analysis on precision, the value of α = 0.3 is found to be higher, which is 46.77%,
3.23%, 20.97% and 6.45% better than the values of α at 0.1, 0.2, 0.4 and 0.5 respec-
tively. Thus, the enhancement of the adopted FM-MBO algorithm was validated
effectively.

References

1. Wan S, Liang Y, Zhang Y (2018) Deep convolutional neural networks for diabetic retinopathy
detection by image classification. Comput Electr Eng 72:274–282
2. Leontidis G (2017) A new unified framework for the early detection of the progression to
diabeticretinopathy from fundus images. Comput Biol Med 90:98–115
3. Arias MEG, Marin D, Ponte B, Alvarez F, Bravo JM (2017) A tool for automated diabetic
retinopathy pre-screening based on retinal image computer analysis. Comput Biol Med 88:100–
109
4. Romero RR, Carballido JM, Capistrán JH, Valencia LJU (2015) A method to assist in the diag-
nosis of early diabetic retinopathy: Image processing applied to detection of microaneurysms
in fundus images. Comput Med Imaging Graph 44:41–53
5. Gupta G, Kulasekaran S, Ram K, Joshi N, Gandhi R (2017) Local characterization of neovas-
cularization and identification of proliferative diabetic retinopathy in retinal fundus images”.
Comput Med Imaging Graph 55:124–132
6. Shanthi T, Sabeenian RS (2019) Modified Alexnet architecture for classification of diabetic
retinopathy images. Comput Electr Eng 76:56–64
568 S. S. Basha and S. J. Badashah

7. Kumar S, Adarsh A, Kumar B, Singh AK (2020) An automated early diabetic retinopathy

detection through improved blood vessel and optic disc segmentation. Opt Laser Technol 121,
Article 105815
8. Zago GT, Andreão RV, Dorizzi B, Salles EOT (2020) Diabetic retinopathy detection using red
lesion localization and convolutional neural networks, Comput Biol Med 116, Article 103537
9. Liu Y-P, Li Z, Xu C, Li J, Liang R (2019) Referable diabetic retinopathy identification from
eye fundus images with weighted path for convolutional neural network. Artif Intell Med 99,
2019, Article 101694
10. Chang Y, Jung C, Ke P, Song H, Hwang J (2018) Automatic contrast-limited adaptive histogram
equalization with dual gamma correction. IEEE Access 6:11782–11792
11. Zhu Y, Huang C (2012) An improved median filtering algorithm for image noise reduction.
Phys Procedia 25:609–616
12. Masood A, Jumaily AAA, Hoshyar AN, Masood O (2013) Automated segmentation of
skin lesions: Modified Fuzzy C mean thresholding based level set method. INMIC, Lahore
2013:201–206
13. Zhu C, Zou B, Zhao R, Cui J, Duan X, Chen Z, Liang Y (2017) Retinal vessel segmentation in
colour fundus images using extreme learning machine. Comput Med Imaging Graph 55:68–77
14. Radhakrishnan M, Kuttiannan T (2012) Comparative analysis of feature extraction methods
for the classification of prostate cancer from TRUS medical images. IJCSI Int J Comput Sci
9(2)
15. Venkat Narayana Rao T, Govardhan A, Badashah SJ (2010) Improved lossless embedding and
extraction—A data hiding mechanism. Int J Comput Sci Inf Technol 2:75–86
16. Gu J, Wang Z, Kuen J, Ma L, Shahroudy A, Shuai B, Liu T, Wang X, Wang G, Cai J, Chen T
(2018) Recent advances in convolutional neural networks. Pattern Recogn 77:354–377
17. Wang S, Deb G-G, Cui Z (2015) Monarch butterfly optimization. Neural Comput Appl, 1–20
18. Gandomia XS, Yang S, Talatahari, Alavi AH (2013) Firefly algorithm with chaos. Commun
Nonlinear Sci Numer Simulat 18:89–98
Students Perception and Satisfaction
Towards ICT Enabled Virtual Learning

Moshina Rahamat and B. Lavanya

Abstract According to the Ministry of Education, 2010 “The Education service is

that which moulds young generation into good citizens, who become conscious with
their responsibilities towards the family, society and country. The present COVID 19
Pandemic has largely affected the lives of scholars round the world. They missed the
possibility to interact one on one basis with their teachers as all the governments
temporarily closed the academic Institutions. Due to this, the importance for tech-
nology based learning has become more prominent which resulted in ICT enabled
learning called Virtual learning. It particularly becomes important to grasp how this
ICT enabled learning helps the scholars, who lay strong foundation for better Society,
therefore the present research aimed to look at the student’s perception towards virtual
class room learning, their satisfaction, besides determining the challenges faced by
virtual learning. The survey method is employed for data collection. The scope of
the study involves only post graduate students from Hyderabad city. The information
collected from a convenient sample size of 142 students is analyzed using SPSS 26.
Descriptive statistics, correlation, ANOVA & T-test are used to draw the inferences
because of this pandemic situation it was found that students have positive percep-
tion towards virtual learning. When measured satisfaction there are few issues that
require to be addressed. So based on the findings, a model has been proposed which
include stakeholders of virtual learning and their prime responsibility through which
ICT enabled learning ecosystem may be made a beautiful learning place.

Keywords ICT · Virtual learning · Responsible citizen · Perception and

satisfaction · Benefits and drawbacks

M. Rahamat (B) · B. Lavanya

School of Management Studies, CBIT, Hyderabad, Telangana, India
B. Lavanya
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 569
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_57
570 M. Rahamat and B. Lavanya

1 Introduction

Information and Communication Technologies (ICT) refers to technologies that

supply access to details through telecommunications it’s almost like Information
Technology (IT), but centre of attention is totally on communication technologies.
This includes the web, wireless networks, cell phones, and other communication
mediums. Telecommunication has been the key aspect to drive the concept of virtual-
learning. Virtual-learning is use of technology to bring solutions that increases knowl-
edge and performance. Usually, these interactions happen through videoconferencing
where it permits live interaction between the tutor and also the learners as they’re
engage in learning activities similar to those kind of standard classroom.
A virtual classroom often includes the following features:
• Videoconferencing that facilitates communication with both the teacher and with
peers.
• A digital whiteboard to offer real-time explanations and/or collaboration.
• Instant messaging for low-bandwidth communication.
• Participation controls so that students can still “raise their hands” or otherwise
participate in lessons.
• Sub-chats or group chats for students to collaborate in small groups online
(sometimes also called breakout rooms).

2 Review of Literature

Through review of the existing literature enables a researcher to gain knowledge on

the topic been selected for research Following are the few papers reviewed to gain
insights about the research that has been carried on virtual learning.
The research suggested that members of the educational management staff have
limited knowledge of institutional strategies; there’s a spot between what academic
managers believe they are doing to support and implement TEL and what other
academic staff perceive them to truly do and Technology Enhanced Learning (TEL)
is seldom discussed during performance assessment review additionally, the study
reveals that academic managers have different understandings of the employment of
educational technology [1].
It is too early to see how COVID-19 will affect the adoption of digital learning,
it is clear that the wheels of change have been set in motion. It is time to prioritize
learning and perhaps, the one time in history when the L&D narrative will reflect a
positive outcome from calamity [2].
It was spotlighted “How virtual classroom are becoming the new normal” and
how it has affected during COVID-19, it has it has opened doors to a new way of
delivering education. The perfect virtual classroom experience for their students can
be done through two key learning experience: synchronous learning which includes
online learning platform like zoom, Skype, micro teams, Google teams etc., and
Students Perception and Satisfaction Towards ICT … 571

Asynchronous learning which includes learning management system like google

classroom with simple interface [3].
It was investigated whether the utilization of virtual, interactive, real-time,
instructor-led (VIRI) online learning can deliver the identical student performance
and engagement outcomes as a face-to-face (F2F) course. The findings show that
a synchronous course delivered using VIRI classroom technology has the identical
level of student performance outcomes as F2F learning. This study suggests that
VIRI technology is a good synchronous learning environment [4].
From a study conducted on “Factors Affecting the Acceptance of E-learning By
Students: A Study of E-learning Programs in Gwalior, India”. Here they identify
the factors that affect the perception of the students towards E-Learning and its
acceptability. The result showed that acceptance of E-learning is governed by four
critical dimensions, which are, E-learner Competency, External Influence, System
Interactivity, and Social Influence. In case of educational qualification categories, it
was found that PG students are more enthusiastic to accept E-learning as compared
to other two categories [5].
The research findings illustrate that e-learning system consists of 4 compo-
nents, mainly; the trainer, learner, course, and data and communication technolo-
gies (ICT), additionally to the context determinants of e-learning system success
and also proposed a model for e-learning success, which includes eight factors,
mainly; e-learning context that include individual, institutional, and environmental
determinants to e-learning success [6].
The study provided practical suggestions for those that are reaching to develop
online courses in order that they’ll make informed decisions within the implementa-
tion process. Supported the findings, the authors argued that effective online instruc-
tion depends upon (1) well-designed course content, motivated interaction between
the trainer and learners, well-prepared and fully-supported instructors; (2) creation
of a way of online learning community; and (3) rapid advancement of technology
[7].
The role of ICT for promoting environmental sustainability in a changing, society,
authors propose Ecological World Systems Theory (WST) as a holistic framework
to assess the environmental impact of ICT. The work is usually supported existing
empirical studies, which could be a limitation. This theoretical framework implies
that unequal environmental degradation in several parts of the globe should be taken
into consideration when assessing environmental impact [8].
The research explored the employment of data and communication technolo-
gies (ICT) in teaching. A theoretical framework has been developed that visual-
izes the spatial implications of developments in education. The findings show that
the normal classroom space is progressively being replaced by a range of learning
settings to support contemporary learning activities. The research findings contribute
to a stronger understanding of the alignment of learning space to the evolving needs
that come from new ways of learning, supported by advanced ICT, and may be
accustomed support space planning in teaching [9].
572 M. Rahamat and B. Lavanya

3 Research Methodology

The current study has been carried out with the following objectives:
1. To examine the students perception towards Virtual Learning.
2. To explore the satisfaction levels of contemporary students over the use of
Virtual Learning system.
3. To understand the challenges faced by the students, propose a model for
enhancing virtual class room learning effective.
This study is an empirical and exploratory in nature. Primary data has been
collected with the help of a self—designed structured questionnaire. Secondary data
is collected from various journals, articles, websites and online resources. Conve-
nient sampling method was used to collect data. The Virtual Learning variables
(Perception, Satisfaction, Drawbacks) were measured using 5-point Likert scale.

4 Data Analysis and Interpretation

The data collected were subjected to reliability analysis, descriptive Statistics, T-Test
and ANOVA and Correlation has been applied using SPSS Version 26.

5 Results and Discussions

Table 1 shows Cronbach alpha coefficients (reliability) of the measures of virtual

class room learning. The reliability of the six aspects of virtual learning that varied
from 0.79 to 0.95 was found to be excellent. This indicates a high level of internal
consistency for scale with this specific sample.
From Table 2, it is observed that most of the students have constructive perception
towards virtual learning as subject taught is relevant and understandable which helps
them to achieve their objectives.
From Table 3, it is interpreted that students are just satisfied with virtual learning
as there is no complete interaction with their teachers to understand and clear their
doubts.

Table 1 Cronbach alpha

Factor Item Cronbach alpha
values
Perception 5 0.88
Satisfaction 9 0.95
Drawbacks 8 0.79
Strategies 7 0.93
Students Perception and Satisfaction Towards ICT … 573

Table 2 Descriptive statistics of perception of students towards virtual learning

Statement Mean Std. deviation
Class: boringn stimulating 3.32 1.14
Learning: useless useful 3.16 1.17
Content: irrelevant relevant 3.78 1.11
Objectives: not achieved achieved 3.42 1.13
Subject: not understood understood 3.43 1.06

Table 3 Descriptive statistics of student’s satisfaction towards virtual learning

Parameters N Mean Std. deviation
The instructor communication with students 142 3.33 1.07
The usage of online platform used by the institution for giving the 142 3.56 1.09
lectures
The content provided during virtual learning is effective 142 3.43 1.10
How satisfied are you with the material provided through virtual 142 3.38 1.12
learning
The understanding of the subject through virtual learning 142 3.48 1.07
The presentation given by the instructor during virtual learning 142 3.35 1.09
The interaction between instructor and students during virtual class 142 3.40 1.05
learning
The case studies/examples explained to you while teaching through 142 3.37 1.02
virtual class
Clearing of doubts with the instructor during virtual learning 142 3.49 1.07

Table 4 Descriptive Statistics for the statement about the drawbacks of virtual learning
Statements N Mean Std. deviation
I think I get socially isolated while using virtual learning 142 3.47 1.22
Virtual class is difficult to handle and therefore frustrated to use 142 3.39 1.22
Virtual learning can damage our health as we use laptop for long 142 3.65 1.31
time
Internet connectivity is the major problem I face using virtual 142 3.83 1.24
learning
I do face technical problem when I use virtual learning to learn 142 3.79 1.17
Do you think virtual class provide the right amount of theoretical 142 2.57 1.16
and practical experience
I prefer to learn from books rather than from the virtual 142 3.59 1.20
learning/course learning
I do not have computer and therefore I find it difficult to learn 142 3.04 1.29
through virtually
574 M. Rahamat and B. Lavanya

From Table 4, it is seen that most of the students neutrally agree virtual learning
is a drawback because they feel isolated, due to the technical problems like internet
connectivity, sitting foe long hours in front of laptop make them difficult to handle
and feel that virtual leaning is a disadvantage as it provides low quality experience
than the practical one.
From Table 5, it can be said that students agree to all statements and strongly
believe that drawing attention to core content, attracting and gaining the students
attention and taking continuous feedback makes virtual learning effective.
Table 6 shows the correlation between the perception and satisfaction of students
towards virtual leaning. It can be seen that Pearson Correlation value 0.827 which
indicates that there exist significant and positive relationship between students
perception and satisfaction with regard to virtual learning. This stipulates that the
student’s are satisfied with the way they perceive in connection with virtual learning.
From Table 7, assuming equal variance the significance (2 tailed) value is greater
than 0.05. Hence we reject H0 and accept H1. Therefore there is a significance
difference male and female PG student towards satisfaction of virtual learning.

Table 5 Descriptive statistics of strategies used for making virtual learning effective
Statement N Mean Std. deviation
Developing a detailed outline and research core content 142 4.21 1.04
beforehand
Focus on friendly content delivery 142 4.38 1.14
Ask engaging and thought-provoking questions that will drive the 142 4.28 1.13
discussion
Draw attention to important content and offer periodic recaps 142 4.36 1.21
Keep it short and simple to avoid information overload 142 4.35 1.16
Monitor the students progress about the understanding of subjects 142 4.27 1.13
through virtual learning
Ask for continuous feedback to improve virtual learning 142 4.37 1.16

Table 6 Correlation between

Correlations
students perception and
satisfaction regarding virtual Satisfaction Perception
learning Satisfaction Pearson correlation 1 0.827**
Sig. (2-tailed) 0.000
N 140 140
Perception Pearson correlation 0.827** 1
Sig. (2-tailed) 0.000
N 140 140
** Correlation is significant at the 0.01 level (2-tailed)
Table 7 Gender of students with satisfaction level towards virtual learning
Independent samples test
Levene’s test for t-test for equality of means
equality of variances
F Sig t df Sig. (2-tailed) Mean Std. error 95% confidence interval of
difference difference the difference
Lower Upper
Satisfaction Equal 0.448 0.504 0.064 140 0.949 0.00992 0.15556 −0.29763 0.31747
Students Perception and Satisfaction Towards ICT …

variances
assumed
Equal 0.064 139.188 0.949 0.00992 0.15533 −0.29719 0.31703
variances not
assumed
H0: There is a no significance difference in satisfaction between male and female students towards virtual learning
H1: There is significance difference in satisfaction between male and female students towards virtual learning
575
576 M. Rahamat and B. Lavanya

Table 8 Age of students with perception of virtual learning

ANOVA
Perception
Sum of squares df Mean square F Sig
Between groups 7.369 3 2.456 2.923 0.036
Within groups 115.980 138 0.840
Total 123.349 141
H0: There is a no significance difference in perception between different age groups towards virtual
learning
H1: There is significance difference in perception between different age group students towards
virtual learning

From Table 8, the significance value (0.036) is less than 0.05. Hence we accept
H0 and reject H1. Therefore there is no significance difference in perception between
different age groups towards virtual learning.

6 Findings of Study

From the study it was found that most of the students have positive perception and are
not completely satisfied with virtual learning as it is new way of learning due to this
pandemic situation, from the analysis it was also found that students neutrally agree
that virtual learning makes them isolated from their friends and technical issue is
the major disadvantage they face while using virtual learning. The analysis disclose
that there exist positive correlation between perception and satisfaction of students
towards virtual learning From the study it was also found that there is difference
in the way male and female students perceive virtual learning but irrespective of
different age groups there is no difference in perception of students with regard to
virtual learning.

7 Model for Effective Virtual Learning

Based on the strategies and suggestions of the students a model is proposed for
making virtual learning effective (Fig. 1).
Institution:
• Institution should frame flexible rules and regulations for virtual classes.
• Institution should either develop or choose good technology platform for
education.
• Should develop user friendly to LMS (easily understandable).
Students Perception and Satisfaction Towards ICT … 577

•INSTITUTION •FACULTY

Upgrading
Flexible Rules skills that
and virtual classes
regulations demand

Expediting
Active
the ward’s
Participation
learning
•STUDENT •PARENT

Fig. 1 Model for effective virtual learning

• Institution should provide access to all the forms of resources to students for
learning.

Faculty:
• Faculty should be trained to conduct virtual learning.
• Special/extra efforts need to be put by the faculty for creating a real class room
experience and for problematic/Quantitative subjects.
• Faculty should provide clear and concise instructions and exercises to follow.
• Faculty should monitor discussions to clarify students’ postings, highlight good or
interesting comments and ideas, provide insight, and ensure every voice is heard.
• Faculty should provide the necessary components of successful interaction:
Lesson and Session plans, Learning material, animated PPTs, Content related
videos, demonstration, practice sessions, feedback, and assessment.

Students:
• Students should be active learner as well as active listener.
• Begin with a flexible schedule.
• They should be participative in all activities.
• Should be responsible in completing their task.
• Students should be motivated and committed to learn and work on their own.
• Students should give constructive feedback for effective improvement in learning.
Teachers should consider the feedback given by the students and take the corrective
actions for making the virtual class in an effective manner. In the same way faculty
should also suggest the institution in using new tools to make the students familiar
to work with virtual learning.
578 M. Rahamat and B. Lavanya

Parents:
• Should monitor their wards participation in the virtual classes.
• Should stay in touch with the faculty and to know the progress.
• Should support their ward and help them to stay physically, mentally and
emotionally strong.

8 Conclusion

Shutdown of the institutions due to covid-19 pandemic has resulted in the greater
use of ICT. Online education is not so easy as speaking into the microphone at one
end, and connecting a laptop and listening on the other; there are challenges faced
at both ends of the spectrum. It is true that ICT is playing a vital role in teaching
and learning but at same time there are many challenges that are to be addressed
for smooth functioning of various services provided through ICT. So, in the current
research an attempt has been made to study the perception and satisfaction levels of
students towards virtual learning and found that there exist a positive relationship
between the two perception and satisfaction. During this pandemic situation, it is very
important that students have positive perception with respect to institution, faculty,
and parents. It is the responsibility of teacher to engage the students during the class
to make it interesting as well as student themselves to have a cooperative attitude to
learn from virtual platform. Positive the perception higher the satisfaction level of
students towards virtual learning. Strengthening teacher-student interaction, utilize
student’s feedback, making more creative and useful learning makes virtual learning
effective. It can be concluded that, during this hard and tough time ICT stood as a
boon to students all around the world.

References

1. Habib L, Johannesen M (2020) The role of academic management in implementing technology-

enhanced learning in higher education. Technol Pedagogy Educ 29(2):129–146, p 18
2. https://elearningindustry.com/covid-19-disrupting-online-learning.
3. Disha K (2020) How virtual classroom are becoming the new normal. Times of India, June 2020
4. Federman JE (2019) Interruptions in online training and their effects on learning. Euro J Train
Dev v43(5–6):490–504
5. Kaurav RPS, Rajput S, Baber R (2019) Factors Affecting the acceptance of e-learning by
students: a study of e-learning programs in Gwalior, India 26(1):76–95, p 20
6. Romi IM (2017) A model for e-learning systems success: systems, determinants and perfor-
mance. Int J Emerg Technol Learn 12(10)
7. Sun A, Chen X (2016) Online education and its effective practice: a research review. J Inf Technol
Educ Res 15:157–190
8. Lennerfors TT, Fors P, van Rooijen J (2015) ICT and environmental sustainability in a changing
society. Inf Technol People 28(4):758–774, p 17
9. Beckers R, van der Voordt T, Dewulf G (2015) A conceptual framework to identify spatial
implications of new ways of learning in higher education. Facilities 33(1/2):2–19, p 18
An Appointment Scheduler: A Solution
to Control Covid-19 Spread

Apeksha M. Gopale

Abstract A system based on android application is proposed in this paper to

manage appointment slots for businesses like Automobile Service Center or Beauty
Parlours/Saloons etc. in order to manage spread through social gathering. The
proposed system has features such as a single app for service providers and customers.
Verification of genuine customers and service providers will be done through OTP and
location based photographs of shop respectively. Owner can maintain details of their
staff through registration and can keep track on regular customers. It facilitates easy
booking and cancellation of appointments. Customer can view non-working days
through an event calendar and services offered with their respective charges, time
required etc. The system also provides customer payment handling option, genera-
tion of invoices, reports for analysis helps in maintaining a database and provides
appointment reminder to customer. Thus the system will do proper scheduling and
reduce efforts and time of customer and owner both.

Keywords Covid-19 control · Online booking · Appointment · Services · Android

application

1 Introduction

Appointment Management for businesses is still an emerging area in India though

it is widely spread in the foreign developed countries [1]. It is the need of an hour
in this pandemic situation to manage business in well-organized manner in order to
maintain hygiene and safety for all.
Being a salon client or automobile owner today is inconvenient. One have to
remember to make an appointment through call or physical visit, then need to
remember during business hours, and finally take time out of busy day to make to
the shop. In traditional system customer visits and wait in Salons, Spa centers beauty
centers and automobile service centers which leads to long waiting hours of crowd in

A. M. Gopale (B)
SNDT Women’s University, Mumbai, India
AIKTC, Mumbai University, Mumbai, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 579
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_58
580 A. M. Gopale

small spaces thus increases chances of Covid-19 spread. During pandemic the main-
tenance of personal vehicle became need of an hour due to increased usage.This is
the experience of countless salon-goers and automobile owners.
Thus a system based on android application is being proposed in this paper to
manage appointment slots in order to manage spread through social gathering. The
proposed system focuses on proper scheduling of appointments such that minimum
number of customers who are under service will be present in the service centers and
will give time to service providers to sanitize the area before the next set of customers
visit.
The app connects clients to service providers, putting all the information 24 * 7,
in one convenient place and on demand as well. The app eases the management of
staff, business growth and track of customers for service providers.

2 Literature Review

Qaffas and Barker [2] proposed a web based appointment management system for
students to take appointment of their advisers. This is web based method developed
using ASP.NET 2008. Nimbekar et al. [3] proposed an idea of an android application
for salon appointment but there work do not include any results.
Sherly et al. proposed web based appointment scheduling for health-care [4]
by taking a case study. Majority of the appointment management applications are
web based and in the field of healthcare [6, 7]. Malik et al. developed an android
application based appointment schdeular for patients [5].
Chai and Wen proposed salon management system based on PHP and MySQL
[8]. Xandaro Scheduling Software [9] focuses on the appointment function. This
software is easy to handle. It’s user interface is simple and direct which is helpful for
naive users. Unique Salon Software [10] is complex software to use and one need to
read the user manual. This homepage is unorganized. It uses search function in all
the modules. It simplifies the retrieval of specific information. So, users can obtain
the information faster and easily. Advantage Salon software [11] is organized and
neat which helps the user to use it easily.
Sale et al. proposed [16] web based solutions for automobile service management.
Considering the need of today more than a web based solution [12–15, 17–19], an
app based solution is any time preferable by users. Thus in the proposed system the
idea is considered keeping needs of users in mind.

3 Proposed System

The proposed system is very efficient and easy to handle for owner and client both.The
speed and accuracy will be maintained in proper way. It has features such as facility
to register staff and regular clients along with their personal information, appoint-
ment handling, view non-working days through an event calendar, handling various
An Appointment Scheduler: A Solution to Control Covid-19 Spread 581

services along with their respective prices, time required etc., providing customer
payment handling option, system generated invoices, generating reports for analysis,
maintaining database, reminders through emails, giving staff rating facility to know
the most efficient staff members.

3.1 System Requirements

This application is built in “Android” as a front end and “Firebase” as a back end.
The required system specification are as follows:
Android Smartphone: version 4.0 (or Above), RAM: 512 MB, Device Memory:
50 MB (or above), Internet Connectivity.

3.2 System Architecture

Figure 1 shows the system architecture for shop owner. Owners can register them-
selves by providing personal and shop information along with GPS enabled photo
which can help to verify genuine service providers. The service owner will provide
details of the various services along with their charges and offers available. Shop

Fig.1 Owner view

582 A. M. Gopale

Fig. 2 User view

adder module can fetch the reviews available online about the shop or staff. Appoint-
ment manager keeps track of available and booked slots, generates report on confir-
mation of a booked slot. Notification engine gives confirmation of booked slot to the
owner of the shop.
Figure 2 shows the system architecture for customer. Clients can create their profile
by providing personal information and genuine customers are verified through OTP.
Clients can check previous ratings of shop or staff available, can check list of services
and charges of specific shop and can book the available slot. On confirm booking they
will get notification though email. All the details of shop, owner, staff, appointments
and client are saved in database.

3.3 System Modules

Account Registration: Registration of user is done which makes it easier for the
user to book appointment. As they don’t need to fill personal details from scratch
every time, they want to make an appointment.
Get Location: User will know where they are located, showing the location of
the shop of their locality.
Customized Account: User can change it personal details according to his
convenience.
An Appointment Scheduler: A Solution to Control Covid-19 Spread 583

Email/Push Notification: This module makes use of email and enable shop owner
to update their customer. The shop owner can let their customer know about any
promotional offer run by them.
Marketing and Promotion: Marketing and other means to promote your products
and services is very convenient and easy to manage.
User Record: Every point of fail transaction is track and tied to each user service
history, providing quick rebooking. No restrictions on uses and user unlimited access
to any number user, staff and devices, to use if to the fullest.
Billing: This module is responsible for the generation of customers billing and
invoice.
Feedback: Collect real-time user feedback which inside to improve overall
experience.

4 Results

The proposed system is implemented as an android app for a Saloon Client and the
screenshots of various activities are as follows.
Main screen shows two options for the user either to register as an Owner or a
Customer for very first time (Fig. 3a).
If user selects Owner as an option he/she need to register as an Owner with personal
and shop information including Geo tagged photo of the shop then on verification
the registration will be confirmed (Fig. 3b–d). On successful verification the Owner
will get log in screen in the next attempt. Owner can add staff details, rates of various
services, can see booked appointments, can change profile photo (Fig. 3e–h).
If user selects Customer as an option on main screen then the user is directed to user
registration and it will be verified through OTP. On Completion of verification user
can log in.User can see various shops providing the same services, can see calender,
time slots, services provided by shops, book or cancel appointments (Fig. 4a–h).

5 Conclusion

It is the need in this pandemic situation to manage business in well-organized manner

in order to maintain hygiene and safety for all.Therefore providing a better service
for the customers at their arrival to the shop premise, will help to make a good
impression on their mind over the shop. However at rush hours, it is very difficult to
manage the crowd, hygiene, social distance and thus the spread of Covid-19 and it is
the concern of health of staff and clients both. Therefore the concept which turned
into an appointment scheduler application at the end of this project will be a great
assistance for the shop owners and clients. Thus the proposed system will help to
reduce social spread of Covid-19 and will help to reduce efforts and time of customer
and owner both.
584 A. M. Gopale

Fig. 3 a Main screen, b Owner registration, c Details for registration, d Message after registration,
e Owner log in, f Staff details, g Services with charges, h Profile and cover photo settings

6 Future Scope

The proposed system is successfully implemented.However there is scope to enhance

the features of the system such as reminder through SMS, create attractive member-
ship schemes, refer and earn sms notification, augmented reality to check best suited
style, Multilingual support through app.
An Appointment Scheduler: A Solution to Control Covid-19 Spread 585

Fig. 4 a Customer registration, b Customer log in, c List of service providers, d Selection of date,
e Selection of slot, f Staff selection and services with charges, g Summary of booking, h Booking
confirmation message

References

1. Salon management system, https://salonmanagementsysten.blogspot.com/2011/12/salon-man

agement-system-for-coiff-it-up.html
2. Qaffas A, Barker T (2015) Online appointment management system
3. Nimbekar P, Raut N, Wankhede B, Kejalkar N, Bagde A (2018) Online appointment booking
android application. Int J Recent Eng Res Dev
586 A. M. Gopale

4. Sherly IS, Mahalakshmi A, Menaka D, Sujatha R (2016) Online appointment reservation and
scheduling for healthcare-a detailed study. J Int J Innov Res Comput Commun Eng
5. Malik S, Bibi N, Khan S, Sultana R, Rauf SA (2017) Mr. Doc: a doctor appointment application
system
6. Symey Y, Sankaranarayanan S, binti Sait SN (2013) Application of smart technologies for
mobile patient appointment system. J Int J Adv Trends Comput Sci Eng
7. Choudhari SB, Kusurkar C, Sonje R, Mahajan P, Vaz J (2014) Android application for doctor’s
appointment. Int J Innov Res Comput Commun Eng
8. Chai A, Wen CC (2017) SHEARS Inc. Salon management system. J Int J Inform Vis
9. Xandaro: Free Salon Software (2017) URL: https://xandaro.com/
10. Unique Salon Software (2017) URL: https://uniquesalonsoftware.com
11. Advantage Salon Software. URL: https://www.aknaf.com/salon
12. B. Plan:Pretty (2011) Quick business plan beauty booking made simple. Confidential
Proprietary Bus Plan 1–52 29
13. Salon Today (2018) https://www.salontoday.com/article/81571/tech-pilots-voga-salon-and-
salon-clouds
14. M. S. International, The Bussiness Side of Beauty, Millennium Systems International. https://
www.millenniumsi.com/blog/65-salon-marketing-ideas (2018)
15. SBDCNe, Small Business Development Center: SBDCNet News Magazine Theme built.
https://www.sbdcnet.org/small-business-research-reports/beauty-salon-2014.
16. Sale HB, Bari D, Dalvi T, Pandey Y (2018) Online management system for automobile services.
J Int J Eng Sci Comput
17. Chavan S (2014) Automobile service center management system. J Int J Sci Res Publ 4(3).
ISSN 22503153
18. Shivasankaran N, Senthilkumar P, Scheduling of mechanics in automobile repair shops. J Indian
J Comput Sci Eng (IJCSE)
19. Selokar N, Masne V, Pimpalkar R, Puranik S, Bhoyar N, 24*7 vehicle management systems
for automobile industry. e-ISSN: 2395-0056
Bandobast Allocation and Attendance
System

Prashant S. Bhandare, Somnath A. Zambare, Amey Bhatlavande,

and Shamsundar Bhimade

Abstract This System (Website and App) is aimed at developing a Bandobast Allo-
cation and Attendance System (BAAS) that is of importance to Pandharpur Police
for various Bandobast like VIP Bandobast, Election Bandobast, and mostly for Wari
Bandobast. This system is used for the management of any Bandobast in Pandharpur
(or in any City). This system is being developed for CO’s to allocate duties to all
police, and further heads can take attendance of their subordinates. Later the control
room can print the entire report in just one click. During any Bandobast it is very
hard to allocate duties manually (informing each one about it) and also it may not
be possible to take attendance manually of more than 5000 police staff. Besides, it
is very hard to change the duty allocation or to know whether everyone is at their
allocated point.

Keywords QR code · Attendance · Controller

1 Introduction

The title of the project is “Bandobast Allocation and Attendance System” is a Hybrid
application (Website + Android APP) that aims at providing easier management of
any type Bandobast like VIP Bandobast or Wari Bandobast. Surely, many times there
are various types of Police Bandobast (like Wari in Pandharpur) and City. The police
station has to manage it properly and has to submit a report of it. But how can the
Police Station manage the Bandobast [1] with less manual work if they are still using

P. S. Bhandare (B) · S. A. Zambare · A. Bhatlavande · S. Bhimade

SVERI’s College of Engineering (Polytechnic), Pandharpur, Maharashtra, India
e-mail: [email protected]
S. A. Zambare
e-mail: [email protected]
A. Bhatlavande
e-mail: [email protected]
S. Bhimade
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 587
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_59
588 P. S. Bhandare et al.

the traditional (manual) method for this? In the traditional method, before two days
of Bandobast City Police Station used to know about list of details of Police staff
coming for the bandobast. Then, they were allocating the police to the Points (All
things were typed manually). On a working day, all Police members used to come to
the Police station to know their duty details. There was no facility for attendance or
checking the allocation. The most difficult task was to replace or remove or to change
a complete point of a member during bandobast. Thus, there were more chances of
human errors, and sometimes due to lack of attendance, system members may leave
their point before the end of their shift. Nowadays, technology has changed many
aspects of life, and people’s daily life is becoming indivisible from the network due
to the development of the Internet. With our hybrid system, the process gets much
faster and more efficient than the traditional way. During any Bandobast it is very
hard to allocate duties manually (informing each one about it) and also it may not
be possible to take attendance manually of more than 5000 police staff. Besides, it
is very hard to change the duty allocation or to know whether everyone is at their
allocated point. After all these CO’s were typing the entire report (minimum 400
pages) of bandobast, which was the most tiresome and time-consuming task. Our
system makes this easier.
Even though the Pandharpur police tried to make it online, but they can’t because
the only single system is not enough to solve the issue and they are facing several
problems during process management. There is also a problem when any police
staff doesn’t know his/her point of duty they cannot check their duty details online.
Police stations provide proper Smart I-cards and instruction on a paper the paper
or Smart I-card may be lost. BAAS system helps any police member to know their
duty details at any given time, and CO’s can easily administrate the bandobast,
record maintenance becomes easy, report generation is easy, I-card generation is
possible. This system also provides a feature of attendance along with the real-time
and location of the user, which keeps all police at their point, and any critical situation
can be managed easily. With all this, there are more advantages like the attendance
without any intervention can even be taken without having an internet connection and
automatically get uploaded to the server when the connection is available. The initial
task of CO’s to create and allocate a bandobast becomes much easier and efficient
with the system. Earlier this was taking more than four days to manage the simplest
bandobast, but now our system reduced this time to only one day or half-day and
also this can be done with less manpower.
The BAAS System is developed to implement the whole process in the City Police
Station (Pandharpur) starting with the registration of all police staff coming for the
bandobast. The registration includes details like Buckle Number, Full name, Name of
posting, Rank, Aadhar number, Duty type. Then COs can allocate duties to them. The
allocated person can check their duty details, team, head name, instruction, Smart
I-card in the App. Then all print heads can take attendance of respective subordinates
by scanning QR codes on their Smart Ecards. Bandobast heads have the authority to
take attendance of anyone who is part of the bandobast. During bandobast, COs can
change the allocation of any point or person.
Bandobast Allocation and Attendance System 589

The most important and advanced concept used in the system is attendance [2].
It is because we have used QR code for attendance (never used before for any Police
Station or any organization in India, only some Maleshian school uses this concept to
take attendance of students) [3]. Every police have provided the Smart I-card having
a QR Code which uniquely identifies the police officer [4]. Whenever the respective
head scan that QR code with ‘BAAS App’ then and then only he is marked as present.
Besides increasing the accuracy in the presence of a member at the required point,
we also record his/her location [5] during scanning. If any member is absent, then
the respective head marks him/her as absent with some remark. One more important
part is, the attendance can be marked even though the user is offline, and whenever
the connection is available, the data will be sent to the server. One more important
part is, the attendance can be marked even though the user is offline, and whenever
the connection is available, the data will be sent to the server. Earlier Control Room
has to decide for the shift of duties like A, B, A and on next day, and it will be like B,
A, B, and so on. But our system itself generates the chart for this by just providing
the start and end date of bandobast [6]. Depending on the same, there is a Bandobast
report of a minimum of 500 pages (typed manually in an earlier system), it was also
automatically generated with more detailed information. Thus, all repeated typing
effort is reduced. At the end of the day, the control Room has to prepare the absentee
report, and the system provides the same in one click only. To use the app, every user
has to login with a valid username and password. The app provides different features
and rights to each user as per the role of the user like sector head, control Room,
point head, etc. Hence lower-level users can’t administrate the bandobast, and hence
security and integrity of critical data are maintained [7]. As the system include both
Website and App, the control room can easily access the necessary details of each
police anywhere with ease. Hence this becomes an advantage of this system.

2 Related Work

In this system, BAAS is very user friendly and dynamic. In this control room can
easily use the Website and App with all facilities. All other special officers like SP,
DYSP, PI have some administration rights. So as per the previous system, all work
can be done with ease as well as some new features are also added to improvise the
system [8]. Any head can know his team instantly after allocation; he/she can easily
scan the QR code to take attendance of the team. General Police user will easily
know their duty details and instructions for the point. By just entering the start and
end date control room will have the report of the entire bandobast. The absentee can
be printed even on a daily basis. By just providing an Aadhar card number, I-card is
generated needs not to fill manually. Any Police person will understand this system
in a few minutes [9]. This system is being developed for CO’s to Allocate duties to
all police, and the further officer can take attendance of their subordinates. Later the
control room can print the entire report in just one click. For this, all police staff has
to register with detailed information so that duties can be allocated to them. Once
590 P. S. Bhandare et al.

all police staff is registered to the system, CO’s can easily allocate them through the
Website to any required point. Then each point’s Head Officer can take attendance of
his/her subordinates using the Android App. Every user can see his/her duty details
in the app. CO’s can change allocation in a few seconds using the app. The app helps
to generate I-Cards for all in a few seconds. The app provides different features
to each user as per the role of the user; hence lower-level users can’t administrate
the bandobast, and hence security and integrity of critical data are maintained. As
the system include both Website and App, the Control Room can easily access the
necessary details of every police anywhere; hence this becomes an advantage of this
system. Now any Bandobast can be managed with ease in less time and less manual
work. Now instead of caring about allocation, attendance and report, the Control
room can control the bandobast properly [10]. The works like bandobast creation,
allocation, and report generation are easy to handle with the desktop, so we have
developed a Website “https://www.pandharpurpolice.in”. Other works like checking
duty details, taking attendance (even offline), I-card generation, allocation changing
and checking, instruction showing can be done easily with a mobile device. That’s
why the “BAAS” App was also developed.

3 System Architecture

BAAS System architecture can be described in Fig. 1.

Fig. 1 BAAS system architecture

Bandobast Allocation and Attendance System 591

4 Experimental Results and Performance

The proposed system is tested by running the system by using it in a VIP Bandobast
and Magh Wari Bandobast with the online host server. We also checked the ability
of the system by registering more than two thousand users. We also created some
testing bandobast. To check the offline feature of the app, we turned o an internet
connection and marked the attendance of some members. All types of results are
properly tested against earlier manual reports. This continuing education course is
intended for any general bandobast like Election bandobast, event Bandobast, etc.
This course introduces a variety of functions and concepts that facilitate the transition
from manual handling of a bandobast to a computerized process of allocation and
attendance system and strengthen the practice for years to come. The SP of Solapur
District Mr Manoj Patil is very satisfied with the overall system as they have got the
result as they expected. Now it is very easy for them to handle any Bandobast.

5 Conclusion

The project entitled “Bandobast Allocation and Attendance System” is the Hybrid
System (APP + Website) that deals with the issues related to all types of Police
Bandobast in any Town. This project is successfully implemented two times in real
situations (in VIP Bandobast for Uddhav Thakare and Magh Wari) with all the
features mentioned in the Application requirements specification. The application
provides appropriate information to police according to the chosen service. The
project is designed to keep a view on all the issues faced during any type of Police
Bandobast. Deployment of our application will certainly help the City Police Station
(specially Pandharpur Police) to reduce unnecessary wastage of time, paperwork, etc.
Now instead of caring about other issues, the Control room can handle the bandobast
properly without having any stress. The system has helped the COs to do their work
in one or two days despite 5 or 6 days. Bandobast is a delicate event that can be
managed now efficiently and with less work. So this serves the right purpose in
achieving the desire requirements of both the police. It will be quite difficult to get
adopted with the system, but not much time it will take.

6 Future Work

Future work of the system is as follows:

A. To have the feature of photo capturing during attendance as proof.
B. Soon we are going to generalize this system so that Traffic Duty Management
will be handled with this.
C. Send a message to the particular officers.
592 P. S. Bhandare et al.

D. We are trying to make it more user-friendly so that old people in the police can
understand it.
E. In the future, we will have NIC certificates for more security of the system.
F. Now we have only some activities in offline mode; we try to get all activities
in offline mode.
G. We will make the system capable of handling bandobast for a crowd of more
than 25 lakh people (like Kumbh Mela).

7 Advantages of the System

1. No need for much manual work.

2. Creation and Allocation of Entire Bandobast is now the work of a few hours.
3. No need to come to the police station to know the duty or team details.
4. Every member is tracked during attendance.
5. All records are maintained properly.
6. Any new instruction can be added at any moment.
7. Smart I-Cards available to each user.
8. No need for tiresome and time-consuming work.
9. Need not type a long report (400 pages).
10. Easiest management of any Bandobast.
11. One-click reports are available.

References

1. Zaharchuk TM, Atcheson RM, Palmar J (1981) Study of police management. ISBN 0-662-
11643-7)
2. Mohammad MH (201l) Attendance using QR code at University of Sulaimaniyan. J Math
Comput Sci 4(3). https://www.unifoglive.com/QRcode
3. Pandya KH, Galiywal HJ (2014) A survey on QR codes: in context of research and application.
Int J Emerg Technol Adv Eng 4(3)
4. Jivasoftt CORP (2010) On duty police scheduling software. www.osocer.com
5. Kasinathan V, Mustapha A, Chandran R (2018) Employee tracking using GPS. MATEC Web
Conf 150:05015
6. Ernst AT, Jiang H, Sier D (2004) StaVT scheduling and roistering. Euro J Oper Res. https://
www.elsevier.com/locate/dsw
7. Dezezhyts Y (2013) Thesis on android application development, Published by HAGAHELIA.
The University of Applied Science
8. Wei X, Manori A, Pasai N (2017) Smart attendance system using QR code. Int J Smart Bus
Technol 5(1). EZSCAN attendance system from SAMMY by IEEEIMS
9. Ghuman SS (2014) Software testing techniques. Int J Comput Sci Mobile Comput 3(10). ISSN:
2320-088X)
Fire Fighting Un-manned Air Vehicle
for Remote Areas

N. Shashank Bhat, K. S. Shashidhara, and Veerendra Dakulagi

Abstract Fire accident results in catastrophic injuries and devastating damage. The
death rate in India due to fire accidents was almost 2.5 times more than in other
parts of the world. Fire fighting is a highly difficult and challenging task for human
beings to access the remote target areas. By using unmanned aerial vehicle, quick
response to the fire affected area can be achieved and also firefighters will get the
visual information of the fire accident. This work focuses on the implementation
of Un-manned Air Vehicles (UAV)s that can extinguish the fire. The proposed fire
fighting UAV system consists of Hexacopter as a platform. Hexacopter is a UAV that
works with six motors to achieve stable flight and better lift loading capability. The
goal of this work is achieved with stable and robust hexacopter along with dropping
mechanism which is used to drop the fire extinguishing ball on fire-affected area
and camera interface for live video footage. The description of the proposed work is
briefly described as well as determines the principle functionality.

Keywords Fire fighting UAV · Hexacopter · Camera · Dropping mechanism ·

GPS · RF radio transceivers · Flight controllers

1 Introduction

The research on Unmanned Aerial Vehicles has always been an interest of engineers
and hobbyists due to its wide range of applications such as rescue mission, mapping,
aerial photography, border patrol, etc. These vehicles do not require any pilot on-
board and can be easily controlled from the base station. These vehicles are equipped
with GPS and hence they can autonomously fly to the desired location when a path
is defined. Hexacopters are equipped with six motors to provide thrust to the vehicle.
With the proper arrangement of these motors, stability can be obtained [1]. Depending

N. S. Bhat · K. S. Shashidhara
Department of ECE, Nitte Meenakshi Institute of Technology, Bangalore, India
e-mail: [email protected]
V. Dakulagi (B)
Department of ECE, Guru Nanak Dev Engineering College, Bidar, Karnataka, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 593
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_60
594 N. S. Bhat et al.

on the type of application different mechanisms/payloads, cameras are installed into

the hexacopter. City Fire causes a huge problem in people’s lives and also damages
their property. With the increase in the area and height of the buildings, the damage
caused by the fire is more. During such situations, it becomes necessary to access
the fire affected area quickly [2]. Forest fire threatens wildlife and causes loss to the
forest ecosystem. Forest fires are prone to spread faster. Therefore it is very important
to extinguish the fire when it’s in the initial stage. In the case of forest fire preventing
the fire is more important than fighting the fire. So if the fire is detected in the initial
stage the damages can be reduced to a greater extent [3]. Fire extinguishing balls
are nowadays used in extinguishing small fires such as grassfire etc. Due to its small
size and less weight, it becomes advantageous over traditional fire fighting in many
situations [4, 5]. Hexacopters can carry these fire extinguisher balls and drop into
the fire affected area thereby providing faster response to the accident. Dropping
mechanisms are implemented on the hexacopters to carry the fire extinguisher ball
and drop it to the desired place. The dropping mechanism is also controlled by the
remote from the base station. The camera is installed on the hexacopter to get live
footage of the fire-affected area. This helps in getting more information about the
fire accident.

2 Related Work

The Design and implementation method of “fire fighting UAV” and methods to
extinguish the fire has been explored [6, 7]. It also contains the control structure
of UAV and mission control architecture. The water spraying technique is used on
the fire-affected area which is complicated, instead of spraying technique using fire
extinguishing balls makes the scenario less complex and more efficient [6].
Two separate drones one forgetting the visual information and another for drop-
ping the fire extinguishing ball [8]. The improvisation in our project is that we are
achieving both of these objectives using single UAV. This includes battery time,
flying capacities, and flight velocity with respect to battery voltages which helps to
build a drone [9]. The suitable materials and component used to make the drones has
been discussed [10] with respect to applications.

3 Proposed System

3.1 Objective

The objective of this project is to give a quick response to the fire-affected area and
to acquire better information and a better view of the fire-affected area. This project
Fire Fighting Un-manned Air Vehicle for Remote Areas 595

also helps to extinguish the fire which is not easily accessible by human firefighters.
It also helps to reduce the risk of the life of human firefighters.

3.2 Existing System

When the fire station gets to know about the fire accident, they plan on the resources
to be taken to extinguish the fire. Fire Extinguisher vans are deployed to reach the
fire accident area. There is a delay in giving a faster response to the accident. More
the delay implies more is the spreading of fire and hence more casualties. Once they
reach the accident area they manually analyze the situation and try to get valuable
information from that. Due to many restrictions such as the height of buildings, dense
forests, etc., it becomes difficult to get information from the accident place. In these
situations, the firefighters need to risk their lives in order to get clear information
about the accident and extinguish the fire.
In order to overcome the above-mentioned drawback, we have implemented a
system which contains a Drone along with dropping mechanism. The Drone helps
in providing quick assistance to the fire accident and the dropping mechanism helps
to drop the fire extinguisher ball on the fire. The camera interfaced with the Drone is
used to take live visuals of the accident, thus acquiring more information about the
incident. The Drone and the dropping mechanism is controlled by a remote in the
base station. The sensors such as accelerometer, compass, GPS present in the Drone
provides stability in order to achieve different flight modes.

4 Design and Implementation

Fire fighting UAV system consists of two main units Drone Unit, and Dropping
Mechanism as shown in Fig. 1, they co-ordinate with each other and respond to the
accident immediately. It helps in gathering more information from the fire accident
area and reduces the risk of firefighters.

4.1 Drone Unit

A Drone uses six motors along with propellers to produce a thrust which enables it to
lift itself from the ground. The flight controller board forms the heart of the hexacopter
which processes different instructions and performs desired actions. Electronic speed
controllers are used to control and regulate the speed of an electric motor. Electronic
Speed Controller (ESC) also provides a reversing of the motor rotating direction and
dynamic braking. It changes DC battery power into 3-phase AC for rotating brushless
596 N. S. Bhat et al.

Fig. 1 Block diagram of the drone

motors. The battery is used to provide power to all the circuitry and motors. Higher
the battery power more is the performance of the drone in-terms of flight time.
The drone consists of six motors of 1400 kV. Three motors rotate in a clockwise
direction and the other three in an anti-clockwise direction. Each motor provides
a thrust of 1000 g. The speed of the motors is controlled by ESCs. With the help
of ESCs speed controlling feature, the direction of the drone can be changed. The
flight controller which is positioned at the center monitors the firmware within the
ESC and controls the rotating rate of the motors. The flight controller takes the
input from the receiver and adjusts the motor speed accordingly via the electronic
speed controller. The flight controller is also equipped with many sensors such as a
gyroscope accelerometer, barometer, and GPS. External GPS is installed in the drone
to provide co-ordinate information and altitude information of the drone. Pixhawk
flight controller is used in this project, since it has additional features when compared
to other flight controllers. An Anti-vibrator kit is installed with the flight controller
in order to get a precise reading from the sensors. The camera along with transmitter
and receiver assists in acquiring video footage of the fire accident.
Fire Fighting Un-manned Air Vehicle for Remote Areas 597

4.2 Mission Planner

Mission Planner is the firmware used to calibrate all the sensors present in the flight
controller. It is used to program different flight modes for the drone, such as Loiter,
Alt Hold, Stabilize, and Return to Launch, etc. This firmware is also used to program
different switches of the transmitter to perform specific actions. It is used for battery
monitoring to enable special functions like, failsafe.

4.3 Dropping Mechanism

This unit is used to hold and drop the fire extinguisher ball on the desired location.
It consists of a container with two doors along with a locking mechanism. The fire
extinguisher ball is placed inside the container and locked. When the drone reaches
the desired location where the ball is to be dropped, with the help of the remote
the lock is made open. When the lock is opened the ball drops and the fire gets
extinguished.

5 Experimental Results

The Drone unit and the Dropping Mechanism unit have been developed and the
system is successfully coordinated. Initially the drone without payload as shown in
Fig. 2 has been tested successfully.

Fig. 2 Built drone without payload

598 N. S. Bhat et al.

Fig. 3 Drone carrying payload

Then the drone was made to fly with a payload of around 1.5 kg (Overall weight
of the drone is around 4 kg) as shown in Fig. 3. During this testing drone flew for
around 2 min, suddenly it crashed from 40 ft height. On this crash, a drone wing,
motor, flight controller, and propellers got damaged.
After installation and proper calibration of GPS, both loiter and RTL mode worked
properly as per our requirements as shown in Fig. 4. Drone flight time testing was
carried out and drone flight time was recorded to be 20 min.
The dropping mechanism system is integrated with the Hexacopter to obtain Fire
Fighting UAV as shown in Fig. 5.
As depicted in Fig. 6, Drone fleet successfully and same has been captured.

6 Conclusion

In this paper, we built a UAV that can carry a payload (fire extinguisher ball) and
deploy it on the fire. The camera integrated on the UAV provides visuals to the UAV
controller. The UAV is controlled remotely with the help of RF Transceiver. The drone
is designed to extinguish the fire on buildings and forests. It also consists of a wireless
camera to get the live video footage of the particular locations in fire-affected places.
The objective of this work is accomplished with a steady and robust hexacopter along
with a dropping mechanism which is used to drop the fire extinguishing ball on fire-
affected area and camera interface for live video footage. The proposed drones could
Fire Fighting Un-manned Air Vehicle for Remote Areas 599

Fig. 4 Drone with GPS

Fig. 5 Fire fighting UAV

600 N. S. Bhat et al.

Fig. 6 Fire fighting UAV during flight

be exploited in the emergencies which incidentally happen, and the fire station needs
sophisticated technologies for the rescue missions. The demand for such drones is
indeed very high in aircraft and fire rescue operations.

References

1. Zakaria AH, Mustafah YM, Mudzakkir M, Hatta M, Azlan MNN (2015) Development of load
carrying and releasing system of Hexacopter. In: Control conference (ASCC), pp 1–6
2. Chen R, Cao H, Cheng H, Xie J (2019) Study on urban emergency firefighting flying robots
based on UAV. In: 2019 IEEE 4th advanced information technology, electronic and automation
control conference (IAEAC), 20–22 Dec. ISSN: 2381–0947. https://doi.org/10.1109/IAEAC4
7372.2019.8997723
3. Chen Y, Zhang Y, Xin J, Wang G, Mu L, Yi Y, Liu H, Liu D (2019) UAV image-based forest
fire detection approach using convolutional neural network. In: 2019 14th IEEE conference
on industrial electronics and applications (ICIEA), 19–21 June 2019. https://doi.org/10.1109/
ICIEA.2019.8833958
4. Aydin B, Selvi E, Tao J, Starek MJ, Use of fire extinguishing balls for a conceptual system of
drone-assisted wildfire fighting. https://doi.org/10.3390/drones3010017
5. Radu VT (2019) Use of drones for firefighting operations. Master thesis report MSc. Risk and
Safety Management, 10/01/2019, Aalborg University, Esbjerg, Denmark
Fire Fighting Un-manned Air Vehicle for Remote Areas 601

6. Qin H, Cui JQ, Li J, Bi Y, Lan M, Shan M, Liu W, Wang K, Lin F, Zhang YF, Chen BM
(2016) Design and implementation of an unmanned aerial vehicle For autonomous firefighting
missions. In: 12th IEEE international conference on control & automation (ICCA) Kathmandu,
Nepal, June 1–3
7. Alagirisamy M (2020) Efficient coherent direction-of-arrival estimation and realization using
digital signal processor, IEEE Trans Antennas Propag 68(9): 6675–6682
8. Suprapto BY, Heryanto MA, Suprijono H, Muliadi J, Kusumoputro B (2017) Design and devel-
opment of heavy-lift hexacopter. In: 2017 international seminar on application for technology
of information and communication
9. Hwang M-H, Cha H-R, Jung SY (2018) Practical endurance estimation for minimizing energy
consumption of multirotor unmanned aerial vehicles
10. Baranek R, Solc F (2012) Modelling and control of a Hexa-copter. In: 2012 13th international
Carpathian control conference
Human Age Classification
and Estimation Based on Positional
Ternary Pattern Features Using Ann

Shamli V. Jagzap, Lalita A. Palange, Seema A. Atole, and Geeta G. Unhale

Abstract In this paper, Positional Ternary Pattern features based Human Age classi-
fication using Artificial Neural Network for Forensic science application. The classifi-
cation of human age from facial pictures plays an important role in pc vision, scientific
discipline, and forensic Science. The various machine and mathematical models, for
classifying facial age together with Principal Component Analysis (PCA), Positional
Ternary Pattern (PTP) are planned yields higher performance. This paper proposes a
completely unique technique of classifying the human age group exploitation Artifi-
cial Neural Network. This is often done by preprocessing the face image initially and
so extracting the face options exploitation PCA. Then the classification of human age
is finished exploitation Artificial Neural Network (ANN). The method of combining
PCA and ANN performs higher rather than victimization separately.

Keywords Facial recognition · Positional ternary pattern · GLCM feature

extraction · Principal component analysis · Artificial neural network

1 Introduction

With the substantial increase of personalized interaction with consumer merchandise,

human age recognition is obtaining a lot of attention meant for various Human–
computer Interaction (HCI) and identification tasks. Hence, an efficient’ Age-specific
Human–computer Interaction’ (AHCI) system designed for consumer merchandise
is required, that has its specific relevance in several fields, such as forensic art, police
investigation observation, security management, net access management etc. because
the face presents perceptible information of ageing, automatic age recognition from
face pictures, has been researched broadly speaking. Throughout ageing, the most
changes in the face area associated with craniofacial growth and skin ageing. With the
craniofacial growth, the entire face step by step gets prolonged, primarily vertically.
Likewise, the skin becomes a lot of thinner, darker, and fibrous together with dynamic

S. V. Jagzap (B) · L. A. Palange · S. A. Atole · G. G. Unhale

SVERI’s College of Engineering Pandharpur, Pandharpur, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 603
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_61
604 S. V. Jagzap et al.

wrinkles with age growth. Therefore, representing said facial changes is imperative
in age recognition.
Human faces, as necessary visual cues, convey a major quantity of nonverbal infor-
mation to facilitate the real-world human-to-human communication. As a result, the
trendy intelligent systems are expected to own the aptitude to accurately acknowledge
and interpret human faces in real-time. Facial attributes, like identity, age, gender,
expression, and ethnic origin, play an important role in real facial image analysis
applications as well as transmission communication, human–computer interaction
(HCI), and security. In such applications, numerous attributes will be calculable from
a captured face image to infer the additional system reactions. For instance, if the
user’s age is calculable by a pc, associate age-specific human–computer interaction
(ASHCI) system could also be developed for secure network/system access manage-
ment. The ASHCI system ensures young children don’t have any access to web pages
with adult materials. A slot machine, secured by the ASHCI system, will refuse to
sell alcohol or cigarettes to the underage individuals. In image and video retrieval,
users might retrieve their pictures or videos by specifying a needed age vary. Ad-
agency will verify what reasonably scroll advertisements will attract the passengers
(potential customers) in what age ranges employing a latent laptop vision system.
Although automatic image-based age estimation is a very important technique
concerned in several real-world applications, it’s still a difficult drawback to estimate
human ages from face pictures. Face pictures of two people with totally different ages.
Since totally different people age quite; otherwise, the ageing method is decided by
not only the person’s factor however conjointly several external factors, like health,
living style, living location, and atmospheric condition. Males and females can also
age otherwise thanks to the various extent in victimization makeups and accessories.
The way to extract general discriminative ageing options, whereas reducing the
negative influence of individual variations still remains an associate degree open
problem.
There exits some work on age synthesis and rendering within the last many
decades, but there are only a few publications on age estimation owing to the
complexity of ageing patterns. The age progression displayed on faces is uncon-
trollable and personalized. Such special characteristics of ageing variation cannot
be captured accurately owing to the prolific and heterogeneous information sent by
human faces. On the opposite hand, the age estimation downside is totally different
from the matter of face recognition with age variation, wherever the goal is to estimate
facial identities whereas no ages are calculable from the input faces. The research
effort on age estimation might facilitate recognizing faces containing age variations.
In this paper, we propose a PTP and PCA feature extraction to detect faces in face
images. The feature extraction first detects the faces using the RGB colour model
and divides the face region into blocks of equal size then extracting the face features
using PCA.
Human Age Classification and Estimation Based … 605

2 Review of Literature

This paper concerns the estimation of facial attributes, particularly, age and gender
from images of faces acquired in difficult, within the wild conditions. This downside
has received way less attention than the connected downside of face recognition,
and especially, has not enjoyed an equivalent dramatic improvement in capabilities
demonstrated by modern face recognition systems. Here, we tend to address this
downside by creating subsequent contributions. First, in answer to at least one of the
key issues of age estimation analysis absence information we provide a singular data
set of face pictures, tagged for age and gender, non-inheritable by smart-phones
and alternative mobile devices, and uploaded without manual filtering to online
image repositories. We tend to show the pictures in our assortment to be tougher
than those offered by alternative face-photo benchmarks [1]. Second, we tend to
describe the dropout-support vector machine approach utilized by our system for
face attribute estimation, so as to avoid over fitting. This technique, impressed by
the dropout learning techniques currently popular deep belief networks, is applied
here for coaching support vector machines, to the most effective of our data, for
the primary time. Finally, we tend to present a robust face alignment technique that
expressly considers the uncertainties of facial feature detectors. We tend to report
in-depth tests analyzing each the issue levels of up to date benchmarks likewise
because of the capabilities of our own system. These show our technique to outstrip
progressive by a good margin [2].
We think about the matter of automatic age estimation from face pictures. Age esti-
mation is sometimes developed as a regression downside relating the facial expression
and therefore the age variable, and one regression model is learnt for all ages [3]. We
have a tendency to propose a hierarchical approach, wherever we have a tendency to
initial divide the face pictures into numerous age teams, so learn a separate regression
model for every cluster. Given a take a look at the image, we have a tendency to initial
classify the image into one amongst the age teams, so use the regression model for
that specific cluster. To enhance our classification result, we have a tendency to use
many alternative classifiers and fuse them mistreatment the bulk rule. Experiments
show that our approach outperforms several states of the art regression ways for age
estimation [4].
Human age will give important demographic data. During this paper, we have
a tendency to tackle the estimation older in face pictures with possibilities. The
look of the planned methodology is predicated on the relative order older labels
within the information. The age estimation drawback is reworked into a series of
binary classifications achieved by convolution neural network. Every classifier is
employed to evaluate whether or not the age of input image is larger than a precise
age, and therefore the calculable age is obtained by adding likelihood values of those
classification issues. The planned method: Deep possibilities (DP) of facial age show
enhancements over direct regression and multi-classification strategies [5]. Design
Sparse Features like wrinkles, dark spots, moles etc. were used to analyze the image
and estimate the age [6].
606 S. V. Jagzap et al.

3 Block Diagram

Block diagram of Human age classification using ANN

4 System Analysis

The block diagram for the proposed algorithm is as shown in Fig. 1. System analysis is
done by steps Input image acquisition, Database creation, Pre-processing, Positional
Ternary Pattern, Principal Component Analysis, GLCM Feature Extraction, Artificial
Neural Network, Result.

4.1 Preprocessing

The aim of preprocessing is associate degree improvement of the image information

that suppresses unwanted distortions or enhances some image options relevant for
more process and analysis task.
Human Age Classification and Estimation Based … 607

Fig. 2 PTP code computation steps

4.2 Positional Ternary Pattern

Figure 2 shows the Positional Ternary Pattern (PTP). Which assigns eight-bit
computer code to every element of a picture? Initially, Kirsch compass masks
compute the sting response of eight neighborhood pixels.

4.3 Principal Component Analysis

Principal components analysis (PCA): PCA seeks a linear combination of variables

such as the most variance is extracted from the variables. It then removes this vari-
ance and seeks a second linear combination that explains the most proportion of the
remaining variance, and so on. This can be referred to as the principal axis method-
ology and leads to orthogonal (uncorrelated) factors. PCA analyzes total (common
and unique) variance.
Eigenvectors: Principal elements (from PCA—Principal elements analysis) repli-
cate each common and distinctive variance of the variables and will be seen as a
variance-focused approach seeking to breed each the overall variable variance with
all elements and to breed the correlations. PCA is much a lot of common than PFA,
however, and it’s common to use “factors” interchangeably with “components”. The
principal elements square measure linear combos of the first variables weighted by
their contribution to explaining the variance in an explicit orthogonal dimension.
Eigenvalues: Additionally referred to as characteristic roots. The eigenvalue for a
given issue measures the variance all told the variables that are accounted for by
that issue. The magnitude relation of eigenvalues is that the magnitude relation of
instructive importance of the factors with relation to the variables.
608 S. V. Jagzap et al.

Fig. 3 GLCM (gray level co-occurrence matrix)

4.4 GLCM Features Extraction

This process using creates a GLCM, use the gray co-matrix function. Particular
function create grey level Co-occurrence matrix GLCM by calculation how often a
particular pixel with the intensity value I occur in a specific spatial relationship to a
pixel with the value j. This process number of grey levels in the image determines
the size of the GLCM..In GLCM can also derive stoical measures, and the derive
statistics from GLCM and more information from plot correlation shown in Fig. 3.
How to grey co-matrix calculate the 1st three value in given GLCM.

4.5 Artificial Neural Network

The performance of the artificial neural network was evaluated in terms of coaching
performance and classification accuracies. Artificial Neural Network offers quick and
correct classification and could be a promising tool for the classification of the result.
The ANN with FF is trained with reference options set and desired output exploitation
‘new’ and ‘train’ command. Here, target one for dataset1, two for dataset2 and
dataset3 area unit taken as the desired output. Once the coaching, updated weight
issue and biases with alternative network parameters area unit hold on to simulate
with input options. At the classification stage, take a look at image options area unit
utilized to simulate with trained network model exploitation ‘sim’ command. Finally,
it returns the classified price as one, two or three supported that the choice is going
to be taken as our age classification.
Human Age Classification and Estimation Based … 609

Fig. 4 Result of child age

5 Result and Discussion

In this work, our own database in spite of the existence of other databases has been
created. In this technique, we are using Gabor filter we can calculate the individual
face and mark the key points after that we are applying LDA feature extraction we
can get the output. The output result should become 75% of accuracy only. But in our
proposed system we are introducing the detecting human face then classifying age
using PTP, PCA and Artificial Neural Network classification we can get the output.
The output result should become 92% of accuracy. In Figs. 4 and 5 shows the result
of the child and young age group. In Figs. 6 and 7 shows the result of the middle and
old age group.

6 Conclusion

In this paper, By using MATLAB 2014, we have a tendency to propose a way for
human age classification using a positional ternary pattern, PCA and ANN classifica-
tion and used this prediction in our automatic system for detecting human behavior.
This work could be a continuous study from previous analysis on the employment
of heterogeneous knowledge for autonomous detection of human behavior. Face
610 S. V. Jagzap et al.

Fig. 5 Result of young age

Fig. 6 Result of middle age

Human Age Classification and Estimation Based … 611

Fig. 7 Result of old age

pictures are a sort of knowledge which might improve the proof in work a suspect or
an occurrence. This technique can cut back the scope in distinctive suspicious persons
within the forensic investigation space. The primary step within the methodology
involves extracting the options from face pictures employing a positional ternary
pattern and PCA. Then, we have a tendency to applied ANN to classify the ages.
The transfer learning was then accustomed with success utilize the already learn
data for a replacement task with the restricted dataset, despite the fact that, once
the dataset is little, the results area unit far better than mistreatment traditional hand-
crafted options like Haar. The planned technique was tested on the Face datasets, and
therefore the results are unit terribly encouraging the odds of accuracy for age clas-
sifications are 92.33% and 80.17% respectively, that demonstrates that the answer is
possible. Within the future work, the results are going to be combined with another
dataset. Knowledge fusion techniques are going to be applied to integrate all the
knowledge to provide an improved overall call which might assist investigators to
find suspicious persons.
612 S. V. Jagzap et al.

References

1. Eidinger E, Enbar R, Hassner T (2014) Age and gender estimation of unfiltered faces. IEEE
Trans Inf Forensics Secur 9(12):2170–2179
2. Choi SE, Lee YJ, Lee SJ, Park KR, Kim J (2011) Age estimation using a hierarchical classifier
based on global and local facial features. Pattern Recogn 44(6):1262–1281
3. Guo G, Fu Y, Dyer CR, Huang TS (2008) Image-based human age estimation by manifold
learning and locally adjusted robust regression. IEEE Trans Image Process 17(7):1178–1188
4. Mahalingam G, Kambhamettu C (2011) Can discriminative cues aid face recognition across
age? In: Proceedings of IEEE international conference on automatic face gesture recognition
workshops, Mar 2011, pp 206–212
5. Zheng T, Deng W, Hu J (2017) Deep probabilities for age estimation. In: 2017 IEEE visual
communications and image processing (VCIP)
6. Suo J, Wu T, Zhu SC, Shan S, Chen X, Gao W (2008) Design sparse features for age estimation
using hierarchical face model. In: Proceedings of 8th 2008 international conference on automatic
face and gesture recognition
Object Recognition Using Fuzzy
Classifier

Seema A. Atole, Shamli V. Jagzap, Lalita A. Palange, and Akshay A. Jadhav

Abstract In this paper, object recognition is proposed using combined DRLBP and
SIFT features for high efficient signal transfer system applications. The aim of this
research is to develop a non-real-life application of a security lock system employing
object recognition methodology. DRLBP is chosen for the object recognition algo-
rithmic program. Arduino microcontroller is employed to represent the response to
object identification. USB serial communication is employed to inter-object between
the MATLAB and Arduino UNO Microcontroller. First, the image of the individual
is captured then the captured image is then transferred to the information developed
in MATLAB during this stage, the captured image compares to the training image
within the database to see the individual standing. If the system acknowledges the
individual as an authentication person or un-authentication person, the result is sent
to the Arduino UNO microcontroller.

Keywords Facial recognition · Facial identification · SIFT features · DRLBP ·

Fuzzy classifier

1 Introduction

Object identification identifies the external body part in recordings and computerized
photos will be viewed as associate degree example order issue and therefore, the
underlying amount of any object acknowledgement framework. The not insignificant
summation of utilization zones, for instance, Human—laptop Inter objects, Security
Systems, closed-circuit television, Content primarily based Image Retrieval, and then
forth demonstrates the importance of object identification and acknowledgement
calculations. As a rule, an object recognition framework can gain an image and
acknowledges the confronts freelance of stance, scale or outward appearances. The
target of our work is to differentiate and acknowledge covering confronts, one object
lies or reaches out over and covers some portion of another face, during an image.

S. A. Atole (B) · S. V. Jagzap · L. A. Palange · A. A. Jadhav

SVERI’s College of Engineering Pandharpur, Pandharpur 413304, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 613
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_62
614 S. A. Atole et al.

Whenever a minimum of two appearances is coated, it is a testing enterprise to

differentiate the countenances severally utilizing the standard object identification
ways. What is a lot of, the degree many-sided quality is going to be modified relying
upon each calculation? In facial part based mostly methodologies, if the territory
of covering is larger than the opposite, the district low believability is going to
be disposed of. This issue could prompt lose of essential knowledge of various
appearances, and it’ll influence the unwavering quality of security frameworks.
In this paper, we propose a SIFT and DRLBP feature extraction to detect objects
in object images. The feature extraction first detects the objects using RGB (Red
Green Blue) colour model and divides the object region into blocks of equal size
then extracting the object features using SIFT. After, the Fuzzy classifier method is
used to classify the human object images are Authenticate or Un-authenticate.

2 Review of Literature

During this paper, we tend to plan a non-contact technique to classify the object form
by victimization Support Vector Machine (SVM) technique [1]. This algorithmic rule
consists of three steps: head segmentation, object plane identification, and object form
classification. The accuracy rate is sixty eight 68% [2]. A compact and reliable room
attending system victimization RFID and object verification is presented during this
paper. The RFID system identifies the student victimization the RFID card, and any
identification of the student is applied victimization object recognition technique [3].
The performance of the system is tested for frontal object verification, head create
varied object verification, and detection of proxy attending is applied. It’s found that
the projected theme verifies the identity of the scholar properly of concerning 98 for
frontal object and to makes an attempt on poses varied object verification [4, 5]. The
proxy attending detection applied for frontal object resulted in Associate in nursing
potency of 73.28% Associate in Nursing for various poses resulted in the potency
of 79.29% [6]. Object recognition is that the core application within the biometric
technology space. It’s wide employed in the advanced application of artificial intelli-
gence and laptop vision. Raising business of object recognition and low social control
makes a call for participation of it will increase within the last decade—the new tech-
nique used as feature extractor addition to ancient PCA [7]. Gathering options by
victimization slope technique and PCA is to search out the optimum objects vectors
because of the inputs to the classifier (NNMLP neural network) [8]. Results have
unconcealed acceptable correct classification. As a knowledge check set, we have a
tendency to used BIO-ID knowledge base within the projected system [9].

3 Block Diagram

Figure 1 shows the block diagram for Object recognition.

Object Recognition Using Fuzzy Classifier 615

Fig. 1 Block diagram of object recognition using combined DRLBP and SIFT features

4 System Analysis

1. Input image
2. Database creation
3. Preprocessing
4. DRLBP and Histogram
5. SIFT Features
6. GLCM Feature Extraction
7. Fuzzy Logic classifier
8. Result.

4.1 Preprocessing

Image pre-processing is the term for operations on images like changing the RGB
image to a grey one by adjusting the resolution of the image as needed. These opera-
tions don’t increase image information content; however, they decrease it if entropy
is associate degree metric. The aim of pre-processing is associate degree improve-
ment of the image information that suppresses unwanted distortions or enhances
some image options relevant for more process and analysis task. Plane Separation
on Red/inexperienced/Blue happens.
616 S. A. Atole et al.

4.2 DRLBP

In Fig. 2 DRLBP process shows the descriptor local binary pattern is used to compare
all the pixels, including the center pixel with the neighboring pixels in the kernel
to improve the robustness against the illumination variation. An LBP code for a
neighborhood was produced by multiplying the threshold values with weights given
to the corresponding pixels and summing up the result. LBP codes are weighed using
gradient vector to generate the histogram of robust LBP, and discriminative features
are determined from the robust local binary pattern codes. DRLBP is represented

Fig. 2 DRLBP process

Object Recognition Using Fuzzy Classifier 617

in terms of the set of normalized histogram bins as local texture features. It is used
to discriminate the local edge texture of object invariant to changes of contrast and
shape.

4.3 SIFT Feature

The SIFT algorithm takes a picture and transforms it into a set of native feature
vectors. Every of those feature vectors is meant to be distinctive and invariant to
any scaling, rotation or translation of the image. Within the original implementation,
these options are often accustomed to notice distinctive objects in different pictures,
and therefore the rework is often extended to match objects in pictures. This report
describes our own implementation of the SIFT algorithmic program and highlights
potential direction for future analysis. The SIFT options represented in our imple-
mentation are computed at the sides and that they are invariant to image scaling,
rotation, the addition of noise. They’re helpfully attributable to their distinctiveness,
which allows the right match for key points between objects. These are achieved by
victimization our Gradient-Based Edge Detector, and therefore the native descriptors
bestowed round the key points. Edges are poorly outlined and typically exhausting
to observe; however, there are still giant numbers of key points can be extracted
from typical pictures. Thus we are able to still perform the feature matching even
the objects are small. Generally, the photographs are too swish to search out that a
lot of options for an identical, and in that case tiny low object may well be unrec-
ognized from the coaching pictures. In the next step, we are going to attempt to
perform some object identification, and that we select the closest neighbor or second-
closest neighbor algorithmic program that could be a smart technique to try to the
key points matching. There’s another helpful technique to acknowledge objects by
learning an applied mathematics model. During this technique, a symbolic logic
model is employed to acknowledge the objects. Associate in Nursing Expectation–
Maximization (EM) algorithmic program is employed to find out the parameters in
an exceedingly most chance framework. Hopefully, we are able to limit the model
to a tiny low quantity of components that is economical for matching objects.

4.4 GLCM Features Extraction

This process using creates a GLCM, use the grey co-matrix function. Particular
function create grey level Co-occurrence matrix GLCM by calculation how often a
particular pixel with the intensity value I occur in a specific spatial relationship to a
pixel with the value j. This process number of grey levels in the image determines the
size of the GLCM. By default, grey co-matrix uses scaling to reduce the number of
intensity values in an image to eight, and you can use the Num Levels and the Gray
618 S. A. Atole et al.

Limits parameters to control this scaling of grey levels. GLCM has revealed certain
properties about the spatial distribution of the grey levels in the texture signals.

4.5 Fuzzy Logic Classification

In the section, Fuzzy rule-based mostly system that is employed for facial features
recognition from object expression. Fuzzy is one helpful approach for fuzzy classi-
fication, which may verify the intrinsic division in an exceedingly set of untagged
knowledge and notice representatives for solid teams. Fuzzy integrals were accus-
tomed to describe the uncertainty of facial features. Facial features house might
be created mechanically and compared for expression classification. The core of
our system could be a Fuzzy Rule-based mostly system that is employed for facial
features recognition from object expression. Fuzzy Logic is often won’t too kind
linguistic models and comes with a solid qualitative base. Fuzzy systems are utilized
in several classification and management issues as well as facial features recognition.

5 Result and Discussion

In the existing system, we are using two databases one is input database, and another
one is storage database. In this technique, we are using Gabor filter we can calculate
the individual object and mark the key points after that we are applying LBA feature
extraction we can get the output. The output result should become 75% of accuracy
only. But in our proposed system, we are introducing the detecting object using
DRLBP, SHIFT and Fuzzy Logic classification we can get the output. The output
result should become 95% of accuracy.
In Figs. 3 and 4 shows the input image and detection of the face after that in
Fig. 5 shows the dominant rated local binary pattern and then shows the histogram
of dominant rotated local binary pattern in Fig. 6. After completion of histogram of
DRLBP process then shows the R-Eye, L-Eye, Nose and Mouth in Fig. 7. In Figs. 8
and 9 shows the authentication result. In Fig. 10 shows the Performance Analysis of
the input image.

6 Conclusion

In our purport technique to agnize an object that is relegated into a dyad of categories.
The strategy introduced during this article was tested on the JAFFE information
which incorporates ten persons. The information consists of ten pictures. 70% of
the information was used for training, and therefore the remaining 30% for testing
pictures. The accuracy of the planned system is compared with those of different
Object Recognition Using Fuzzy Classifier 619

Fig. 3 Input image

Fig. 4 Detection of face

ways. It might be seen that the popularity of the planned technique performs higher
than the opposite technique. In the future scope of my project is we are working
video to video processing for surveillance and biometric application.
620 S. A. Atole et al.

Fig. 5 DRLBP

Fig. 6 Histogram of
DRLBP
Object Recognition Using Fuzzy Classifier 621

Fig. 7 DRLBP of R-Eye, L-Eye, nose and mouth

Fig. 8 Object landmark

detection

Fig. 9 Result
622 S. A. Atole et al.

Fig. 10 Performance analysis

References

1. Sarakon P, Charoenpong T, Charoensiriwath S (2014) Object shape classification from 3D human

data by using SVM. In: The 2014 biomedical engineering international conference (BMEiCON)
2. Kar N, Debbarma MK, Saha A, Pal DR (2010) Study of implementing automated attendance
system using face recognition technique. In: 2010 32nd global conference on IEEE information
technology inter objects (ITI)
3. Pss S, Bhaskar M (2016) RFID and pose invariant face verification based automated classroom
attendance system. In: International conference IEEE
4. Ahmady M, Ghasemi R (2013) Local weighted pseudo Zernike moments and fuzzy classification
for facial expression recognition. In: 13th Iranian conference on fuzzy systems (IFSC)
5. Alazzaw A, Uçan ON, Bayat O (2017) Object recognition based on multi features extractors.
In: International conference on engineering and technology (ICET)
6. Wang X, Cai Y, Abdulghafour M (2015) A comprehensive assessment system to optimize
the overlap in DCT-HMM for object recognition. In: 2015 11th international conference on
innovations in information technology (IIT). IEEE
7. Kuriakose RB, Vermaak HJ (2015) Developing a java based RFID application to automate
student attendance monitoring. In: 2015 pattern recognition association of South Africa and
robotics and mechatronics international conference (PRASA-RobMech) Port Elizabeth, South
Africa, Nov 26–27. IEEE
8. Slavkovic M (2013) Face recognition using Gabor filters, PCA & neural networks, pp 978-1-
4799-0944-5/13/$31.00© IEEE
9. Zhoa W (2011) Fully automated object detection & facial feature extraction using Local Gabor
Filter point. 978·1·4577·0308-9/11/$26.00 © IEEE
An Effective Approach for Accuracy
of Requirement Traceability in DevOps

Vinayak M. Sale, Somnath Thigale, B. C. Melinamath, and Siraj Shaikh

Abstract Requirement fulfilment is an essential factor in the success of software.

The various stakeholders specify the requirements should be satisfied with each point
of the development of the software. DevOps is a mutual directorial endeavor to auto-
mate continuous rescue of new software renew while assurance their accuracy and
consistency. Requirement traceability helps software engineers to trace the require-
ment from its starting point to its completion. In the software development process,
traceability helps in various ways, like change management, software maintenance,
and confusion prevention. However, many of the challenges can be overcome through
organizational policy, quality requirements traceability tool support remains the open
problem. The traceability links become outdated throughout software updating and
maintenance since the developers can modify or remove some features of the source
code. The proposed method is based on automatically find out characteristic prin-
ciple as of explicit links. The presented system proves to give superior quality results
by comparison. It is also a low cost, the very flexible method to apply regarding
preprocessing the source code and documentation.

Keywords DevOps · Requirements · Traceability · Management · Requirement

traceability approach (RTA) · I.R. technique (IRT)

V. M. Sale (B) · S. Thigale · B. C. Melinamath

Department of Computer Science and Engineering, SVERI’s COE, Pandhapur, India
e-mail: [email protected]
S. Thigale
e-mail: [email protected]
B. C. Melinamath
e-mail: [email protected]
S. Shaikh
Mind I.T. Services, Miraj, Sangli, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 623
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_63
624 V. M. Sale et al.

1 Introduction

Basically, for any software development task, a developer must know the project
background, particularly the architecture of the system, design, working, and the rela-
tions between the various components using any available documentation. Program
conception occurs in a bottom-up method, a top-down approach or some combination
of both [1, 2]. The traceability is an essential factor for any software project, and if
we use it, it could be valuable from different viewpoints for the development. When
we develop source code for any system that source code can be traced and become
identical with the requirement and analysis because we develop a source code as per
the requirements [3, 4]. A traceability link is a connection between the source code
and requirement.
Requirement traceability helps software engineers to trace the requirement from
its emergence to its fulfilment [3, 5]. Traceability may not help us to know how
different components of systems are interlinked and dependent on each other in the
same system. We may also fail to find the impact of change on the software and system
[6]. A major goal of traceability is a link of, in the lack of originality. Requirements
and other artifacts traceability links [4]. Therefore, we should look at traceability
from all the aspects of traceability regarding scope and coverage. Requirements
traceability has proved much important over the past decade in the scientific literature.
It is defined as “the ability to illustrate and go after the life of a requirement, in both
the onward and toward the back direction” [4, 5, 7]. Traceability links among of a
system and its source code help us in reducing comprehension system attempt.
The developers can add, remove, or modify features as per the users’ demand while
updating the software. At the same time, software maintenance and development,
requirement traceability links become marginal because no developer can devote
effort to update it. However, to recover traceability links later is a very painful and
tedious task; also it is costly for developers too [2]. A developer usually does not
update requirement-traceability links with source code. Requirements and source
codes are different from each other, which decrease the textual similarity [3, 6].

2 Reasons for Requirement Traceability

It is most significant to verify that the requirements are properly executed in the
design. This is done with requirements traceability which is usually referred to as
[2, 5, 8] “the capability to demonstrate and go after the life of a requirement, in both
onward and toward the back direction” [9]. Requirements traceability confines the
relationships between the requirements and source code. The traceability is one of
the needs of different stakeholders—project sponsors, project managers, analysts,
designers, maintainers, and end-users, because of their need, priority, and goal [1,
2, 4]. The requirements traceability is a point of a system in which the requirements
are linked to their sources and the objects formed during the system development
An Effective Approach for Accuracy of Requirement … 625

life cycle based on these requirements [7]. The important part of requirements engi-
neering and elicitation phase is that the underlying principle and sources to the
requirements are captured to understand requirements evolution and verification [3].
Modifications in design appear. During design phase requirements, traceability
helps keep track of when the change request is implemented before a system is
redesigned. Traceability can also give information about the validation, significant-
conclusion, and postulation behind requirements [3, 7]. After the delivery of the
system, [1] modifications occur due to various reasons (e.g. to a changing envi-
ronment). The traceability helps us complete, more accurate cost and schedule of
change(s) can be resolute, instead of depending on the engineer or programmer who
is expert [7].
Traceability information allows answering:
1. What is the effect when the requirement is changed?
2. Where is a requirement applied?
3. Are all requirements assigned?
4. Which requirement is deal with by a requirement?
5. Is this requirement essential?
6. What design decisions affect the implementation of a requirement?
7. What are the benefits of this technique and what were the further options?
8. Is the implementation compliant with the requirements?
9. Is this design element necessary?
10. How do I interpret this requirement?

Benefits of Traceability
1. Prevents losing of knowledge
2. Helps for the verification process
3. Impact analysis
4. Change control
5. Process monitoring
6. Better software quality
7. Reengineering
8. Reusability
9. Reduction of Risk.

Basic Traceability Links

Traceability links depend on the traceability information, the linking of maybe
1. One-to-one—one design element to one code module.
2. One-to-many—one functional requirement verified by multiple test cases.
3. Many-to-many—a use case may lead to multiple functional requirements, and
a functional requirement may be common to several use cases.
626 V. M. Sale et al.

3 Background and Related Work

This section presents an environment on the I.R. technique and a review of the related
work. Traceability approach can be separated into three main categories, i.e. dynamic,
static, and hybrid. The dynamic approach gathers and examines execution traces [10]
to recognize the technique that a software link has belen carrying out in the particular
scenario. However, it couldn’t help to differ in overlapping circumstances, because
there are some limitations to a single method [2]. The legacy system may not be
applicable, due to bugs and/or some other issues. Thus, to collect execution traces is
not possible.
Static traceability approaches [11, 12] use source code structure and/or textual
information for recovering traceability associations among high-level and low-level
software artifacts. The combination of static and dynamic information is hybrid
traceability. The study shows that a combination of dynamic and static information
can perform better than the single I.R. technique [5].

3.1 Information Retrieval Technique (IRT)

Information Retrieval (I.R.); refers to a method that would calculate textual simi-
larities of different documents. The textual similarity is calculated using the terms
that occurred in the documents. If two documents have a number of general terms,
those documents are measured to be similar. The analysis of different I.R. methods
can be in three steps [13]. First, after preprocessing such as stop word removal and
stemming, a corpus is made from the documents.
Second, each document is represented as access in an index. The term-by-
document matrix is a common index, where the document as rows and each term as a
column. The incidence of the term arising in the document is the values in the matrix.
Third, by using a cosine similarity formula, the similarity among the index entries is
calculated [14]. The presentation of the key entries and the formula for calculating
the similarity varies depends on the I.R. method. We use the VSM IR method in this
paper and briefly describe it in the following paragraph.
In the Vector Space Model (VSM) [13], each document represents the vector of
terms. In the term-by-document matrix, each row can be measured as one document’s
vector in the space of terms that occur in all documents. The calculation of similarity
of two documents is based on the cosine angle between vectors of each document.
In general, the cosine angle between vectors of the two documents will reduce as the
different documents share more terms. Hence, the higher similarity of the documents
will occur.
An Effective Approach for Accuracy of Requirement … 627

4 System Architecture

4.1 Outline of System

For any software evolution, the essential task is, a developer must understand the
project background [2, 8], in particular, the system planning, propose, how to imple-
ment, and the relations among the different artifacts using any available documen-
tation. Program understanding occurs in a bottom-up way, a top-down way, or some
mixture thereof. Different types of data, ranging from domain-specific knowledge
to general programming knowledge can be used throughout program conception
[3]. Traceability links between source code and part of the documentation, e.g.,
requirements, abet both top-down and bottom-up conception.
Requirement traceability is defined as, “the capability to demonstrate and go
after the life of a requirement, in both onward and toward the back direction” [9].
Traceability links are also necessary to make sure that source code is reliable with
its requirements and that all and only the specified requirements have been imple-
mented by developers. Traceability links are useful in decreasing understanding
effort between the requirements of a system and its source code [4–6]. The trace-
ability information is also useful for software maintenance and development tasks.
For instance, once a developer has traceability links, a user can easily trace what
software artifacts must be changed for the development of a new requirement.
Even with the importance of traceability links, in software maintenance and devel-
opment, as developer update features, requirement traceability links become outdated
because developers do not dedicate effort to update them later [2, 6, 8]. This lacking
traceability information is one of the main issues that contribute to project failure
and difficult to sustain. Unsatisfactory traceability information results in the need
for costly and painstaking tasks of manual recovery and maintenance of traceability
links [9]. These manual tasks may be frequently required depending on how normally
software systems evolve or are maintained.
As a result, the literature proposed methods, techniques, and tools to improve
traceability links automatically [2, 8, 9, 15]. Researchers used information retrieval
(I.R.); techniques, to recover traceability links between high-level documents, e.g.,
requirements, instruction booklet pages, and plan documents, and low-level docu-
ments, e.g., source code and UML diagrams. I.R. techniques compute the textual
similarity between each two software artifacts, e.g., the source code of a class and
a requirement [2–5, 8]. A high textual similarity means that the two artifacts most
likely share numerous concepts and that; therefore, they are likely linked to one
another.
628 V. M. Sale et al.

5 System Block Diagram

The proposed work is based on the IR-based RTAs process is typically divided into
three main steps. Figure 1 shows the IR-based RT links revival process.
First, every the textual information with the requirements and source code is taken
out and preprocessed by splitting terms [3–5], removing stop words and remaining
words are then steamed to its grammatical origin. Second, all the stemmed terms
are weighted using a term weighting system. Last, an I.R. technique calculates the
similarity between requirements and source code documents [6]. Lastly, it creates
a ranked list of probable traceability links. An elevated comparison between two
documents shows a probable semantic connection between them.

5.1 Preprocessing

To generate traceability links, we remove all the identifiers from source code and
terms from requirements. In this, I.R. techniques are used as an engine to create links
between requirements and source code. I.R. techniques imagine that all documents
are in the textual format [8, 16]. To remove source code identifiers, a source code
parser is used. The parser throw-outs extra information, e.g., primary data types and
keywords, from the source code and gives only identifier names. The identifiers and
terms are removed by filtering, stopper, and stemmer process [10, 17].
The primary step is term splitting [2–4]. A text normalization step renovates
all upper-case letters into lower-case letters. This step eliminates non-textual, i.e.,
some numbers, mathematical symbols, brackets, etc., information and extra white
spaces, from the documents. Some identifiers/terms could be united with some special
characters, e.g., underscore, and/or Camel Case naming reunion. Therefore, divide
all the united terms to make them separate. For example, Hello India and hello India
are split into the terms “hello India” [2].

Fig. 1 System block diagram

An Effective Approach for Accuracy of Requirement … 629

The following step is the stop word removal [2, 8]. The input for this step is the
normalized text that could include some general words, e.g., articles, punctuation,
etc. These general words are measured as noise in the text because it does not be a
symbol of the semantics of a document. Hence, in this step, a stop word list is used
to eliminate all the stop words.
The next step is stemming. An English steamer, for example, would recognize the
terms “excel,” “excellence,” and/or “excellent” as based on the root “excel” [2, 5].
An I.R. technique calculates the similarity between two documents based on similar
terms in both documents. Still, due to different postfix, I.R. techniques would judge
them, e.g., add and addition are like two different documents, and the result would
be a low similarity between two documents. Thus, it becomes important to perform
the morphological investigation to exchange plural into the singular and to take back
infected forms to their morphemes [3, 18].
Following two main factors are considered important [2]:
1. Term frequency (T.F.): T.F. is often called home frequency. If a term appears
several times in a document, then it would be allocated higher T.F. than the
others.
2. Global frequency (G.F.): If a term appears in various documents, then the term
is considered global. It is also known as inverse document frequency (IDF).

5.2 Term Weighting/Assign Weights

An I.R. technique (IRT) changes all the documents into vectors to calculate their
similarities [12, 19]. To change documents terms into vectors, each term is allocated a
weight. A variety of schemes for weighting terms have been proposed in the literature.
Widely used weighting schemes are differentiated as probabilistic. In the following,
the term identifiers to refer to all source code entities, i.e., class name, method name,
variable name, and comments [5].
If a term comes out multiple times in single or multiple documents, then IRT
would propose that document as a relevant document to a query [5, 8, 11]. However,
multiple amounts of a term do not show that it is an important term.

5.3 IR Techniques

To create sets of traceability links, various I.R. techniques are used, to identify
concepts in the source code, carry out experiments using different I.R. techniques to
recover traceability links [6, 8].
630 V. M. Sale et al.

6 Proposed Algorithm

6.1 Data Processing

The data preprocessing is prepared to eliminate needless content from the text and
to find out the original form of the words [3, 5]. The preprocessing of the data is
completed by a valid method such as Stop word removal and Stemming to the data
composed of the customer.

6.2 Stop Word Removal

For work out, stop words are words that are filtered out preceding to, or following,
processing of text [3, 4]. Stop words are ordinary words that take less significant
meaning than keyword. These stop words are a few of the most common, short func-
tion words, such as the, a, an, is, at, which, that, and on, etc. Stop-word elimination is
the method of eliminating these words. To find out the words from a text, all needless
content must be removed, so it is needed to remove the stop words from the text put
into an array [5, 12].
Algorithm:
The following is an algorithm for stop word removal
1. Acquire the input
2. Establish the glossary of stop words
3. Divide factors into words
4. Assign new word list to store words
5. Collect outcome in the String Builder
6. Loop during the entire terms
7. Come again, string with words detached.

6.3 Stemming

Words get from the input of the data are created to be too sparse to be useful as
features for categorization as they do not simplify well. The presence of a large
number of inflexions of the same word, this is the common reason for stemming.
Hence, the origin form of the word is to be taking out as a feature [2, 3, 5].
Stemming is the method, for decreasing derived words to their origin form. Stem-
ming program is commonly known as stemming algorithms or stemmers [3]. Even
as writing the sentence for a grammatical basis, it contains various forms of a word,
for example, collect, collection, collecting and/or collected. In many circumstances,
it would be helpful for a finding for one of these words to revisit the word in the
An Effective Approach for Accuracy of Requirement … 631

set to take away the required content from a given sentence [3, 5, 20]. The goal of
stemming is to decrease variation form and sometimes derivationally related forms
of a word to a common base form [2, 5, 8].

For instance : car, cars, → car

Stemming algorithm:
The stemming algorithm consists of different steps of stemming applied sequentially.
Within each stage, there are various principles to select rules, such as choosing the rule
from every rule group that applies to the longest suffix. The algorithm of stemming
works as follows:
Rules Illustrations
S → cats → cat
EED → E.E. agreed → agree
(∗v∗)E.D. → plastered → plaster
(∗v∗)ING → cutting → cut
There are three main reasons for stemming algorithm, or stemmer, as follows.
The first reason of a stemmer is to cluster the words according to their theme.
Many words are the root from the same stem, and we can consider that they belong
to the same concept (e.g., act, actor, action) [3, 4, 8]. The different forms are created
by attaching affixes (prefixes, infixes, and/or suffixes) but, in English considering
only suffixes, as normally prefixes and infixes change the meaning of the word, and
a bit of them would lead to errors of bad topic resolve [18, 21].
The next reason of a stemmer is openly associated to the [3–5, 11] I.R. process,
as containing the stems of the words agree to some point of the I.R. process to be
better, among which we can stress the ability to index the documents according to
their theme, as their terms are clustered by stems or the extension of a query to obtain
to a greater extent accurate results.
The extension of the query permits it, for refining by replacing the terms, it covers
the related topics, which are also there in the collection [4, 8, 22]. This alteration
can be done routinely and obviously to users, or the system can propose one or more
superior method of the query.
Finally, the conflation of the words allocation the same stem leads to a decrease
of the vocabulary to be taken into the process, as the entire terms contained in the
natural input collection of documents can be decreased to a set of topics [3, 5, 6].
This directs to a decrease of the space needed to store the formation used by an I.R.
system and after that also lightens the computational weight of the system.
632 V. M. Sale et al.

7 How to Represent Traceability

Program conception occurs in a bottom-top way, a top–bottom way, or a mixture

of them [6, 20]. Developers use knowledge throughout program comprehension,
from domain-oriented knowledge to common programming knowledge. Traceability
linkage between source code and sections of the documentation, e.g., requirements,
aid both top-down and bottom-up comprehension [1]. Traceability linkage between
the requirements of a system and its source code is useful in reducing comprehension
effort.
Requirement traceability is defined by [2, 9], “the capability to demonstrate and
go after to the life of a requirement, in both onward and toward the back direction”.
This traceability information also supports in software maintenance and evolution
tasks. For traceability links, it is essential to represent them in a form that is suitable
for its purpose [1]. The different ways (traceability matrices, graphical models, cross-
references) exist to represent traceability links, which are also supported by tools.
a. Traceability matrices: Traceability links are represented in matrix form. The
traceability matrix is the association between, horizontal and vertical dimensions
are the values in the matrix stand for links between the artifacts in the matrix
[23].
b. Graphical models: Entity-Relationship Model (ERM), various UML diagrams
support the representation of traceability links embedded in the different
development models [23, 24].
c. Cross references: Traceability associations between different parts are repre-
sented as links, pointers, or annotations in the text [23].

8 Data Analysis and IR

The information retrieval problem is to trace appropriate documents in a docu-

ment collection based on a user’s query, which is often some keywords relating
an information need, although it could also be an example relevant document.
1. Precision: It is the ratio of intersection of relevant and retrieved to retrieved.
|{Relevant}Ç{Retrieved}|
Precision =
|{Retrieved}|

2. Recall: It is the ratio of intersection of relevant and retrieved to relevant.

|{Relevant}Ç{Retrieved}|
Recall =
|{Relevant}|
An Effective Approach for Accuracy of Requirement … 633

9 Experimental Results

Our goal is to develop software which should be able to trace the source code
according to requirement by applying steaming algorithm. Traceability links are
useful in decreasing understanding effort linking the requirements of a system and
its source code. The traceability information is also useful to the software mainte-
nance and development tasks. For instance, with traceability links, a user is able to
simply trace what software parts must be changed for the development of a new
requirement. As, show in Table 1, comparison for different files with similarity IR
accuracy, time complexity.
The experimental outcomes of the work are given below. Here we have given
the result of the proposed work, which implements the data preprocessing with a
stemming algorithm to identify the similarity between requirement and source code
(Figs. 2 and 3).

Table 1 Comparison of different files with similarity IR accuracy, time complexity

Input similarity (%) IR accuracy (%) Time complexity (%)
File 1 80 File 1 86 File 1 86
File 2 76 File 2 89 File 2 89
File 3 94 File 3 91 File 3 91
Table 1.1: input similarity Table 1.2: IR accuracy Table 1.3: time complexity

Fig. 2 Comparison graph of precision

634 V. M. Sale et al.

Fig. 3 Comparison graph of recall

The proposed system can be evaluated based on traceability links such as I.R.,
and VSM. The analysis of the proposed system results and discusses observations
from the experiential evaluation of the Trustrace.

10 Data Set Quality Analysis

The experiential evaluation of the proposed system supports the Trustrace united
with IRT is helpful in increasing the precision and recall principles of some baseline
requirements traceability links.
Achieving Target
Program conception arises in a bottom to top way, a top to bottom way, or mixture
them. Developers use knowledge throughout program conception, from particular
domain knowledge to common programming knowledge. The linkage of traceability
among source codes with requirements of different stakeholders associated with
goals, help both top-down and bottom-up conception, which reduces the under-
standing exertion. Requirements Traceability is related to requests and different
improvement factors. The possible objectives are:
1. The best excellence of the system which is in the development phase;
2. To classify the product which are in the development phase with different
artifacts’; and
3. The facility to deal with change.
The requests and correlation with all the different parts related to it, such as
different modules of the system, investigation of outcomes, execution of the system
with different test cases, test actions, test outcomes and requirements of the system
all kinds should be traced. For any software development, the essential task is, a
An Effective Approach for Accuracy of Requirement … 635

designer has to know the development background, such as the development plan-
ning, purpose, how to implement, and the associations among the different parts
using any available documentation. Structures of system understanding take place in
a start to end way, an end to start way, or grouping of both. Different types of data,
ranging from particular knowledge to broad programming knowledge can be used
throughout program conception.
Requirement traceability helps software engineers to trace the requirement from
its emergence to its fulfilment. Traceability links are useful in decreasing developing
effort between the source code of a system and its requirements.
The traceability of requirements is a point of a technique. The requirements and
their sources are obviously correlated to each other and the objects formed in the
life cycle of system development depend on the requirements. During design phase
traceability of requirement helps to keep track of when the modification of request is
executed before a system is redesigned. Traceability is also able to give notification
about the validation, significant conclusion, and postulation of requirements.

11 Future Work

In future work, a broader survey, a more diverse set of stakeholders for multiple
domains. Different repositories and traceability recovery activities will be used for
replicating the study of traceability. The overall goal is to enhance the state-of-the-art
by addressing the shortcomings in a current research that uses I.R. technique to mine
vague software repositories and to persuade future researches to follow the lead.
Furthermore, a number of methods to get better the accuracy of the proposed trace-
ability improvement technique. An appropriate cost-beneficial model for traceability
needs further study as regards the costs of sustaining traceability and regarding the
benefits of traceability for other tasks in the process of software development. This
should be just the focal point of future work.

12 Conclusion

The traceability is the most important factor and precious from different points of
view for any software project development. We have developed a source code for
the system that source code can be traced and become identical with the requirement
analysis because we develop a source code as per the requirements.
From the results, it shows that our approach is performing better than other
techniques. It provides better consistency compared to existing systems. Thus,
this technique is useful for to minimize the execution time during the software
development.
For development of any software, requirement traceability plays a vital role in the
maintenance of software. Creating traceability links manually is one of the costly
636 V. M. Sale et al.

and lengthy works. The major finding of evaluation that is the proposed system
generates accurate traceability links during development with high precision and
recall. Requirements specification for requirements traceability is formed alongside
all the investigations, which drives both their direction and focus.

References

1. Shinde SK, Sale MVM (2015) A survey on accuracy of requirement traceability links
during software development. In: International conference on innovations & technological
developments in computer, electronics and mechanical engineering (ICITDCEME), 2015
2. Ali N, Guéhéneuc Y-G, Antoniol G (2011) Trust-based requirements traceability. In: Sim SE,
Ricca F (eds), Proceedings of 19th IEEE international conference program comprehension, pp
111–120, June 2011
3. Divya KS, Subha R, Palaniswami S (2014) Similar words identification using naive and TF-IDF
method. Int J Inf Technol Comput Sci 11:42–47. Published Online Oct 2014 in MECS https://
www.mecs-press.org/. https://doi.org/10.5815/ijitcs.2014.11.06. Copyright © 2014 MECS
4. Khetade PN, Nayyar VV (2014) Establishing a traceability links between the source code
and requirement analysis. A survey on traceability. In: International conference on advances in
engineering and technology (ICAET-2014) 66, Page (IOSR-JCE), e-ISSN: 2278-0661, p-ISSN:
2278-8727, pp 66–70
5. Ali N, Guéhéneuc Y-G, Antoniol G (2011) Factors impacting the inputs of traceability recovery
approaches. In: Zisman A, Cleland-Huang J, Gotel O (eds) Springer, Berlin
6. Muthamizharasi S, Selvakumar J, Rajaram M (2014) Advanced matching technique for
trustrace to improve the accuracy of requirement. Int Innov Res Sci Eng Tech (ICETS’14)
3(1)
7. Ramesh B, Jarke M (2001) Toward reference models for requirements traceability. IEEE Trans
Softw Eng 27(1):58–93
8. Ali N, Guéhéneuc Y-G, Antoniol G (2013) Trustrace: mining software repositories to improve
the accuracy of requirement traceability links. IEEE Trans Softw Eng 39(5):725–741
9. Ramesh B, Edwards M (1993) Issues in the development of a requirements traceability model.
In: Proceedings of the IEEE international symposium on requirements engineering, pp 256–259
10. Poshyvanyk D, Guéhéneuc Y-G, Marcus A, Antoniol G, Rajlich V (2007) Feature location
using probabilistic ranking of methods based on execution scenarios and information retrieval.
IEEE Trans Softw Eng 33(6):420–432, June 2007
11. Hayes JH, Antoniol G, Guéhéneuc Y-G (2008) PREREQIR: recovering pre-requirements via
cluster analysis. In: Proceedings of 15th working conference on reverse engineering, pp 165–
174
12. Antoniol G, Canfora G, Casazza G, Lucia AD, Merlo E (2002) Recovering traceability links
between code and documentation. IEEE Trans Softw Eng 28(10):970–983
13. Marcus A, Maletic JI (2003) Recovering documentation-to source-code traceability links
using latent semantic indexing. In: Proceedings of 25th international conference on software
engineering, pp 125–135
14. Panichella A, McMillan C, Moritz E, Palmieri D (1993) When and how using structural
information to improve IR-based traceability recovery
15. Gotel O, Finkelstein (1994) An analysis of the requirements traceability problem. In:
Proceedings of the first international conference on requirements engineering, pp 94–101
16. Schwarz H, Ebert J, Winter A (2009),Graph-based traceability: a comprehensive approach.
Softw Syst Model
17. Heindl M, Biffl S (2005) A case study on value-based requirements tracing. In: Proceedings
of the 10th European software engineering conference. Lisbon, Portugal
An Effective Approach for Accuracy of Requirement … 637

18. Lindvall M, Sandahl K (1998) How well do experienced software developers predict software
change? J Syst Softw 43(1998):19–27
19. Cleland-Huang J, Chang CK, Ge Y (2002) Supporting event-based traceability through
high-level recognition of change events. In: Proceedings of the 26th annual international
computer software and applications conference on prolonging software life: development and
redevelopment. Oxford, England, pp 595–602
20. Winkler S, Pilgrim J (2010) (2010), A survey of traceability in requirements engineering and
model-driven development. Softw Syst Model 9(4):529–565
21. Clarke S et al (1999) Subject oriented design: towards improved alignment of requirements,
design, and code. In: Proceedings of the 1999 ACM SIGPLAN conference on object-oriented
programming, systems, languages, and applications, Dallas, TX, pp 325–329
22. Von Knethen A (2002) Change-oriented requirements traceability. Support for evolution of
embedded systems. In: Proceedings of international conference on software maintenance, pp
482–485
23. Wieringa R (1995) An introduction to requirements traceability. Technical Report IR-389,
Faculty of Mathematics and Computer Science
24. Lehman M, Ramil J (2003) Software evolution—background, theory, practice. Inf Process Lett
88(1–2):33–44
Clustering of Fruits Image Based
on Color and Shape Using K-Means
Algorithm

Vidya Maskar, Kanchan Chouhan, Prashant Bhandare, and Minal Pawar

Abstract Clustering Algorithm is an unsupervised machine learning technique.

Unsupervised Machine learning well defined unknown patterns in data. Clustering
is the process of organizing data into specific groups. Clustering is mainly deals with
finding a structure or pattern in a collection of uncategorized data. Clustering has
been studied for a long time by many researchers with different methods. In this
project work, we deal with object clustering problem. Here, we proposed a K-means
algorithm. K-means algorithm is the easiest and prominent unsupervised machine
learning algorithm. We apply the K-means algorithm for grouping the fruits as per
these features. The experiment conducted on small clustering dataset and results
found that the K-means algorithm s help for clustering object.

Keywords Unsupervised machine learning · Clustering · K-means algorithm

1 Introduction

Machine learning is a branch of statistics and computer science in which the machine
predicts analyzing the data provided to it. In machine learning, there are various
techniques and algorithms: Supervised learning, Unsupervised learning, Clustering,
Classification, etc. [1]. The clustering algorithm is an unsupervised machine learning
technique [2]. Unsupervised machine learning well defined unknown patterns in data.
Here, we study the object clustering problem by using the K-means algorithm. K-
means algorithm is the elemental and most popular unsupervised machine learning
clustering algorithm. In clustering algorithm K variable represented by data set of
cluster group. In all cluster, centroid is define. The centroid is a data point present
at the center in all clusters. The tricks to define the centroids far away from each
other so that variation is less. After all this process, each and every data point in

V. Maskar (B) · P. Bhandare · M. Pawar

SVERI’s College of Engineering (Polytechnic), Pandharpur, Maharashtra, India
e-mail: [email protected]
K. Chouhan
SVERI’s College of Engineering Pandharpur, Pandharpur, Maharashtra, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 639
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_64
640 V. Maskar et al.

the cluster is assigned to the nearest centroid. The K-means clustering algorithm is
very expedient in grouping new data [3]. Clustering analysis is the organization of
patterns into clusters based on similarity. Intuitively, patterns within a valid cluster are
more similar to each other than they are pattern belonging to a different cluster. It is
important to understand the difference between clustering and discriminant analysis.
Clustering is unsupervised classification and discriminant is supervised classifica-
tion [4]. In supervised classification we provided with collection of labeled data and
the unsupervised learning is provided data into unknown pattern [5]. A learner can
cash in of examples (data) to capture characteristics of interest of their unknown
underlying probability distribution. Data are often seen as examples that illustrate
relations between observed variables. A major focus of machine learning research is
to automatically learn to acknowledge complex patterns and make intelligent deci-
sions supported data. The ability of a program to find out from experience that’s, to
switch its execution on the idea of newly acquired information.

2 Literature Review

Number of clustering algorithms which continue to appear in the literatures; each

new clustering algorithm performs slightly better than the existing ones on a specific
distribution of patterns. Data mining techniques, Machine learning algorithms, and
Image Processing techniques are applied to clustering the object. Different authors
have proposed different methods for the classification of the object.

2.1 Unsupervised Machine Learning for Clustering

the Infected Leaves based on the Leaf-colors

This proposed paper help to understand the clustering techniques which is a defined
by Dr. P. Tamilselvi and Dr. K. Ashok Kumar. In this paper, the plant leaves are
grouped based on the colors in the leaves. Totally, three categories are specified to
represent the leaf with more green, leaf with yellowish shades and leaf with reddish
shades. The task is performed using image processing. The leaf images are processed
in the sequence such as image pre-processing, segmentation, feature extraction and
clustering [1].

2.2 Color, Shape and Texture based Fruit Recognition System

The Paper present an automated system for classification of fruits. Ruaa Adeeb
Abdulmunem Al-falluji have been proposed an automated system for classification
Clustering of Fruits Image Based on Color … 641

of fruits was constructed using an ordinary camera. The fruits classify based on
this color, Shape, and Texture. Gray Level Co-occurrence Matrix (GLCM) is used
to calculate texture and Best accuracy was achieved by using the Support Vector
Machine (SVM). In this paper all the fruits were analyzed on the basis of their color,
shape and texture and then classified using different classifiers to find the group that
gives the best accuracy [2].

2.3 Fruits Classification Using Image Processing Techniques

Chithra and Henila have been applied image processing system for fruits classifi-
cation. Machine vision system for fruit identification and disease identification of
fruits or vegetables under the research in the agriculture industry. Machine vision
system for fruit identification and disease identification of fruit or vegetable is one
among the current topic that is under research in the agriculture industry. As a part of
this current research area, this fruits classification using image processing techniques
was developed [6].

2.4 A Short Survey on Data Clustering Algorithms

Won have been proposed different types of the clustering algorithm, clustering
paradigms used in the machine learning system. They have been applied to a wide
range of domains; for instance, bioinformatics, speech recognition, and financial
analysis [3].

2.5 Supervised Machine Learning Approaches: A Survey

Muhammad and Yan have been applied the basic idea of machine learning and its
various techniques. The algorithms of machine learning are useful in areas where
deploying explicitly written in high- speed performance. Machine Learning can be
considered as a subset of Artificial Intelligence. Machine learning has many more
algorithm; those algorithms can be seen as building blocks to make computers learn
to behave more intelligently by somehow generalizing rather than just storing and
retrieving data items from a database system and other applications [4].
642 V. Maskar et al.

2.6 Comprehensive Review on Supervised Machine Learning

Algorithms

Gianey and Choudhary have been proposed the idea about machine learning and in
that paper also learn machine learning techniques and tools. In machine learning
empowers system with the ability to learn automatically and get better with expe-
rience without being explicitly programmed. In machine there are many algorithms
are useful in areas where deploying explicitly designing application with high speed
performance are unfeasible [5].

2.7 Machine Learning Techniques and Tools: A Survey

Obulesu et al. have been proposed the current world lots of electronic data is generated
in every field using machine learning algorithms and techniques. In that using the
different techniques like as the decision tree, Support Vector Machine, Naive Bayes
Classification, Linear Regression, etc. In the current world many data is generated
in each and every field. These data contains useful information to predict the future
[7].

2.8 Object Clustering by K-means Algorithm with Binary

Sketch Templates

Mei have been proposed Object clustering is a very challenging unsupervised learning
problem in machine learning and pattern recognition. In this paper defined the object
clustering technique it is an unsupervised machine learning technique. The experi-
ment conducting on a small clustering dataset shows that the k-means with binary
sketch templates for object clustering is very promising and the learned mixture
of templates is also meaningful for understanding the results. Number of clustering
algorithms which continue to appear in the literatures; each new clustering algorithm
performs slightly better than the existing ones on a specific distribution of patterns.
Data mining techniques, Machine learning algorithms, and Image Processing tech-
niques are applied to clustering the object. Different authors have proposed different
methods for the classification of the object [8].
Clustering of Fruits Image Based on Color … 643

3 Methodology

Methodology is the flow the content of the project in that we define the actual process
of the project. In order to classify the procedure of work. In that chapter we define
the system architecture, proposed algorithm, Flowchart and simple example of the
clustering technique.

3.1 System Architecture

In Fig. 1 shows that the how the system are work. Take picture of fruit image using
the camera. Images are there then extract the feature of that image as like color and
shape. After extracting feature we divide that image either shape or color wise. Using
the K-means clustering algorithm for distinguish fruits image. When we apply this
technique we got result in color or shape wise.

3.2 Proposed Algorithm

The goal of this algorithm is to find groups in data, with the number of groups
represented by the variable K. The algorithm works step by step to assign each data

Fig. 1 System architecture

644 V. Maskar et al.

point to one of K groups based on the features that are provided. Here, Data points
are based on feature similarity.

Algorithm
(1) K points are placed into the object data space representing the initial group of
centroids. At beginning decide the value of k.
(2) Each object or data point is assigned into the closest k. Randomly put initially
partition that classifies the data into k cluster. Assign value of k randomly.
(i) Take the first k training sample value which is single-element of clusters.
(ii) Assign each of the remaining training sample value to the cluster with
the nearest centroid. After each assignment, recomputed the centroid of
the gaining cluster.
(3) After all objects are assigned; the positions of the k centroids are recalculated.
Take each sample value in sequence and compute its distance from the centroid
of each of the clusters. If a sample is not currently in the cluster with the closest
centroid, switch this sample to that cluster and update the centroid of the cluster
gaining the new sample and the cluster losing the sample.
(4) Steps 2 and 3 are repeated until the positions of the centroids no longer move. K
centroids are created randomly (based on the predefined value of K). K-Means
allocates every data point in the dataset to the nearest centroid (minimizing
Euclidean distances between them), meaning that a data point is if the decided
data point is closet to centroid than it is considered. Then K-means recalculates
the centroids by taking the mean of all data points assigned to that centroid’s
cluster, hence reducing the total intra- cluster variance in relation to the previous
step. In the K-means refers to averaging the data and finding the new centroid.
The algorithm step by step process between steps 2 and 3 until some criteria is
met (e.g. the sum of distances between the data points and their corresponding
centroid is minimized, a maximum number of iterations is reached, no changes
in centroids value or no data points change clusters).

3.3 Flowchart

Initially, the K-means algorithm is works based on a data point. In Fig. 2 shows how
the K-Means clustering algorithm works.
In that first decide the number of cluster means K = 1, K = 2 or K = 3 etc.
Choose the centroid of an object and select the distance of that particular object. If
the distance is a minimum grouped cluster in the same cluster otherwise it gives in
another cluster. And repeat the process until a got current cluster.
Clustering of Fruits Image Based on Color … 645

Fig. 2 Flowchart

3.4 Simple Example

Figure 3 is the example of K-means clustering algorithm. In figure, different data

points are present which convert in the group with color-wise. It is a simple example
of the k-means algorithm.

3.5 Feature Extraction

Color feature and shape feature are extracted in complete feature vector.
646 V. Maskar et al.

Fig. 3 Simple example

Color Features: The color feature that can be used for showing different color
or mixer of color. In color feature there are different type’s colors these are also
known as RGB color. The main purpose of the this color model is for the sensing,
representation, and display of images in electronic systems as per the different color,
such as televisions and computers, though it has also been used in conventional
photography RGB is a device detect or reproduce a given RGB value differently [2].
Shape Features: The second feature that can be used as an attribute for recognition
is shaped. In machine learning, a feature is a piece of information that is relevant
for solving the computational task related to a certain application. Shape Features
may be specific structures in the image such as points, edge or objects. The shape
feature used is area, several pixels in an image is used for determining the area of a
particular object [2].

4 Results and Analysis

This section presents the computational analysis for investigating the performance
of the k-means algorithms for image clustering. Given algorithm is coded in python
language. The performance evaluation in the project is done for the purpose of
minimize fitness function.
Initialization
Select k cluster centroids randomly Here we apply k = 3 as shown in Fig. 4 and
stated as per Table 1.
Clustering of Fruits Image Based on Color … 647

Fig. 4 Initialization

Table 1 Calculate distance table

S. No X Y Distance from Closet Distance Distance Distance Color
from_1 from_2 from_3
1 12 39 26.925824 1 26.925824 56.0803 56.727418 r
2 20 36 20.880613 1 20.880613 48.373546 53.150729 r
3 28 30 14.142136 1 14.142136 41.761226 53.338541 r
4 18 52 36.878178 1 36.878178 50.990195 44.102154 r
5 29 54 38.118237 3 38.118237 40.804412 34.058773 b

Assignment
Assign each point to its nearest centroid cluster in this stage we assign the each and
every point nearest to centroid and calculate the distance of the point. After assigning
values to respective points we have Figs. 5, 6, 7 and 8.

5 Conclusions and Future Work

In this report, we proposed a method based on clustering of the k- mean algorithm. K-

means clustering is one of the most popular clustering algorithms and usually apply
for the solving clustering task using different dataset. In this we use K-means for
improving the quality of formed clusters. This report has been proposed an approach
648 V. Maskar et al.

Fig. 5 Assignment

Fig. 6 Update centroid of

the data point

using K-Means clustering for classified fruits images from their colors and shapes.In
feature we use different technique of machine for fruits clustering. And also generate
different module for clustering which is less time consuming.
Clustering of Fruits Image Based on Color … 649

Fig. 7 Repeat assignment

step

Fig. 8 Result of clustering

dataset

References

1. Gianey HK, Choudhary R (2017) Comprehensive review on supervised machine learning

algorithms. In: International conference on machine learning and data science
2. Obulesu O, Mahendra M, ThrilokReddy M (2018) Machine learning techniques and tools:
a survey. In: Proceedings of the international conference on inventive research in computing
650 V. Maskar et al.

applications ICIRCA, IEEE Explore Compliant Part Number: CFP18N67-ART; ISBN:978-1-

5386-2456-2
3. Won K-C (2015) A short survey on data clustering algorithms. arXiv:1511.09123v1 [cs.DS]
4. Muhammad I, Yan Z (2015) Supervised machine learning approaches: a survey. ICTACT J Soft
Comput 5(3)
5. Tamilselvi P, Ashok Kumar K (2017) Unsupervised machine learning for clustering the infected
leaves based on the leaf-colours. In: 2017 third international conference on science technology
engineering & management (ICONSTEM)
6. Chithra PL, Henila M (2019) Fruits classification using image processing techniques. l J Comput
Sci Eng 7. Open Access Reserach Paper
7. Al-falluji RAA (2016) Color, shape and texture based fruit recognition system. J Adv Res
Comput Eng Technol (IJARCET) 5(7)
8. Mei X (2006) Object clustering by K-means algorithm with binary sketch templates. 978-1-
5090-3484-0/16/$31.00 ©2016 IEEE.
Modern Education Using Augmented
Reality

Vishal V. Bandgar, Ajinkya A. Bahirat, Gopika A. Fattepurkar,

and Swapnil N. Patil

Abstract Augmented reality (AR) is an virtual experience of the real world, some-
times across multiple sensory modalities, including visual, auditory, and haptic
(Umeda et al. in ICIIBMS 2017, track 3: bioinformatics, medical imaging and neuro-
science, Okinawa, Japan, pp 146–149, 2017 [1]) Augmented reality is expounded
largely to synonymous terms: mixed reality and computer-mediated reality. So we are
thinking to vary over the instruction into present day or computerized training with
the help of Augmented Reality. Nowadays everybody has mobile phone addition-
ally children utilize the cell phones, with the help of mobile phone we are going
to propose our framework. Considering a book there are numerous pictures yet
kids can’t envision these items in genuine so with our application we’ll founded
a framework which tells this specific picture and also the comparing data to find out
understudies effectively and adequately with no issues.

Keywords Augmented reality · Digital school · Education · Machine learning

1 Introduction

The first value of augmented reality is that the manner during which components
of the digital world blend into an individual’s perception of the important world,
not as a straightforward display of information, but through the combination of
immersive sensations, which are perceived as natural parts of an environment. The
earliest functional Augmented Reality systems that provided immersive mixed reality
experiences for users were invented within the early 1990. Presently coming toward
our exploration, right off the bat we learned about the present framework, there are
numerous ways we are able to utilize Augmented Reality in education however we are

V. V. Bandgar · S. N. Patil
SKNCOE Vadgaon, Pune, India
A. A. Bahirat (B)
SVERI’s COE (Poly), Pandharpur, India
G. A. Fattepurkar
VIT, Pune, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 651
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_65
652 V. V. Bandgar et al.

going for the change of 2D pictures into 3D for better comprehension of understudies
with the assistance of virtual teaching. The essential estimation of enlarged the reality
is that the way where segments of the advanced world mix into a human impression
of this present reality, not as a simple showcase of knowledge, however through the
incorporation of vivid sensations, which are seen as normal pieces of a situation.

2 History and Background

The most punctual useful AR frameworks that gave vivid blended reality encounters
to clients were designed within the mid-1990s, beginning with the Virtual Fixtures
framework created at the U.S. Commercial increased reality encounters were first
presented in stimulation and gaming organizations. During this manner, increased
reality applications have adjoin business ventures, for instance, instruction, inter-
changes, prescription, and stimulation. In training, substance can be gotten to by
filtering or review an image with a telephone or by utilizing marker less AR tech-
niques. A model significant to the event business is an AR protective cap for devel-
opment laborers which presentations data about building destinations. The mixture
of AR technology with the academic content creates new style of automated applica-
tions and acts to boost the effectiveness and attractiveness of teaching and learning for
students. Key technologies and methods are discussed within the context of education
[1]. Some of the areas where Augmented Reality plays very vital role:

2.1 Medical Treatment

Medicinal imaging has clad to be imperative for determination and treatment in our
general public. We’ve got been adequately pushing past the cutoff points of innova-
tion we had with the progression of PC innovation. As of late, medicinal preparing
frameworks utilizing computer generated realities, VRs, and blended reality, MR [2]
are created. Medicinal school understudies are required to require various restora-
tive trainings in process and achieve therapeutic aptitudes. It’s normal that utilizing
the PC vision innovation encourages improvement in therapeutic training. We built
up a restorative preparing framework utilizing enlarged reality, AR. By and huge,
there are two sorts of AR usage, marker based AR and markerless based AR [2]. In
our framework, the marker based AR was utilized in light of the actual fact that the
previous technique is easier than the latter, and therefore the last strategy is altogether
constrained by computational power. During this paper, we present our medicinal
preparing framework utilizing AR, including 3D anatomical articles and manipulable
natural interface.
Modern Education Using Augmented Reality 653

2.2 Academic Research

Current academic research within the MR sector has targeting technology and user-
interface research but there’s a hunt gap in studying user experiences and decision-
making, technology advancement and application development side-by-side so as to
grasp their value-in-use. The user-value drivers are numerous and will drive appli-
cation development. So far, the key value drivers are identified to be cost-saving
through out-of-home and out-of-office access, total control and high level of person-
alization, going beyond reality, personal efficacy experiences.From now of view, the
main challenge for both VR and AR technologies is to convince users that the added
value is high enough to compete with the present systems and offerings in desktops,
notebooks, tablets, smartphones and related video and game-like applications [3].

2.3 Robotic 3D Printer

As a planner makes another model utilizing RoMA AR CAD editorial manager,

highlights are developed simultaneously by a 3D printing automated arm having
an identical structure volume. The somewhat printed physical model at that time
fills in as an unmistakable reference for the planner as she adds new components
to her structure. RoMA’s proxemics-motivated handshake component between the
creator and therefore the 3D printing mechanical arm enables the planner to rapidly
obtrude upon printing to urge to a printed territory or to point out that the robot can
assume full responsibility for the model to complete the method of printing. RoMA
gives clients an opportunity to coordinate true requirements into a structure quickly,
enabling them to form proportional unmistakable curios or to broaden existing items.
We reachable showing the qualities and confinements of our present structure [4].

3 Design Issues

Presently allow us to see the present framework. The accompanying image demon-
strates the use of Augmented Reality in education however there’s just transformation
of 2D picture into 3D picture. So there are some shots that some understudies might
not comprehend what’s happing. So now here in Fig. 1 as you’ll see that the image of
brain in 2D is just converted in 3D. Some students could also be capable of learning
from this but some student are {going to be are} bit confused in some part so to
address this issue, we are going to propose an new system. In the wake of discussing
current structure presently it is a great opportunity to debate our proposed framework.
Our primary objective is to accomplish a full computerized help to class and easy
learning for understudies. Presently consider a bit school from rustic zone there aren’t
654 V. V. Bandgar et al.

Fig. 1 Augmented reality in

education [5]

any such an offices accessible like Augmented Reality. So our point is to manufacture
an application for those understudies from rustic zone for viable learning.
Now consider an example of a state like Maharashtra there are around 65,000
government schools are digitized but it includes digital equipment like LCD screens,
projectors, and computer machines so we are thinking grade up to create schools
digitized. Right off the bat, we want to assemble an application for our framework
we are assuming to make android application with assistance of android studio on
the grounds that the majority normal cell phones are android following making an
application now its opportunity to check it. After some review we understand that the
instruction utilizing Augmented Reality the reality is simple to grasp and effectively
able to get a handle on. For more splendid understudy there’s enormous chance to
analyze the unexplored things splendid understudies yet as poor getting a handle on
limit understudies will have the choice to induce a handle on the ideas. Presently we
are going to see some models, now consider a book of images containing pictures
of old creatures with the help of Augmented Reality understudies can see the real
3D picture of that creatures. On the off chance that an understudy can’t comprehend
the nearby planetary group with the help of increase reality we will demonstrate that
understudy the 3D liveliness of close planetary system so these are some instances of
Augment Reality in instruction. Enlarged and computer generated reality are seeing
increasingly noticeable usage in study hall and instructive settings. The configuration
and even the world of the training procedure may be changed through AR within the
years to come. Remembering that 71% of people matured 16–24 all have a mobile
phone, AR may be the subsequent huge thing in instruction. In Fig. 2 the Apple 3D
model is shown supported a target image and display it on our application.

4 Implimentation

There are main two stages in application first stage is Scanning and second one is
Displaying 3D module with some data. Figure 3 describes the system architecture.
Modern Education Using Augmented Reality 655

Fig. 2 Augmented reality in

3D printing

4.1 Scanning

In the implemented system Vuforia plugin is used to map a perfect image targets.
Like most of the work is done by vuforia like the image scanning matching with
the database. To use vuforia need to register on it’s official site and get the key for
development. Also vuforia provides its own database so after storing data we able to
get the Database file also.

4.2 Displaying 3D Models

For display purpose or 3D rendering the IDE used is Unity 3D, after getting the
development key and database it’s easy to handle in unity. Firstly need to create some
Targets or ImageTargets on that we can assign the images from Vuforia Database.
And then we have to take specific 3D model and place on specific ImageTarget, so
when we scan that target we will get desired 3D model. Also we playing specific
audio on specific targets.
656 V. V. Bandgar et al.

Fig. 3 System architecture

5 Result and Analysis

From this project the training curiosity of scholars are increased and that they learn
efficiently at school similarly as reception. It creates a learning by doing environment.
The following image describes how system is working when we scan ImageTarget
the specific 3D model is displayed (Fig. 4).
The following are some other screenshots of implemented system. If the Letter
K Image Target is Scanned the the Kangaru 3D model displayed and so on (Figs. 5
and 6).
Modern Education Using Augmented Reality 657

Fig. 4 Augmented reality

own created module

6 Conclusion

At every new inflection in communication technology, the marketing and advertising

community brought their existing models with them in their first efforts to adapt to
the new medium. Radio, television, Internet, Web, and Mobile have all seen this
progression play out AR will likely follow that model too. Eventually there’ll be
innovations round the ways an advertiser can bend the messaging, strategies and
tactics of AR marketing to convey to focus on audiences a persuasive, emotionally
compelling narrative leading to some measurable level of name affinity and buying
outcome. The new AR ecosystem won’t happen all told the ways foreseen but it
definitely will happen in a number of those ways. This next big thing in marketing is
bound to render obsolete many existing marketing and advertising business models
and replace them with new ones. It’s reasonable to conclude that marketers must be
able to adopt this new ecosystem or see competitors gaining brand recognition and
market share at their expense if they don’t have a go at it first.
658 V. V. Bandgar et al.

Fig. 5 K ImageTarget
scanned
Modern Education Using Augmented Reality 659

Fig. 6 S ImageTarget
scanned

References

1. Botella C, Baños RM, Perpiñá C, Villa H, Alcaniz M, Rey A (1998) Virtual reality treatment of
claustrophobia: a case report. Behav Res Ther 36(2):239–246
2. Umeda R, Seif MA, Higa H, Kuniyoshi Y (2017) A medical training system using augmented
reality. In: ICIIBMS 2017, track 3: bioinformatics, medical imaging and neuroscience, Okinawa,
Japan, pp 146–149
3. https://www.emergingedtech.com/2018/08/multiple-uses-of-augmented-reality-in-education/
4. Peng H, Briggs J, Wang C-Y, Guo K, Kider J, Mueller S, Baudish P, Guimbretière F, In RoMA:
ınteractive fabrication with augmented reality and a robotic 3D printer. Cornell University Ithaca,
NY, USA {hp356, jeb482, cw776, kg344, fvg3}@cornell.edu
5. Augmented Reality in Education, https://thinkmobiles.com/blog/augmented-reality-education/
OSS Features Scope and Challenges

M. K. Jadhav and V. V. Khandagale

Abstract Today oss becomes most popular in software community due to its wide
use. But some issues are also identified by users while using such softwares. Due
to free availability major advantage of such software is even if vendors stop devel-
oping such softwares other programmers can contribute to develop such softwares.
Nowdays there is big challenge to outline the policy for usage of open source soft-
ware. This paper includes challenges faced in the use of oss, features of oss, and
emerging applications of oss.

Keywords OSS (open source software) · Stakeholders · Forums

1 Introduction

Closed source and open source software are two genres of software. But open source
software are become most popular nowadays. Any user can detect open source soft-
ware code, can refine and can enrich the code to meet the growing needs of stake-
holders. Normally users do not have wisdom of code of application they use. Appli-
cation working can be directed by developers as they can operate the code due to its
free availability. Due to code availability programmers can point out flaws in it, can
add features to it in order to increase usability of that software. Most popular open
source software are linux, libreoffice etc. Oss allows programmers collaboratively
work on software to share their knowledge of code, to make changes into the source
code in order to add required features into their project. But open source software
also have issues in its use.

M. K. Jadhav (B) · V. V. Khandagale

SVERI’S College of Engineering (Polytechnic), Pandharpur, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 661
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_66
662 M. K. Jadhav and V. V. Khandagale

2 Literature Survey

In this section we see how computer programmers need to develop open source soft-
ware for their real time projects. In early days of information technology, computers
are with already installed os and software applications. But as the software is going
to become standalone product, software vendors restricts to access their code freely
by other in order to avoid inspection of their important software details. So computer
users are become dependent on software vendors for maintenance, technical support
and inclusion of additional features of software. For this they are pay high cost to
software vendor. As a result of this there was great need of accessing the source
code freely in order to meet the increasing needs of application by modifying and
enhancing that code. So, in early 70s open source software movement was start.
Trainor [1, 2] Richard M Stallman play important role in this movement. He
introduced concept “copy left” which consists General Public License for software
which allows anyone to freely use, distribute and modify the source code at no charge.
Also introduction of Linux by Linus Torvalds forwards the open source software
movement in 1991 [2]. He implements Linux os on Intel based computers. 25% of
servers and 2.8% of desktop computers were use Linux in 2002 [3]. Still companies
were not interested for investing in open source software. In 1998 Netscape releases
source code of its web browser which motivates free software developers to improve
quality of software by working collaboratively together in a way that no company
can achieve this individually. Major open source software project Apache web server
software occupies 60% of the market [4, 5].

3 Applications of OSS

3.1 Libra Office

It is open source alternative to Microsoft office. It becomes popular because

it preserves Microsoft office format with support for documents, spreadsheets,
databases, presentations and mathematical formulae [6].

3.2 VLC Media Player

It is created by open source development group known as videogame project. Audio

and video file in any format can be opened by VLC media player. It is also used for
streaming media such as online radio stations [6].
OSS Features Scope and Challenges 663

3.3 Shotcut

It is open source program for advanced video editing. It supports no destructing audio
and video editing without any quality loss [6].

3.4 Linux

It is open source operating system. It is released in 1991. Most of servers are devel-
oped in Linux. It becomes popular due to its security and less vulnerability to virus
[6].

4 Features of OSS

4.1 Community Control

Most of people prefer loss because they have more control on it. As source code is
available to all freely people can examine the code to ensure that code do not doing
anything that they do not want it to do. They can make changes in source code as per
requirement of application [7].

4.2 Security

As source code of loss is freely available to all, many people are working on same
piece of code. So if any error is there in loss they all are working on that error to
remove that error. As many eyes on same piece of software, it is less vulnerable to
threats [8].

4.3 Stability

Usually many people prefer loss for long term projects. As source code of oss is
distributed publicly, so even if original vendor [8] of project stop working on source
code then also other programmers can rely on that software tools. So oss gives more
stability as compared to closed source software.
664 M. K. Jadhav and V. V. Khandagale

4.4 Free Use

Non-programmers also get advantage from oss as they can use software for any
purpose as they wish which is not happen in case of closed source software users
have to use software in the way as its vendor thinks [8].

5 Challenges of OSS

5.1 Lack of Customer Support

It is found that one issue with oss is lack of customer support. One way to overcome
this problem is by giving answers to the questions asked by end user in forums. If
someone found something about software he must share it in forum [9].

5.2 Lack of Policy

It is challenge for business owner to outline a policy for their open source usage.
Without policy developers may use any component for any purpose which may cause
issues down the line. So establishing clear written policy is the best way to avoid
such issues.

5.3 Threats in Security by Mystery Resources

As source code of oss is freely available to community, it may be risk for business
owner that hackers may make unwanted changes in that code. So security of that oss
may be challenged [10].

5.4 Consumption is More Than Security

As oss is freely accessible to all, people are more consuming the software for their
business but people are less contributing towards further development of software
[9].
OSS Features Scope and Challenges 665

References

1. Trainor C (2009) Open source, crowd source: harnessing the power of the people behind our
libraries. Program-Electron Lib 43:288–298
2. https://medium.com/fossmec/the-history-of-free-and-open-source-software-for-the-third-gen
eration
3. https://portal.stupica.com/Linuxpedia/Linux_adoption.html
4. https://yourbusiness.azcentral.com/list-different-servers-3142.html
5. Appelbe B (2003) The future of Open Source Software. J Res Pract Inf Tech 35:227–236
6. https://www.techradar.com/in/best/best-open-source-software
7. Kemp R (2009) Current developments in Open Source Software. Compute Law Secure Rev
25:569–582
8. https://opensource.com/resources/what-open-source
9. https://dzone.com/articles/executive-insights-on-the-current-and-future-state-2
10. Abdullah R, Lakulu M, Ibrahim H, Selamat MH (2009) The challenges of Open Source
Software development with collaborative environment 2009. In: International conference on
computer technology and development
Text Summarization and Dimensionality
Reduction Using Ranking and Learning
Approach

Dipti Bartakke, Santosh Kumar, Aparna Junnarkar, and Somnath Thigale

Abstract Because of the exponential increment of records on the web, clients need
all the related information in a difficult situation. Finding the significant data from
such huge information is examining undertakings, hence the data recovery turns out
to be progressively crucial for looking through the applicable information viably
These oversee via Automatic content summarization. It is a procedure that perceives
the significant focuses from all the important reports to display a compact summary.
The proper text summarization and dimensionality reduction (TSDR) of summarized
text can lead to a notable reduction in accessing time for the input elements. The
proposed method produces the summarization task by dimensionality reduction with
rank (TSDRR) using training methodology. The consequence of a string of words
in a data text is appraised by the assistant of the PageRank algorithm. The subject
is first pre-processed to tokenize the determinations and perform stemming opera-
tions. Then descent-based text summarization involves selecting determinations of
high connection according to the level of the report depends on word also determi-
nation characteristics and established them collectively to make a report. The test
results determine that this approach has more reliable production than other current
classifications.

Keywords Text summarization · Extractive method · Natural Language

Processing · DUC 2002 Dataset (Document Understanding Conferences)

1 Introduction

In Human, Summarization one has a tendency to abridge a single article by summing

up the most vital thoughts and requesting to guarantee they are coherent. Notwith-
standing for humans, this undertaking would take a ton of work. Synopses created
by two distinct individuals would normally be unique. Distinctive people may have

D. Bartakke · S. Kumar · A. Junnarkar (B)

Mahrishi University of Information Technology, Lucknow, India
S. Thigale
CSE Department SVERIs CoE, Pandharpur, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 667
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_67
668 D. Bartakke et al.

an alternate view of what is imperative. This motivated the requirement for having
an Automatic Summarizer that can play out the activity in less time and with the
slightest exertion. This drove for the examination of Automatic Summarization to
begin over 50 years back [1]. Text Summarization when all is said in done is the
way toward abridging a single article or an arrangement of related ones by summing
up the most imperative occasions, ensuring the occasions succession is coherent by
requesting them chronically. Then again automatic Text Summarization is the way
toward delivering an abbreviated variant of a text by the utilization of computers [2].
The synopsis ought to pass on the key commitments of the text. In Automatic Summa-
rization, there are two primary methodologies that are extensively utilized Extractive
and Abstractive. The primary strategy, the Extractive Summarization, removes up to
a specific utmost the key sentences or passages from the text and requests them in a
way that will create a coherent outline. The removed units contrast starting with one
summarizer then onto the next. Most summarizers utilize sentences as opposed to
bigger units, for example, sections [3, 4]. Extractive Summarization techniques are
the attention strategy on Automatic Text Summarization. The other strategy, Abstrac-
tive Summarization, includes more language-dependent tools and Natural Language
Processing (NLP) innovation. Such summarizers can incorporate words not present
in the first document [5], illustrated in Fig. 1.
Automatic summarization is the means of decreasing a collection of data compu-
tationally, to produce a subset (a summary) that describes the common significant or
appropriate message inside the opening this show in Fig. 1. The aim of this research
work is to proposed automatic text summarization using dimensionality reduction.
The objectives are:
1. To study and analyze different summary readability and learning-based text
summarization and reduction methods.
2. To assess the importance and applications of text summarization and reduction.
3. To design a novel approach for accurate, quick, and meaningful abstractions
extraction from the large text.
4. To propose a novel Text summarization framework based on effective abstrac-
tion extraction.
5. To propose the text reduction technique and combined with text summarization
to enhance accuracy.

Fig. 1 Automatic summary generation process

Text Summarization and Dimensionality Reduction … 669

6. To implement and evaluate the proposed methods with state-of-art methods.

During this activity, we propose a programmed subject rational summarizer to
create a release version.
The record has five divisions. Section 2 presents an amazing of the literature
reviews. Division III describes the recommended procedure for pre-processing prac-
ticing feature descent and the principal report production. District IV presents the
preliminary design, outcomes, and get the measure with a human summary. Section 5
proposes completion and future role.

2 Literature Review

In [1] the author performed a methodology for clustering utilizing WordNet

including lexical chains. A qualified WordNet-based grammatical similarity model
was intended for word understanding disambiguation, and lexical connections were
manipulated to extract core semantic comments that reveal the topic of reports. It
worked four difficulties in paper clustering. The difficulties were disambiguating the
polysemous and corresponding words, mastering high dimensionality, restricting the
representation of clusters, and allowing relevant descriptions for the produced clus-
ters. They produced a hybrid design for resolving these difficulties in text clustering
at the related time.
In [2] exhibited an extricate text summarization operation that can improve human
analysts in receiving individual capacity and PICO benefits of full-text PDF state-
ments. The practice is constituted of two foremost elements: determination ranking
including keyphrase removal. Indecision ranking, they confirmed that applying a
robot training classifier on full text to emphasize determinations completed evenly
or entirely than eliminating rights and summaries.
In [3] the author entered computerized originated extricate summaries; under this
following item, instruction created reports collected by different English profes-
sors continued practice. The automatic reports were collected utilizing the Fuzzy
classification and Vector method. Managing the Rouge examine method, they exam-
ined the manually-produced minutes and the computerize-produced ones. Rouge
evaluation of produced reviews registered the perfection of reports designed by
people over the computerize generated summaries. On the other sizes, the asso-
ciation among the produced surveys determined that summaries displayed by the
Fuzzy opinion occurred leading further satisfactory including acceptable analyzed
to reports provided through the Vector method.
In [6] author completed the valuable sentence scoring techniques for computerized
extractive text summarization algorithms to be conditional on the variety of document
individuals requires to summarize, the scope of reports, the variety of information
practiced, and their organization. Variant sequences of a string of words scoring
estimate yield inconsistent events both in the position of the summaries received
and the period transpired in producing them. The recommended augmentation of the
670 D. Bartakke et al.

exposition was obtaining the best sequences of a string of words securing systems
during three brands of records: news, blogs, furthermore articles.
In [7] the founder completed the recommended entrance to make the summariza-
tion assignment by unsupervised training approach. The meaning of a decision in
a data extract was decided by the representatives of the Simplified Lesk algorithm.
As an online semantic dictionary WordNet was practiced. Original, this procedure
assesses the importance of everything individual decisions from a text indepen-
dently utilizing the Simplified Lesk algorithm and provides them in contracting
methods according to their importance. Following, according to the provided section
of summarization, a remarkable character of a string of words was picked from that
required list.
Wikipedia articles obtained provided as input to the operation and extractive text
summarization was performed [8] by recognizing text features and obtaining the
decisions accordingly. The text was first pre-processed to tokenize the decisions
and conduct stemming progress. Then score the sentences using the modified text
features. Two strategies presented were preparing the commands already in the text
and identifying synonyms. These comments along with the conventional techniques
were practiced to obtain the sentences. The numbers were practiced to classify the
sentence to be in the summary outline or not with the maintenance of a neural network.
In [9] author clarified “Enhanced connected and discrete multi-objective shred
swarm optimization for text summarization” Examining down this colossal mate-
rial would be useful in the original drive for numerous assistants. So text summa-
rization structures end up being significantly in separating this huge substance.
The summaries are created in perspective on basic highlights using multi-target
approaches where sufficient composing isn’t open. Noteworthy imprisonments of
substance layout systems are versatility and execution.
In paper [10] a query-oriented text Summarization procedure used for deci-
sion extraction was introduced. In the extricate summarization method, the various
instructive decisions in the text were recognized and decided to revisit in the report.
To recognize decisions containing relevant information, a set of suitable pieces were
extricated of that text. Whatever the selected pieces of the decisions were also suit-
able, the numerous informative decisions were more precisely recognized and the
property of the produced report increases. They have determined that the application
of more becoming features commands to enhanced summaries created. To estimate
the automatically created summaries, the ROUGE pattern has been practiced.
In [11] author recommended the application of the Firefly algorithm (FF) through
that descent of reports of original Arabic records. The recommended procedure was
connected with two evolutionary methods that use genetic algorithms and harmony
exploration. A compound of informational and grammatical numbers was practiced.
The suggested FF approach was applied to decide the optimal sub-path from appli-
cant ways in the design. Several routes in the plan describe a hopeful summary, so
the FF algorithm engaged to select the optimal track. The EASC Corpus moreover
significant ROUGE toolkit obtained practiced for the examiner of the recommended
entrance.
Text Summarization and Dimensionality Reduction … 671

In [12] authors completed a Vietnamese text summarization organization based

on a sentence extraction procedure utilizing a neural network for training merge
decreasing dimensional characteristics to defeat the value while creating term
positions and decrease the calculation complication. The advanced techniques to
exact summary decisions in math Vietnamese text practicing controlled training
approaches toward neural systems, besides, people also complete particular means
of element compression to decrease processing complexity and that added certain
efficient arrangements occurred accessibly.
In [13] authors performed a summarization method for text documents utilizing
the grammatical similarity between decisions to exclude the repetition from the
text. Semantic association scores were measured by mapping the string of words on
a grammatical term doing random indexing. Random indexing, in association with
distinct semantic space algorithms, offers a computationally effective way of implicit
dimensionality modification. It includes reasonable vector estimates such as interest.
It, therefore, presents an efficient way to calculate connections between words,
sentences, and documents. Random indexing has denoted managed to calculate the
semantic similarity transcripts of decisions and graph-based ranking algorithms have
been contracted to compose an essence of the distributed text.
Automated Text Summarization (ATS) does “decreasing the reference document
within a smaller version while maintaining its learning text and overall interest”.
In [14] suggested automatic feature-based extractive traveling forward document
summarizer to develop the connection through developing the comprehensibility
of the report text. It decreases the presented information record handling social
scoring including social ranking that is it performs progressing perceptive summary.
Specifications of a statement give reference information and approval visual checking
of the record to discover the pursuit substance. It positions the requests heading
astute and picks top n sentences from different heading any place n depends against
pressure proportion. The last heading sentiment rundown created by this method
was an aggregation of an outline of explicit headings. Since the heading insightful
rundown controls the equivalent relationship of sentences from each heading, it
conquers the intelligent hole of the conceptual content. Likewise, it improves the
general application and attitude of the rundown content. The results of the analysis
unmistakably show that heading savvy summarizer delivers better over the head
summarizer, Ms-word summarizer, free summarizer, moreover Auto summarizer.
In [15] proposed a manifold education based on the system, entitled Mutual
message preserving mapping (MIPM), to investigate the low-dimensional, commu-
nity and common knowledge preserving embeddings of high element informa-
tion. The MIPM was practiced to produce a transient summarization operation for
powerful handing the report connected beside a result preserving liberty.
In [16] empirically examined infrequent well-developed complex education calcu-
lation for performance in extractive expression summarization, including isometric
feature mapping (ISOMAP), locally linear embedding (LLE) furthermore Lapla-
cianeigenmaps. Preliminary evidence establishes that the different systems moment
from his summarization structure exceeds various living widely-used acting under
supervision. systems and very equivalent among remarkable start-of-the-art outlines.
672 D. Bartakke et al.

3 Methodology

Text Summarization implies an existing range of analysis identified together with the
Information Retrieval including Natural Language Processing districts. Text Summa-
rization occurs frequently remaining worked in the marketing department so as the
data mining, Communications communication management about text databases, for
web-based data retrieval, in word Processing accessories. Various methods contrast
in the presence of their weight formulations. Automatic text summarization holds
an essential measure for information management responsibilities. It explains the
predicament of choosing the usual relevant divisions of some text. In this paper
proposed method, TSDR without ranking sentences system performed high-quality
summarization requires sophisticated NLP techniques. The following steps to text
summary generation are:
1. Examine the Data
2. Divide Text into Sentences
3. Divide Text into Sentences
4. Text Preprocessing Vector Representation of Sentences
5. Similarity Matrix Preparation
6. Applying the PageRank Algorithm
7. Summary Extraction.
Title word: An unusual rate remains provided before the decision if this includes
information transpiring in that string of words as the principal text of the report is
dispatched via the right information. That characteristic is measured as regards:

N o.o f titlewor d
Scor e(S1) =
N o.o f totalwor d

Sentence Length: Reduce the string of wordsthat are likewise small such as time-
line or designer handles. During each string of words, the normalized period of a
string of words denotes determined because:

T otal N o.o f W or ds ∈ sentences

S2 = 
T otal N o.o f W or ds lar estsentences

Sentence Position: The sentences happening beginning in the section have the
most powerful score. If the paragraph becomes p sentences, the average of every
sentence is determined:
p
S3(Sentences1) =
p
4 3 2
S3(Sentences2, Sentences2, Sentences3) = , ,
5 5 5
Text Summarization and Dimensionality Reduction … 673

Numerical data: The string of words should statistical results can emulate relevant
numerical of the report and suddenly chosen for a report. Its amount is computed as:

T otal N o.o f N umericData ∈ sentences

S4 =
Lengtho f Sentences

Inflected words: These are field-specific information including the greatest poten-
tial relationship. The degree of the product of thematic reports that happens in a
string of words over the highest quantity of thematic information in a string of words
provides the number of individual characters as:

T otal N o.o f N umericdata ∈ sentences

S5 =
Max N o.o f I n f lectedW or ds

Sentence to Sentence Similarity: The symbol test system is applied to calculate

the correlation within individual sentence S furthermore each additional sentence.
The matrix [N][N] is created. N means the entire product of the sentence into the
record. The oblique components about a matrix do determined to zero because the
string of words should negatively be associated including itself. The relationship of
separately determination combination remains determined to be regarded:

[(Sx, Sy)]
S6 =
M AX [(Sx, Sy)]

where the values from x = 1 to N furthermore y = 1 to N.

Phraseweight: The ratio of summation of phrase(p) repetitions of all expressions
in a string of words over the maximum of summation conditions of all decisions
in a document provides the number of term weight feature. It is determined by the
resulting comparison.

P Sx
S7 = 
M AX P Sx

Frequent Nouns: The relevant string of words implies that string of words that
include the greatest amount of individual names. Its composition is provided through,

N o.o f N oun ∈ sentences Sx

S8 =
lengtho f sentences Sx

The architectural flow diagram (explained in Fig. 1) of the process of text summary
generation and dimensionality reduction represent in Fig. 1. In the first phase are
preprocessing that acquired the input document from datasets and these read data
transform separate each sentence into words using segmentation and tokenization
strategy. After that remove the unnecessary word and highlighted important words
store in the matrix array list form. The analysis of input data and match similar
674 D. Bartakke et al.

keywords including each word assign a decreasing string of words to score for
word and sentences put together to generate a summary by select proper words
and sentences.
Pre-processing: In the pre-processing state is much unnecessary and excessive infor-
mation instant or noisy and misleading data, and then knowledge acquired during
the preparation period is more complicated. Data preparation and filtering measures
can take a substantial amount of processing time. Data pre-processing comprises
cleaning, Instance collection, normalization, conversion, feature extraction, and
reading, etc. The output of data pre-processing is the final training set.
There are four levels in preprocessing. The division is a strategy for dissemi-
nating a given report into sentences. Prevent words are released from the content.
Stop words are generally happening words, for example, ‘an’ a’, the’ that produces
less significance and controls clamor. The Stop words are predefined and gathered
in a cluster. Tokenization will circulate the information content into discrete tokens.
Accentuation checks, spaces, and word eliminators are the words uncovering noto-
rieties. Word Stemming is acknowledged to all novices’ words into its root structure
by pushing its prefix and postfix for a relationship with different words.
Feature Extraction: In feature descent state obtained from the original data, but
the selected content is not changed in any way. Representatives of selected content
introduce key-phrases that can be related to “tag” or index a text report or key
sentences (including headings) that collectively compose an abstract, and illustrative
images or video segments, as declared above. For text, ancestry is comparable to
the method of skimming, where the report (if available), headlines and subheadings,
personalities, the first including last paragraphs of a section, and optionally the front
and last decisions in a paragraph are read ere one wants to read the complete report
in detail.
The writing report is forwarded by collection, D = {S1, S2, …, Sk} wherever
Si imports a string of words checked in report D. The report is controlled to high-
light descent. The significant information and string of words characteristics to be
applied are determined. That commitment applies highlights such as Title infor-
mation, Sentence season, Sentence place, numerical data, Term weight, sentence
agreement, the occurrence of Thematic reports furthermore Proper Terms.
Semantic analysis: Semantic analysis is the responsibility of guaranteeing that the
reports and statements of a schedule are semantically accurate, i.e., that their purpose
is clear and compatible with how to control buildings and data types are assumed to
be practiced.
Latent Semantic Analysis (LSA): LSA does a mathematical representation of
term mode which connects the grammatical relationship among sections of textual
knowledge. LSA is produced to promote the benefits of message retrieval arrange-
ments by utilizing the “semantic” text of messages in a query as engaged in giving
through name equivalent. LSA circumvents the difficulties of synonymy, in which
various words can be applied to represent the likewise grammatical thought.
Text Summarization and Dimensionality Reduction … 675

Sentences Selection: Based upon the string of words includes all the string of words
are ranked in settling position. Sentences with the most leading number are selected
as a release summary. Certainly, the string of words is completed in the sequence
they arrive in the primary record.
Based on sentence records all the decisions are ranked in decreasing string of
words. Sentences with the most distinguished score are selected as a document
summary. Ultimately, the sentences, in summary, are organized in the string of words
they transpire in the original document.

Algorithm 1: Pre-Processing Algorithm

Step 1: Create the instruction information set D including n records

D = {d1, d2, . . . , dn}.

Step 2: Division of each sentence and assign each sentence to a symbol. The
selected sentences marked as x and non-selected sentences marked as y moreover
the text of each sentence into datasets.
Step 3: Determine the amount of topic name data meaningful in both marked x
string of words and y string of words. Assemble these utilities through the database
report.
For scoring sentences, rundown portrayal approaches figure scores dependent on
how well the sentence communicates a portion of the significant subjects in the report
or joins the points. Marker portrayals score sentences dependent on joining proof
from various pointers utilizing machine learning strategies. For instance, PageRank
doles out loads to sentences by applying stochastic techniques to the chart portrayal
of the archive. Approaches for choosing sentences for the synopsis incorporate best n,
maximal minor importance, and worldwide determination. In the principal approach,
the top n positioned sentences that have the ideal rundown length are chosen. In
the second, sentences are chosen to utilize an iterative insatiable system, which
recomposes sentence scores after each progression. The third methodology chooses
sentences utilizing improvement techniques.

Algorithm 2: PageRank Algorithm

1. Value of each topic word to match into sentences
2. The relation between any two sentences is accepted as an equivalent to the report
spread probability
3. The comparison scores are put in a square matrix, related to the matrix X applied
for PageRank
4. The first level would be to concatenate all the text included in the document
5. Split the document into the sentences
6. In the next measure, we will discover vector representation (word embeddings)
for each sentence
7. Similarities among sentence vectors are then computed and collected in a matrix
8. The correlation matrix is then transformed into a graph, with sentences as
vertices and comparison scores as edges, for sentence rank estimation
676 D. Bartakke et al.

9. Subsequently, a specific product of top-ranked sentences forms the conclusive

summary.

Algorithm 3: Feature Score (Weight)

Here, in PageRank algorithm, we calculate the score of each Sentence of each word
as denoted by S: sentences, W: word, I: token, M: Matrix, p: paragraph.
Input S: List of sentences; V: word tokenized;
Output W’: a set of sentences and value corresponding to it
Initialization W’ = ϕ; I = ϕ; P = ϕ; M = ϕ; x = 0;y = 0;

For i=1 to count(S) do

For k=1 to length(si) do
If (si[k] is noun)
x=x+I(si[k]);
I[k] ← x;
x=0;
For k=1 to length(si) do
If (si[k] is noun) calculate
y++;
M[k] ← y;
y=0;
P[k]= ←MATCH(si,d);

The suggested framework TSDRR centers around the extractive based worth posi-
tion utilizing the summarization strategy. We will talk about two new methodologies
for outlining the content. The primary strategy is displayed in Fig. 2 the recurrence
of the words dependent on the root type of the word, and furthermore the recur-
rence of its equivalent words present in the content. The subsequent strategy is
to recognize sentences containing references or references and give them a higher
weight. As appeared in Fig. 3 the summarization framework accepts contribution
as DUC2002 datasets, forms it and gives the rundown sentences. The information
record comprises of crude information to be prepared by the framework. The initial
step of the summarization is known as tokenization, which separates the sections into
sentences and every one of these sentences is additionally broken into a lot of words.
The information structure got after tokenization of sections is a rundown, containing
every component as a sentence and information structure got after tokenization of
sentences is a rundown of the rundown, containing sets of words. Tokenization is
performed through example coordinating utilizing ‘customary articulations’. Infor-
mation acquired as a lot of words is additionally broke down and stop words or most
normally happening words, similar to articles, are expelled from the arrangement of
words.
Text Summarization and Dimensionality Reduction … 677

Fig. 2 Process of generate summary and dimensionality reduction using the TSDR system

4 Results and Discussion

The analyses were executed applying a Python environment among an 8 GB RAM

core2 DUO Intel Processor. The outcomes are associated with the nature of art
evolutionary strategies and also TSDR outdoors regarding procedures practiced.
The dataset consists of examine in the region of document summarization called the
DUC called Document Understanding Conference. The experiment result for DUC
678 D. Bartakke et al.

Fig. 3 Architecture of generate summary and dimensionality reduction using the TSDRR system

2002 produced 60 reference sets. The particular set sustained documents, single-
document summaries, furthermore multi-document outlines/selections, including
attitudes described by many types of models such as issue positions, biograph-
ical collections, etc. The reports were prepared including and without “sentences”
designated as determined by a transcription of the single sentence severance soft-
ware utilized for DUC 2001. These datasets general summary of the report with a
length of around 100 words or more limited was conceived. (Whitespace-delimited
tokens). The coverage metric selected ranges into account including compensated
Text Summarization and Dimensionality Reduction … 679

conciseness. The summaries were produced undividedly of a perfect string of words

(Fig. 4).
In Fig. 2 shows the comparison results of recall, precision and f-score utilizing
the datasets ROUGE CNN, ROUGE blog, Fuzzy logic, and ANFIS. Table 1 shows to
perform the values of our proposed TSDR system having more efficient and reliable
as compared to the existing system.
Figures 5 and 6 observation graph represented to apply proposed TSDR and
TSDRR methodsutilizing DUC2001 and DUC2002 datasets. Its results show as
compared to our proposed TSDRR methods having an efficient score of recall, preci-
sion, and f-score. The system has a high recall and precision consequence experiment
with manual evaluation outcomes. We have used two different data sets.

Fig. 4 Comparison graph results using recall, precision and f-score

Table 1 Comparison results

Datasets/methods Recall Precision F-score
using precision, recall, and
F-score TSDR (DUC2002) 0.7934 0.7981 0.7921
TSDR (DUC2001) 0.7841 0.7956 0.7818
ROUGE (CNN) [6] 0.73 0.36 0.69
ROUGE (Blog) [6] 0.77 0.63 0.48
Fuzzy logic 0.42 0.2979 0.3479
ANFIS 0.77 0.7906 0.7801
680 D. Bartakke et al.

Fig. 5 Comparison graph using proposed TSDR and TSDRR methods

Fig. 6 Comparison graph using proposed TSDR and TSDRR methods

5 Conclusion

In this paper proposed the text summarization and dimensionality reduction TSDR
and advanced TSDRR with a ranking of sentences. In the extractive summarization
Text Summarization and Dimensionality Reduction … 681

strategy, the most educational sentences in the content are recognized and chose to
go to the outline. To distinguish sentences containing significant data, a lot of fitting
highlights are extricated from the content. The result shows the proposed system
TSTR and TSDRR are more efficient and reliable as compared to the ROUGE, Fuzzy
logic and ANFIS system. In assessment, we have utilized estimates like accuracy,
review, and F-score which are utilized essentially with regards to data recovery. Later
on, we mean to utilize more summarization-explicit systems to quantify the efficiency
of our scheme.

References

1. Wei T, Lu Y (2015) A semantic approach for text clustering using WordNet and lexical chain.
Expert Syst Appl 42
2. Bui DDA, Del Fiol G (2016) Extractive text summarization system to aid data extraction from
full text in systematic review development. J Biomed Inform 64
3. Kiyoumarsi F (2015) Evaluation of automatic text summarizations based on human. Proc Soc
Behav Sci 192:83–89
4. Johna A (2015) Vertex cover algorithm based multi-document summarization using information
content of sentences. Proc Comput Sci 46:285–291
5. Babara SA, Patil PD (2015) Improving performance of text summarization. Proc Comput Sci
46:354–363
6. Ferreira R, Freitas F, Cabral L de S, Lins RD, Lima R, Franca G, …, Favaro L (2014) A context-
based text summarization system. In: 2014 11th IAPR international workshop on document
analysis systems
7. Pal AR, Saha D (2014) An approach to automatic text summarization using WordNet. In: 2014
IEEE international advance computing conference (IACC)
8. Hingu D, Shah D, Udmale SS (2015) Automatic text summarization of Wikipedia articles.
In: 2015 international conference on communication, information & computing technology
(ICCICT)
9. Priya V, Umamaheswari K (2018) Enhanced continuous and discrete multi-objective particle
swarm optimization for text summarization. Received: 12 February 2018/Revised: 20 March
2018/Accepted: 22 March 2018 Springer Science+Business Media, LLC, part of Springer
Nature 2018
10. Afsharizadeh M, Ebrahimpour-Komleh H, Bagheri A (2018) Query-oriented text summariza-
tion using sentence extraction technique. In: 2018 4th international conference on web research
(ICWR)
11. Al-Abdallah RZ, Al-Taani AT (2019) Arabic text summarization using firefly algorithm. In:
2019 amity international conference on artificial intelligence (AICAI)
12. Thu HNT, Huu QN, Ngoc TNT (2013) A supervised learning method combines with dimen-
sionality reduction in Vietnamese text summarization. In: 2013 computing, communications
and it applications conference (ComComAp)
13. Chatterjee N, Mohan S (2007) Extraction-based single-document summarization using random
indexing. In: 19th IEEE international conference on tools with artificial intelligence (ICTAI
2007)
14. Krishnaveni P, Balasundaram SR (2017) Automatic text summarization by local scoring
and ranking for improving coherence. In: 2017 international conference on computing
methodologies and communication (ICCMC)
15. Yang Z, Yao F, Fan K, Huang J (2017) Text dimensionality reduction with mutual information
preserving mapping. Chin J Electron 26(5):919–925
682 D. Bartakke et al.

16. Liu S-H, Chen K-Y, Chen B, Wang H-M, Hsu W-L (2017) Leveraging manifold learning for
extractive broadcast news summarization. In: 2017 IEEE international conference on acoustics,
speech and signal processing (ICASSP)
17. Meena YK, Gopalanib D (2015) Domain-independent framework for automatic text summa-
rization. Proc Comput Sci 48:722–727
Properties of Extended Binary Hamming
[8, 4, 4] Code Using MATLAB

N. S. Darkunde, S. P. Basude, and M. S. Wavare

Abstract The main aim of this paper is to study various properties of extended binary
Hamming [8, 4, 4] code, when we know its generator matrix. Using MATLAB, we
can study syndrome decoding, weight of a codeword, error correction and error
detection of binary Hamming [8, 4, 4] code.

Keywords Linear code · Generator matrix · Parity check matrix · Hamming

code · Syndrome decoding

1 Introduction

In the late 1940s, Claude Shannon has started development of information theory
and coding theory as a mathematical model for communication. At the same time, R.
Hamming found a necessity for error correction in his work on computers. Already
Parity checking was used to detect errors in the calculations of the relay-based
computers of the day, and Hamming realized that a more sophisticated pattern of
parity checking allowed the correction of single errors along with the detection of
double errors. The codes that Hamming put forth, were important for theoretical
and practical reasons. In [1, 2], algorithmic approach for error correction has been
studied for few codes.
Each binary Hamming code [3] has minimum weight and distance 3 and that of
extended binary Hamming code has weight and distance as 4. In this paper, we are
going to study some properties of extended binary Hamming [8, 4, 4] code.

N. S. Darkunde (B) · S. P. Basude

School of Mathematical Sciences, SRTM University, Nanded, Maharashtra, India
M. S. Wavare
Department of Mathematics, Rajarshi Shahu Mahavidyalaya, Latur, Maharashtra, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 683
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_68
684 N. S. Darkunde et al.

2 Preliminaries

2.1 Linear Codes [4]

Let Fq denote the finite field with q elements, where q is some power of a prime. A
linear [n, k]q —code is a k-dimensional subspace of Fqn . The parameters n and k are
referred to as the length and dimension of the corresponding code.

Example The subset C = {000, 001, 010, 011} of vector space F23 is [3, 2]2 linear
code. Similarly, C = {0000, 1100, 2200, 0001, 0002, 1101, 1102, 2201, 2202} is
[4, 2]3 linear code.

2.2 Hamming Distance and Hamming Weight [5]

Let x, y ∈ Fqn , The Hamming distance from x to y, denoted by d(x, y), is defined
to be the number of places at which x and y differ.

Example Consider x = 01010, y = 01101, z = 11101 in F25 , Then d(x, y) = 3,

d(x, z) = 4. For any x ∈ Fqn , the support of x, denoted by supp(x), is defined to be the
set of nonzero coordinates in x = (x1 x2, x3 , . . . , xn ), that is supp(x) := {i : xi = 0}.

For a [n, k]q code C containing at least two words, the nonnegative integer given
by min{d(x, y) : x, y ∈ C, x = y} is called minimum distance of C. It is denoted by
d(C).

Example For a code C = {0000, 1000, 0100, 1100} in F24 , we see that d(C) = 1.

Definition 1 [3] Let u be a positive integer. A code C is u—error-detecting if,

whenever a codeword incurs at least one but at most u errors, the resulting word is
not a codeword.

A code is exactly u—error-detecting if it is u—error-detecting but not (u + 1)

error-detecting.

Example Consider C = {000000, 000111, 111222} ⊆ F26 . This code is 2-error-

detecting, because changing any codeword in one or two positions does not result
in another codeword. In fact, C is exactly 2-error-detecting, as changing each of the
last three positions of 000000 to 1 will result in the codeword 000111(so C is not
3-error-detecting).

Theorem 1 [3] A code C is u—error-detecting if and only if d(C) ≥ u + 1, that is,

a code with distance d is an exactly (d − 1)—error-detecting code.
Properties of Extended Binary Hamming [8, 4, 4] Code Using MATLAB 685

Definition 2 [3] Let v be a positive integer. A code C is v—error-correcting if

minimum distance decoding is able to correct v or fewer errors, assuming that the
incomplete decoding rule is used. A code C is exactly v—error-correcting if it is
v—error-correcting but not (v + 1)—error-correcting.

Example Consider C = {000, 111} in F23 . It is easy to see that, C is 1-error-

correcting.

Theorem 2 [3] A code C is v -error-correcting if and only if d(C) ≥ 2v + 1, that is

a code with distance d is an exactly (d − 1)/2—error correcting code, where x
denote the greatest integer less than or equal to x.

Definition 3 (Dual Code) [3] Given a [n, k]q code C in Fqn , the subspace

C ⊥ = {x = (x1 , x2 , x3 , . . . , xn ) ∈ Fqn : x.c = 0 for all c = (c1 , c2 , . . . , cn ) ∈ C},

We have following properties of C and C ⊥ .
1. |C| = q dim(C) , i.e. dim(C) = logq |C|;  
2. C ⊥ is a linear code and dim(C) + dim C ⊥ = n;
 ⊥ ⊥
3. C = C.

2.3 Generator Matrix and Parity-Check Matrix

Definition 4 [5] A matrix G of order k × n is said to be the generator matrix for a

[n, k]q code C, if its rows form basis for C.

Definition 5 [4] A parity-check matrix H for a code C is a generator matrix for C ⊥ .

2.4 Extended Binary Hamming [8, 4, 4] Code

Let G = (I4 |A) where I4 is a 4 × 4 identity matrix and A is a 4 × 8 matrix given by

⎛ ⎞
10000111
⎜ 01001011 ⎟
A=⎜ ⎟
⎝ 00101101 ⎠
00011110

The binary linear code generated by matrix G is called extended binary Hamming
[8, 4, 4] code.
686 N. S. Darkunde et al.

2.5 Syndrome Decoding [3]

An efficiency of decoding technique works well, when length n of a given code is

small, but it can take a more time when, n is very large, so this time can be saved
by using the syndrome to identify the coset from which the word is taken. In [3], the
procedure of syndrome decoding has been demonstrated.
Step 1: Let w be received word in the transmission and for this received word w, first
compute the syndrome of w denoted by Syn(w) which is given by, syn(w) = w H T ,
where H , is parity check matrix of a given code.
Step 2: After constructing Syndrome look up table, we will find the coset leader u
next to the syndrome, syn(w) = syn(u).
Step 3: Finally decode the received word w as v = w − u.
Now, let us study the properties of extended binary Hamming [8, 4, 4] code using
MATLAB [6].

3 Properties of Extended Binary Hamming [8, 4, 4] Code

Using MATLAB

Communication fails due to the error in the communication channel. In this process,
receiver receives the original message with error in it. If the system knows how much
error has come with original message then that error can be removed. It is difficult
task to correct the errors in the communication; hence we have used MATLAB to
accomplish this task.

3.1 MATLAB Program for Syndrome Decoding Using

Extended Binary Hamming [8, 4, 4] Code
Properties of Extended Binary Hamming [8, 4, 4] Code Using MATLAB 687

3.2 MATLAB Program for Encoding Message Using

Extended Binary Hamming [8, 4, 4] Code
688 N. S. Darkunde et al.

4 Conclusion

Once we know the generator matrix of any linear code, using MATLAB we can have
many of the things that can be discussed. In this paper, we have studied extended
binary Hamming [8, 4, 4] code with the help of MATLAB. The parity check matrix in
standard form, for extended binary Hamming [8, 4, 4] code is calculated and using it,
syndrome of a particular codeword is calculated. We have encoded the message and
decoded it correctly by using MATLAB. Using MATLAB, it has been verified that,
the distance of an extended binary Hamming code is 4 and it is 1-error-correcting
code.

References

1. Patil AR, Darkunde NS (2019) Algorithmic approach for error-correcting capability and
decoding of linear codes arising from algebraic geometry. In: Information and communication
technology for competitive strategies. Springer, Berlin, pp 509–517
2. Patil AR, Darkunde NS (2017) On some error-correcting linear codes. Asian J Math Comput
Res 56–62
3. Ling S, Xing C (2004) coding theory-a first course, 1st edn. Cambridge University Press
4. Huffman WC, Pless V (2010) Fundamentals of error-correcting codes. Cambridge University
Press
5. MacWilliams FJ, Sloane NJA (1977) The theory of error-correcting codes. Elsevier
6. Pratap R (2017) Getting started with MATLAB: a quick introduction for scientists and engineers,
7e. Oxford University Press.
Identification of Fake News on Social
Media: A New Challenge

Dhanashree V. Patil, Supriya A. Shegdar, and Sanjivini S. Kadam

Abstract Because of the uncommon development of data on the web, it is getting

difficult to decode reality from the bogus. Accordingly, this demonstrates the serious
issue of phony news. This test thinks about past and current strategies for counterfeit
news identification. The executed framework manages the uses of NLP (Regular
Language Handling) methods for recognizing the ‘Phony News’, that is, misin-
forming reports that originate from the offensive sources. Just by developing a
model dependent on a tally victimiser (utilizing word counts) or a (Term Recur-
rence Backwards Record Recurrence) tfidf lattice, (word counts comparative with
how frequently they are utilized in various articles in your dataset) can just contact
you up until this point. Yet, these models don’t concentrate on significant charac-
teristics like word requesting and setting. It is truly practical that two articles that
are comparative in their promise will be totally extraordinary in their importance.
Restricting the phony news is the exemplary content order venture with a direct
hypothes. So a proposed chip away at bunching a dataset of both phony and genuine
news enlist a Credulous Bayes classifier to make a model to characterize an article
into phony or genuine dependent on its words. Right now two techniques are utilized
Credulous Bayes, Bolster Vector Machine (SVM). The standardization system is a
fundamental advance for refining information before utilizing AI strategies to order
information.

Keywords AI · NLP · Fakenews

1 Introduction

Fake news poses various problems in several media from derisive stories to fake news
and policy information preparation. Fake news causes huge issues for our culture. A
fabricated story by default is “fake news” but the discourse of social media latterly

D. V. Patil (B) · S. A. Shegdar · S. S. Kadam

SVERI’s College of Engineering, Pandharpur, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 689
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_69
690 D. V. Patil et al.

Fig. 1 Genaralized flow

diagram

shifts its meaning. A number of them are currently using the phrase to refute the
evidence against their most common points of view.
Today, detecting fake news on social media is a major challenge in the world.
First, fake news is deliberately published to misinform readers into believing false
facts, making it impossible to identify news content based on it. But, to help distin-
guish it from the real news, we need to provide contextual details, such as user
social interactions on social media. Second, leveraging this additional information
is non-trivial in and of itself as the social connections of consumers of false news
generate data that is immense, incomplete, unstructured and noisy. The Internet and
social media made it much easier and more convenient to access news information.
Internet users may also follow the events of their interest in online mode, and mobile
device spreading makes this process much more convenient. Many scientists think
machine learning will solve the fake news problem. There’s an explanation for this:
artificial intelligence algorithms have recently started to work much better on a lot
of classification issues (image recognition, voice detection, and so on) because the
equipment is cheaper and there are bigger data sets.
Figure 1 shows general flow of followed algorithm.

2 Literature Review

What are false news? The intentional propagation of disinformation by conventional

news media or by social media is fake news. False knowledge spreads out abnormally
rapidly. False news can also become impalpable from objective coverage, because it
spreads too quickly. Users will download newspaper articles from blogs, exchange
information, re-share information from others and the false information by the end
of the day the false information has gone so far from its original site that it becomes
impalpable from real news [1].
Identification of Fake News on Social Media: A New Challenge 691

A. Existing System

Today, for days, various academic bodies as well as the government working on the
topic of machine learning techniques for fraud detection have concentrated mainly on
classifying user comments and publicly accessible social media messages. In 2016,
the issue of deciding on fake news has been the subject of specific attention within
the literature during the American presidential election.
Conroy et al. [2] outline several approaches that seem optimistic towards the
aim of correctly identifying the misguiding news. They note that straightforward
content-related n-grams and shallow piece-of-speech (POS) tagging have well-tried
lean for the classification task, usually failing to account for crucial discourse data.
Preferably, these methods have been shown useful only in tandem with more complex
ways of analysis. Feng et al. [3] can attain 85–91% accuracy in deception related
classification tasks using online review corpora.
B. Proposed System

Because of the multifaceted nature of phony news location in web-based social

networking like Twitter, Facebook, and so on., It is clear that an attainable technique
must contain numerous parts of handling the issue precisely. This is the reason the
proposed technique is a mix of Help Vector Machines, Gullible Bayes classifier,
and semantic investigation. The proposed technique is entirely made out of Man-
made consciousness draws near, which is basic to precisely characterize between
the genuine and the phony, rather than utilizing calculations that can’t emulate
intellectual capacities.
The three-section strategy is a mix of AI calculations that subdivide into normal
language preparing and regulated learning systems techniques [4]. Albeit every one
of these methodologies can be just used to group and recognize counterfeit news,
to build the exactness and apply to the online networking space, they have been
converged into a coordinated calculation as a strategy for counterfeit news discovery.
Furthermore, SVM and Naïve Bayes classifiers tend to “rival” each other due to
the fact they are both supervised learning algorithms that are efficient at classifying
information. Both these techniques are moderately accurate at categorising fake
news in experiments. This is why this proposed method focuses on combining SVM
and Naïve Bayes classifier to get even more perfect results [5]. In “Blend Gullible
Bayesian and Bolster Vector Machine for Interruption spotting Framework,” the
creators coordinate the two techniques for SVM and Guileless Bayes classifier so as
to make a more exact strategy that characterizes better than every strategy especially.
They found that their “half and half calculation” successfully limited “bogus positives
just as boost balance discovery rates,” and performed somewhat better than SVM
and Gullible Bayes classifiers did especially [6, 7]. Despite the fact that this test was
applied to Interruption Location Frameworks (IDS), it shows that blending the two
strategies would be applicable to counterfeit news recognition.
692 D. V. Patil et al.

Acquainting semantic examination with Credulous Bayes and SVM classifier can
improve the calculation [7]. The greatest disadvantage of Guileless Bayes classifier
is that it thinks about all highlights of an archive, or whichever literary arrangement
being utilized, to be free despite the fact that more often than not excessively isn’t
the circumstance. This is an issue because of brought down precision and the way
that connections are not being educated if everything is thought to be isolated. As
we referenced before, one of the primary preferences of semantic investigation is
that this strategy can discover connections among words. In this manner, including
semantic investigation helps fix one of the principle shortcomings of the Innocent
Bayes classifier [8].
Adding semantic examination to SVM can improve the working of the classifier.
In “Support Vector Machines for data divided Based on Latent rules Indexing, “the
creator shows that combine the two strategies improves the productivity because of
“concentrating consideration on Help Vector Machines onto information subspaces
of the element spaces,” [7]. In the examination, semantic investigation had the option
to catch the “hidden substance of the report from a semantic perspective,” (Huang
2001), [9]. This will improve the effectiveness of SVM since the strategy would lose
less of its time characterizing good for nothing information and invest more energy
arranging pertinent information with the assistance of semantic investigation. As
laid out before, a more advantage of semantic investigation is its capacity to sepa-
rate basic information through connections between words; subsequently, semantic
examination can utilize its essential favorable position to improve SVM further [1].

3 Methodology

3.1 Problem Statement

Online life for news utilization is a twofold edged sharp edge. From one perspective,
its minimal effort, simple access, and quick scattering of data lead individuals to
search out and devour news from online life. Then again, it empowers the across
the board of “counterfeit news”, i.e., Low-quality news with intentional bogus data.
The abundant spread of phony news has the potential for amazingly negative effects
on people and society. Along these lines, counterfeit news identification of online
networking has as of late become a developing examination that is drawing in gigantic
consideration.
Counterfeit news discovery of online life presents one of a kind qualities and
difficulties that make existing location calculations from customary news media
failed or not material. Initially, counterfeit news is deliberately composed to mislead
perusers to accept bogus data, which makes it troublesome and nontrivial to identify
dependent on news content; along these lines, we have to incorporate transitory data,
for example, client social commitment via web-based networking media like Twitter,
Facebook, and so on, to help make an assurance. Second, using this assistant data
Identification of Fake News on Social Media: A New Challenge 693

is trying all by itself as clients’ social commitment with counterfeit news produces
information that is huge, inadequate, unstructured, and loud.

3.2 Definition

(Fake News Detection) the social-news, engagement E among n users for the news
article, the task of fake news detection is to predict whether the news article a is a
fake news part or not, i.e., F: E {0, 1}.
Such that,

F(a) = 1 if a is a part of the fake news,

0 otherswise

where F is the prediction function, we have to learn.

3.3 Collecting Data

• There are two pieces of information expansion process, “counterfeit news” and
“genuine news”. Social occasion the phony news was basic as Kaggle gave a phony
news dataset comprising 13,000 articles uncovered all through the 2016 political
race cycle. It includes huge work around a few Locales because of it had been
the sole gratitude to do web scratching a large number of articles from changed
sites. By scratching an aggregate of 5279 articles, the genuine news dataset was
made, generally from media associations (New York Times, WSJ, Bloomberg,
NPR, and the Watchman) which are distributed around 2015–2016.
• Online news can be assembled from various sources, for example, news asso-
ciation landing pages, web indexes, and internet based life. Be that as it may,
physically choosing the truth of stories might be a troublesome errand, ordinarily
difficult to-satisfy annotators with space experience who plays out a cautious
investigation of Cases and additional confirmation, setting, and reports from Legit-
imate sources. Generally, news, information with explanations can be gathered in
the accompanying manners: Master columnists, Actuality checking sites, Industry,
indicators, and Publicly supported Laborers. In any case, there are no settled upon
benchmark Datasets for the phony news identification issue.
694 D. V. Patil et al.

3.4 Technology Used

Python is a phenomenal language for doing information investigation on account of

the incredible environment of information driven python bundles. Python is an open
source language that was made to be anything but difficult to peruse and incredible.
It is a deciphered language. Deciphered dialects shouldn’t be incorporated to run. A
software engineer can change the code and speedier outcomes. This implies Python
is more slow than an assembled language like C since it isn’t running machine code
straightforwardly.
Numpy is the fundamental package, and it is used for scientific computing with
Python. It contains, among other things:
• A powerful N-dimensional array object
• Sophisticated (broadcasting) functions
• Tools for integrating C/C++ and Fortran code
• Useful linear algebra, random number and Fourier transform capabilities.
Pandas are one of those bundles and make breaking down information and bringing
in a lot simpler. Pandas work for bundles like matplotlib and Numpy to give you a
helpful, single, spot to do the majority of your information investigation and repre-
sentation work. Matplotlib is a plotting library for the Python programming language,
and it is numerical science expansion NumPy. It gives an item arranged Program-
ming interface to inserting plots into applications utilizing universally useful GUI
toolboxs like Qt, Tkinter, wxPython, or GTK+. Sklearn is likely the most helpful
library for AI in Python. It is on SciPy, NumPy and matplotlib, this library contains a
great deal of proficient apparatuses for AI and measurable demonstrating including
grouping, characterization, relapse, and dimensionality decrease.

3.5 Data Prepossessing

Data preprocessing is a data mining technique that involves transforming raw data
into an understandable format. Real-world data is often incomplete, inconsistent,
and/or lacking in certain behaviors or trends, and is likely to contain many errors.
Data preprocessing is a proven method of resolving such issues. Data preprocessing
prepares raw data for further processing.
It is an integral step in Machine Learning as the quality of data and the useful
information that can be derived from it directly affects the ability of our model to
learn; therefore, it is extremely important that we preprocess our data before feeding
it into our model as shown in Fig. 2.
A confusion matrix is a summary of prediction results on a classification problem.
The number of correct and incorrect predictions are summarized with count values
and broken down by each class. This is the key to the confusion matrix.
Identification of Fake News on Social Media: A New Challenge 695

Fig. 2 Confusion matrix

Fig. 3 Classification of fake and real news-1

The confusion matrix shows the ways in which your classification model is
confused when it makes predictions. It gives us insight not only into the errors
being made by a classifier but more importantly the types of errors that are being
made as shown in Figs. 3 and 4.

3.6 Feature Extraction

News content highlights portray the Meta data identified with a bit of news. A
rundown of Illustrative news content characteristics are recorded beneath:
696 D. V. Patil et al.

Fig. 4 Classification of fake and real news-2

• Source: Writer of the news story

• Headline: Short title message that centers around grabbing the eye of perusers
and portrays the primary subject of the article.’
• Body Content: Fundamental content that depicts the subtleties of the report; there
is typically an essential case that is especially featured and that shapes the point
of the distributor.

In light of these crude substance qualities, various types of highlight depictions

can work in Concentrate segregating attributes of phony news. Ordinarily, the news
data we are seeing will for the most part be semantic based and visual-based.

3.7 Model Construction

Since counterfeit news endeavors to spread bogus cases in news content, the most
direct method for recognizing it is to check the honesty of incredible cases in a
news story to choose the news veracity. Information based methodologies mean to
utilize outside sources to actuality check proposed to demand news content. The
objective of certainty checking is to designate a fact incentive to a case in a particular
setting. Reality checking has pulled in raising consideration, and numerous endeavors
have been made to build up a practical computerized certainty checking framework.
Existing reality checking can arrange as master situated, publicly supporting focused,
and computational-arranged.
Identification of Fake News on Social Media: A New Challenge 697

Fig. 5 System architecture

3.8 System Architecture

Extracting Data from the Dataset

As shown in Fig. 5, the task required informational collection is accessible on GitHub.
On the GitHub site, there are two CSV documents accessible, for example Train
Information and Test Information. In excess of 52,000 articles. 12,000 of the articles
were marked as phony or genuine and downloaded from Gitub.com. Train.csv: This
preparation dataset with the accompanying properties: Id: a novel id for a news story.
• Title: The title of a news story
• Author: writer of the news story
• Text: The content of the article
• Label: A name that denotes the article as possibly deceitful
1: untrustworthy.
0: trustworthy.
Test.csv: Testing training dataset with all the same attributes same as train.csv
without the label.
Generated TFIDF
TF: Term Frequency, which counts how frequently a term occurs in a document.
Thus, the term frequency is often divided by the length of the document (total no. of
terms in the documents).
TF(t) = (Number of times term t appears in a document)/(Total no of term in
document).
698 D. V. Patil et al.

Suppose Consider a document containing 100 words wherein the word apologist
appears 3 times.
The term frequency (i.e., tf) for apologist is then
(3/100) = 0.03.
IDF: Inverse Documents Frequency, which measures how important a term is.
While calculating TF, all terms are considered equally important.
However, it is known that specific terms, such as “is”, “of”, and “that”, may appear
many times but have little importance.
IDF(t) = log_e(Total number of file (documents)/Number of documents with term
T in it).
Assume we’ve ten million documents and also the word seems in one thousand
of those. Then, the inverse document frequency (i.e., idf) is computed as:
log(10,000,000/1,000) = 4.
Thus, the Tf-idf is the product of these two quantities: 0.03 * 4 = 0.12.
Count Vectorizer
The Count Vectorizer provides a simple way to both tokenize a collection of text
documents and form vocabulary of known words, but also to encode new documents
using that vocabulary. You can use it as follows build an instance of the Count
Vectorizer class.
If you implemented CountVectorizer on one piece of documents and then you want
to use the set of features of those documents for a new set, use the vocabulary_attribute
of our original Count Vectorizer and pass it to the new one.
Test Train Split
Train-Test Splits. The data we use is usually divided into training data and test
data. The training set contains a known output, and the model learns on this data in
sequence to be generalised to other data later on. We have the test dataset (or subset)
in series to test our model’s prediction on this subset.
Test Output
Better fitting of the training dataset, as opposed to the test dataset, usually point in
fitting. Test pieces are therefore a set of examples used only to assess the performance
(i.e. Generalization) of a fully specified classifier. Holdout sample is a sample of data
not used to fit a model, used to assess the performance. Of that model; this book uses
the term validation pieces or if one is used in the problem, test set. Instead of a
holdout sample.
Identification of Fake News on Social Media: A New Challenge 699

4 Result

As you can see by running the cells below, both vectorizers extracted the same tokens,
but obviously have different. Likely, changing the max_df and min_df of the TF-IDF
vectorizer could alter the result and lead to different features in each. Compare TF-
IDF versus bag-of-words. My intuition was that bag-of-words (CountVectorizer)
would perform better with this model.
700 D. V. Patil et al.

So fake news is solved. We achieved 93 accuracy on our dataset Next output is

which new is fake or real.
Identification of Fake News on Social Media: A New Challenge 701

5 Conclusion and Future Scope

With the increasing quality of social media, a lot of people use news from social
media rather than traditional news media. However, social media have also been
used to spread fake news, which has strong adverse effects on individual users and
broader society. In this project, we explored the fake news problem by reviewing
existing literature in two phases: characterization and detection.
In the characterization phase, we introduced the basic concepts and principles of
fake news in both traditional media and societal media. In the detection phase, we
reviewed existing fake news detection approaches from a data mining perspective;
it includes feature extraction and model construction. We also further discussed the
data sets, evaluation metrics, and promising future directions in fake news detection
research and spread the field to other applications.

References

1. Rubin V, Conroy N, Chen Y, Cornwell S (2016) Fake news or truth? Using satirical cues to
detect potentially misleading news. In: Proceedings of the second workshop on computational
approaches to deception detection. https://doi.org/10.18653/v1/w16
2. Conroy NJ, Rubin VL, Chen Y (2015) Automatic deception detection: methods for finding fake
news. Proc Assoc Inf Sci Technol 52(1):14
3. Chen Y, Conroy NJ, Rubin VL (2015) News in an online world: the need for an automatic crap
detector. Proc Assoc Inf Sci Technol 52(1):14
4. Fake news websites (n.d.) en.wikipedia.org/wiki/Fakenewswebsite. Accessed on 6 Feb 2017
5. Conroy N, Rubin V, Chen Y (2015) Automatic deception detection: methods for finding fake
news. Proc Assoc Inf Sci Technol 52(1):1–4
6. Naive Bayes spam filtering (n.d.) https://en.wikipedia.org//wiki/NaiveBayesspamfiltering.
Accessed on 6 Feb 2017
7. Wang AH (2010) Dont follow me: spam detection in Twitter. In: Proceedings of security and
cryptography international conference (SECRYPT)
8. Memon AA, Vrij A, Bull R (2003) Psychology and law: truthfulness, accuracy and credibility.
Wiley
9. Joulin A, Grave E, Bojanowski P, Mikolov T (2016) Bag of tricks for efficient text classification
A Smart and Secure Helmet for Safe
Riding

Ramesh Kagalkar and Basavaraj Hunshal

Abstract Life is becoming more fast and hazardous while driving. Moreover life is
valuable so, we need to have some automation techniques to secure life. In this paper
we attempted to plan our thought in which the system that can recognize the person
worn or not the helmet, spot accident place and immediate response, finding whether
the person consumed liquor while riding, identify petrol level of tank and predicting
the crash or collision between the vehicles in order to avoid road side accidents. The
interaction between bike and helmet part takes place wirelessly by making utilization
of RF transmitter and receiver. RF transmitter is appended at helmet and receiver at
bike.

Keywords Alcohol sensor · Vibration sensor · Magnetic sensor ·

Microcontroller · IR sensor · RF module · GPS · GSM

1 Introduction

Road accidents are incredibly increasing day by day, and in countries like India
where bikes are more in use due to low price when compared to other vehicles many
people die due to carelessness by not wearing helmet. Even though there have been
continuous awareness program from the government authorities regarding helmets
and seat belts a majority of the riders not follow the rules. In order to put an end to
accidents of bike we have planned to develop the smart helmet for bike, in which the
helmet not allow rider to ride the bike without wearing helmet. This is the best way
to make helmet is compulsory for bike riders. Other than this there are four more
features alcohol detection, identifying accidents, petrol level indication, and obstacle
detection to avoid the collision between the vehicles. Nearly a third of all fatal bike
accidents happen after riders have been drinking. Head on collisions: More than 75%
of all motorcycle crashes due to head on collision. Exceeding speed of limit: Some
32% of motorcycle accidents that resulted in death involved speeding. Left turns:
The road can turn into a danger zone when a car makes a left-hand turn. In 42% of

R. Kagalkar · B. Hunshal (B)

KLE College of Engineering and Technology, Chikodi, Karnataka, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 703
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_70
704 R. Kagalkar and B. Hunshal

all fatal car accidents, the other vehicle was turning left while the biker was either
heading straight, passing or overtaking. Running into objects: In some cases accident
happen without knowledge of rider due to objects or vehicles around bike.
The rest of the paper is outlined into 5 sections, where Sect. 2 presents the outline
of literature survey. The overview of proposed system and it’s irs flow diagram
given in Sect. 3. In Sect. 4 the methodologies used to developed is discussed. The
implementation and results are discussed with case studied where presented in Sect. 5.
Finaly the conclusion is made for development of work.

2 Literature Survey

The paper [1] aim is to provide a in time treatment to person by making use of
ambulance. Many a times when the rider needs an ambulance that time ambulance
may not be available and no one available at the time of accident it often causes deaths.
So present system is prepared in such a way that it gives the notifications at the earliest
so the any responsible person may take any required action for this Aurduino used
ad micro-controller. İn paper [2] aim is to save the people when the rescue workers
get harm in the disaster the Smart helmet respond to the current accident. Using
TCP and UDP socket connection communication is made between Smart helmet
and the device. It provides the disaster safety and secure helmet for the workers. In
real time problems the Smart helmet is useful through which the other person can get
the rescue worker location at the time of disaster event occur. In paper [3] building
a shield which ensures safety of the rider by implementing feature such as alcohol
detection, accident identification, location tracking and fall detection. Force sensing
resistor is used to sense the actual human touch, to ensure if the helmet is worn by the
rider. Accelerometer ADXL345 is used to measure the static acceleration of gravity.
This unit senses if the bike is falling. And bike will take decision that accident has
occurred or not. The paper [4] gives solution to help a rider in need, after the rider has
met with an accident. Raspberry Pi zero it supports multiple os such as linux based,
Raspbian OS, Windows 10 IoT cosre, Google’s Android Things, etc. Pressure sensor
it measures the pressure with which object falls. MCP 3008 IC it is a 10-bit analog
to digital converter. Google’s Firebase Cloud messaging is a crossplatform solution
for messages and notifications for Android, iOS and web applications, which is free
of cost. E Android is a mobile operating system designed for smart phones. In [5]
build protective helmet for miners. The system is used by the miners and this safety
helmet is used for detecting the health of the person and also it provides the same
information to the central office outside the mines. Helmet consists of the wireless
sensor devices which consists of the force sensor and that sensor detects the dangerous
load experienced by the miners on his head if that was heavy load then the wireless
sensors inform about the danger to the center with the help of the room manager. After
knowing the information center can help miners by sending team. The model can
be consists of the miners, Room manager and center. The center is connected to the
GUI which helps to send/Receive message from room manager. The sensor node will
A Smart and Secure Helmet for Safe Riding 705

implement routing protocol in which route is calculated based on distance vectors

using the Bellman ford algorithm. The paper [6] is to reminding the riders about
wearing the helmet properly and also informs about the speeding over speed limit
and also detecting the locked buckle of the vehicle. The model consists of the two
modules which are mounted on the side of the helmet and also on the motorcycle
dashboard. These module are controlled by the Arduino and nRF24L01+ is used
for communication between the modules. The response time is calculated on four
smart helmet functions which are the helmet usage, helmet strap locking system,
speed detection, and shock detection [7]. Author’s aim at building safety helmet and
wearing detection by detecting a single shot multi-box detector (SSD) using deep
learning. In this dataset, image, SSD, detection result, label, feature and safety-helmet
precise detecting are used. Here experiments are conducted by using own dataset.
Overall testing set contains 1076 images, validation set contains 1258 images and
training set is divided and contains 2895 images, then final dataset contains 5229
images.
The paper [8] aim is to propose fall detecting using barometric pressure sensor
and whoes accuracy is high. Another goal is implementing IoT app where the person
can speak and helps to monitor old age people from any where. İn [9] informs about
theMicrocontroller Arduino NANO this is used to calculate the ‘tilt’ of the helmet
while riding the motorbike. Flex Sensor is used which is a strip attached to the interior
of helmet, which bends upon wearing a helmet, to detect if the helmet is worn. Impact
sensor detects the vibration when the helmet falls on the ground and the impact is
sensed by the impact sensor. The paper [10] aim is to build a model that gives the
safety like car for example characters like built-in black box and GSM system, so in
emergency time you can find the rider’s location easily.

3 Overview of Proposed System

Figure 1 shows complete functioning of proposed system that is how the decision
is made. Input to system is helmet and processing starts from this part. When the
helmet is worn then helmet is detected, if yes bike starts else bike won’t starts. Same
way if alcohol is detected if yes then bike stops else bike will be in running state. In
accident detection, the system continuously monitors and observes for an accident.
At the time accident detected if yes message is quickly sent to pre stored number
else bike is in running state. Petrol level is monitored if low petrol bike stops. And
if any obstacle is detected message is displayed on LCD. This process of execution
is represented in the form of flow chart.
706 R. Kagalkar and B. Hunshal

Fig. 1 Flow diagram of smart helmet

4 Methodologies

There are five modules are considered in the system. The heart of the system is Micro-
controller, that is Atmega 328 which has 28 pins. All sensors used in implementation
are connected to input pins of microcontroller. Figure 2 shows sytem setup of the
smart and secure helmet system.
By taking input from different types of sensors the microcontroller will produce
the output. To set some conditions and requirements data, we have developed code in
embedded C language which will be dumped on to the ATmega328 microcontroller.
Arduino software is used in order to execute embedded C code.
A Smart and Secure Helmet for Safe Riding 707

Fig. 2 System setup of smart and secure helmet system

• Helmet Detection Module

As shown in Fig. 2, push limit switch is used in order to detect the helmet is worn or
not. The push switch is attached in such a way that when person wear the helmet the
push switch is pressed at that time RF transmitter send signal that analog signal is
converted to digital by ADC after that binary 1 is generated that is transmit to bike
part where in RF receiver receives after that bike is going to start. And if RF receiver
of bike part receives binary 0 then bike will not start.
• Alcohol Detection Module

Alcohol detection is done by the MQ3 gas sensor which senses the alcohol fumes
from the breath of the person. So, if alcohol is detected then immediately bike stops
there is no chance of riding bike by consuming alcohol. MQ3 gas sensor is as shown
in Fig. 3. MQ3 is sensor which senses the gases produced by alcohol.
• Accident Detection System

Magnetic sensor will helps to detect the accident and GSM modem will immediately
send the SMS to the predefined numbers. By this we can provide proper facility to
the person who got in to an accident. When a person got into an accident at that time
helmet send the signal to bike part and then GSM receives signal to send the message
to pre-defined number to give a in time treatment to injured person.
708 R. Kagalkar and B. Hunshal

R
e
c
o
g
ni
Ɵ
o

Fig. 3 Helmet recognition time of the system

• Petrol Level Detection Module

In this module it helps the rider to know about petrol level of bike by indication. And
the levels are low level, middle level, full level and over full level. If petrol tank has
less petrol level then indication is made by buzzer and display message on LCD.
• Obstacle Detection Module

Collision of vehicle also leads to accidents to avoid that IR sensor is used at backside
of bike in order to detect obstacles at back side. So, here sensor will detect obstacle
and alert is given by the buzzer sound and message on LCD display.

5 Implementation and Results

Execution is done to check whether the system is working properly or not. The
performed tested and evaluated by testing the system for various focibilities of
outputs. In this testing each unit is tested to check the performance and efficiency
of the unit. To show the efficiency of whole system, with respect to all objectives
processing time and response time are calculated in the following sections.
• Helmet Detection and start of the Bike:

The rider can start the bike only if he wears the Helmet properly. Processing time of
Helmet may very according to the model as well as its size. Therefore, processing
time can be defined as, the time required to recognize the Helmet’sworking and
glowing of the Blue LED over Helmet from RED. Recognition rate is considered in
percentage, and it is with respect to successful start of the bike when a rider wears
the Helmet. In our system, the Helmet as well as bike both are fixed with Red and
A Smart and Secure Helmet for Safe Riding 709

Blue LED’s for different combinations of outcomes. The different cases considered
here are,
– If the applied Key of bike is proper then Blue LED will glow, else Red LED will
glow at Bike.
– For proper wear of Helmet, Blue LED over Helmet glows, else Red LED will
glow.
Cases to consider the Recongition rate:
– If a LED over Helmet and Bike both are Blue, it is 100% recongition rate.
– If LED of Helmet is RED or rider not having helmet and Bike is Blue, then it has
been considered as 50% recognition rate.
It means the rider either not having helmet or not properly weared it. The bike
LED is Blue because, the Key applied to start the bike is proper one.
– In case, if the rider had weared the helmet properly, but applied wrong key then
it is 50% recognition rate.
– In case, if the rider nighter weared the helmet properly, nor applied the proper key
it will be considered as 0% recognition rate.
In Fig. 3 shows an helmet recognition time of the system and in Fig. 4 shows an
Helmet processing time of the system.
• Alcohol Detection and Result Analysis:

Alcohol sensor MQ3 is used in the system to guard the rider from an accident. When
a person is trying to ride the bike at that time if alcohol fumes are detected from
breath of the person then LCD should display “ALCOHOL DETECTED” and bike

Fig. 4 Helmet processing time of the system

710 R. Kagalkar and B. Hunshal

automatically have to stop. The below graphs shows the test cases with respect to
percentage of alcohol the rider has consumed. The alcohol consumption exceeds
the configured threshold value of alcohol then, bike wouldn’t start. Threshold value
considered here is 25%. The bike is configured with two LED’s. If alcohol detected
is above the threshold value, then Red LED will glow, else Blue LED will glow and
rider can run the bike.
Alarm and Indicating Unit
An alarm is used as buzzer thatbuzzes when a rider is detected by alcohol. A PS
series buzzer is used. A DC motor is used for demonstration of engine locking. For
the same DC motor is joined to pin number 9 of microcontroller. When alcohol is
detected during exheling time by the ride, the DC motor stops the working and motor
will keep on running if the sensor not detects alcohol.
Engine Locking Unit
Engine locking was developed by the using a DC motor to exhibit like a bike engine.
The DC motor operates with specific conditions; when the alcohol level crosses 40%
of threshold value the engine automatically stops. The engine runs continuously if
the alcohol level comes less than 39%. The LED is used to show when the sensor
detect alcohol. in the simulation, when the logic state is 1 the led goes on to indicate
that alcohol is present and off to show the absence of alcohol.
Figure 5 graph x-axis represents the rider samples, tested for 8 instances and the
y-axis represents the processing time in seconds.
Figure 6 graph x-axis represents the rider samples, tested for 8 instances and the
y-axis represents the recognition rate in seconds.

Fig. 5 Processing time of alcohol detection

A Smart and Secure Helmet for Safe Riding 711

Fig. 6 Recognition rate of alcohol detection

• Accident Detection and Result Analysis

Due to increase in trafic as well as rider’s fast move, accidents may happen. But,
some precotionary techniques are needed to avoid the lose of life. Therefore, in
our system, vibration sensor have been used by configuring the threshold value i.e.,
15 mV/g. When the hammering over helmet crosses the set threshold value, an
alert messege is sent automatically to defined phone numbers. Recognition rate is
calculated depending on the threshold value to notify the seviority of cause happened
to the rider. The system identifies the obstacle present in front and bike of a bike. For
this, IR sensor is used to detect the threats during back and forward turns or riding
over the road. To identify an obstacle the distance considered here is four feets for
our experimentation. It can be advanced by adopting the good one.
Obstacle Detection
An IR sensor is used for detecting the obstacles arriving at back side. The sensor
is mounted over Helmet. After detecting obstacle at the back side LCD should
display “OBSTACLE IS DETECTED AT THE BACK”. The system will prompt
by displaying as in Fig. 7.
Petrol Level Detection
Continuously system should monitor petroleum level in fuel tank LCD displays
FULL, MIDDLE and LOW level. At the time of low level petrol bike should stop
automatically. Figure 8 shows the LCD display of the output.

Fig. 7 LCD display for

obstacle detection
712 R. Kagalkar and B. Hunshal

Fig. 8 LCD display shows the level of fual tank

Depending on the level of petrol in the tank we can calculate how much KM more
a rider can travel on the bike.

6 Conclusion

The research over the idea fetches us the knowledge, that the technique plays yet
the vital role in society. The systemhas an effective smart and secure helmet system
which is based on RF transmitter and receiver. There are five modules and those
are considered as the objectives of thesystem. ATmega328 microcontroller is used
for executing the objectives of whole system. The experimental test results have
been obtained and observed that the system efficiently protects and awakes the rider
by indicating the LED signals. Smart and secure helmet is usefull and efficient
application for socity to reduce the accidents.

References

1. Ahuja P, Bhavsar K (2018) Microcontroller based smart helmet using GSM and GPRS. In: 2nd
IEEE ınternational conference on trends in electronics and ınformatics. IEEE Press
2. Jeong M, Lee H, Bae M, Shin D-B, Lim S-H, Lee KB (2018) Development and application of
the smart helmet for disaster and safety. IEEE Press
3. Tapadar S, Ray S, Saha AK, Karlose R, Saha HN (2018) Accident and alcohol detection in
bluetooth enabled smart helmets for motorbikes. IEEE Press
4. Biswas JR, Kachroo S, Chopra P, Sharma S (2018) Development of an app enabled smart
helmet for real time detection and reporting of accidents. IEEE Press
5. Revindran R, Vijayaraghavan H, Huang M-Y (2018) Smart helmets for safety in mining
industry. IEEE Press, Year
6. Budiman AR, Sudiharto DW, Brotoharsono T (2018) The prototype of smart helmet with safety
riding notification for motorcycle rider. In: 3rd IEEE ınternational conference on ınformation
technology, ınformation systems and electrical engineering. IEEE Press, Indonesia
7. Long X, Cui W, Zheng Z (2019) Safety helmet wearing detection based on deep learning. In:
3rd IEEE ınformation technology,networking,electronic and automation control conference.
IEEE Press
8. Rajan PTB, Mounika M, Nivetha S, Olive Sherine J (2019) Smart helmet based on ıot for
accıdent detectıon and notıfıcatıon. Int J Emer Technol Innov Eng 5(9)
9. Shravya K, Mandapati Y, Keerthi D, Senapati RK (2019) Smart helmet for safe driving. E 3S
web of conferences
10. Gupta S, Sharma K, Salvekar N, Gajra A (2019) Implementation of alcohol and collision sensors
in a smart helmet. In: IEEE ınternational conference on nascent technologies in engineering.
IEEE Press
On Some Properties of Extended Binary
Golay [24, 12, 8] Code Using MATLAB

N. S. Darkunde, S. P. Basude, and M. S. Wavare

Abstract The main aim of this paper is to study various properties of extended
binary Golay [24, 12, 8] code, when we know its generator matrix. Using MATLAB,
we can study syndrome decoding, weight of a codeword, error correction and error
detection of extended binary Golay [24, 12, 8] code.

Keywords Linear code · Generator matrix · Parity check matrix · Extended binary
golay code · Syndrome decoding

1 Introduction

In the late 1940s Claude Shannon was developing information theory and coding as
a mathematical model for communication. In mathematical sciences and electronics
engineering, a binary Golay code is a type of linear error-correcting code used in
digital communications. The binary Golay code, along with the ternary Golay code,
are named in honour of Marcel J. E. Golay whose 1949 paper [1] introducing them has
been called, by E. R. Berlekamp, the “best single published page” in coding theory.
There are two closely related binary Golay codes. The extended binary Golay code
encodes 12 bits of data in a 24-bit word in such a way that any 3-bit errors can be
corrected or any 7-bit errors can be detected.
The other, the perfect binary Golay code, has codewords of length 23 and is
obtained from the extended binary Golay code by deleting one coordinate position.
In standard coding notation the codes have parameters [24, 12, 8] and [23, 12, 7],
corresponding to the length of the codewords, the dimension of the code, and the
minimum Hamming distance between two codewords, respectively [2, 3]. In [4, 5],
algorithmic approach for error correction have been studied for few codes. In this

N. S. Darkunde (B)
School of Mathematical Sciences, SRTM University, Nanded, Maharashtra, India
S. P. Basude
SRTM University, Nanded, Maharashtra, India
M. S. Wavare
Department of Mathematics, Rajarshi Shahu Mahavidyalaya, Latur, Maharashtra, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 713
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_71
714 N. S. Darkunde et al.

paper, we are going to study some properties of extended binary Golay [24, 12, 8]
code.

2 Preliminaries

2.1 Linear Codes [6]

Let Fq denote the finite field with q elements, where q is some power of a prime. A
linear [n, k]q —code is a k-dimensional subspace of Fqn . The parameters n and k are
referred to as the length and dimension of the corresponding code.

Example The subset C = {000, 001, 010, 011} of vector space F23 is [3, 2]2 linear
code. Similarly, C = {0000, 1100, 2200, 0001, 0002, 1101, 1102, 2201, 2202} is
[4, 2]3 linear code.

2.2 Hamming Distance and Hamming Weight [7]

Let x, y ∈ Fqn , The Hamming distance from x to y, denoted by d(x, y), is defined
to be the number of places at which x and y differ.

Example Consider x = 01010,y = 01101, z = 11101 in F25 , Thend(x, y) =

3,d(x, z) = 4. For any x ∈ Fqn , the support ofx, denoted by supp(x), is defined to be
the set of nonzero coordinates in x = (x1 ,x2, x3,..., xn ), that is supp(x) = {i : xi = 0}.

For a [n, k]q code C containing at least two words, the nonnegative integer given
by min{d(x, y) : x, y ∈ C, x = y} is called minimum distance of C. It is denoted by
d(C).

Example For a code C = {0000, 1000, 0100, 1100} in F24 , we see thatd(C) = 1.

Definition 1 [8] Let u be a positive integer. A code C is u -error-detecting if, when-

ever a codeword incurs at least one but at most u errors, the resulting word is not a
codeword.

A code is exactly u-error-detecting if it is u-error-detecting but not (u + 1) error-

detecting.

Example Consider C = {000000, 000111, 111222} ⊆ F26 . This code is 2-error-

detecting, because changing any codeword in one or two positions does not result
in another codeword. In fact, C is exactly 2-error-detecting, as changing each of the
last three positions of 000000 to 1 will result in the codeword 000111 (so C is not
3-error-detecting).
On Some Properties of Extended Binary Golay … 715

Theorem 1 [8] A code C is u -error-detecting if and only if d(C) ≥ u + 1, that is,

a code with distance d is an exactly (d − 1)-error-detecting code.

Definition 2 [8] Let v be a positive integer. A code C is v -error-correcting if

minimum distance decoding is able to correct v or fewer errors, assuming that the
incomplete decoding rule is used. A code C is exactly v-error-correcting if it is v
-error-correcting but not (v + 1)-error-correcting.

Example Consider C = {000, 111} in F23 . It is easy to see that, C is 1-error-

correcting.

Theorem 2 [8] A code C is v -error-correcting if and only if d(C) ≥ 2v + 1, that

is a code with distance d is an exactly (d − 1)/2-error correcting code, where x
denote the greatest integer less than or equal to x.

Definition 3 (Dual Code). [8] Given a [n, k]q code C in Fqn , the subspace
C ⊥ ={x = (x1, x2, x3,..., xn ) ∈ Fqn : x.c = 0 for all c = (c1 , c2 , . . . , cn ) ∈ C},
We have following properties of C and C ⊥

1. |C| = q dim(C) , i.edim(C) = log q |C|;

 
2. C ⊥ is a linear code and dim(C) + dim C ⊥ = n;
⊥
3. (C ⊥ ) = C.

2.3 Generator Matrix and Parity-Check Matrix

Definition 4 [7] A matrix G of order k × n is said to be the generator matrix for a

[n, k]q code C, if its rows form basis for C.

Definition 5 [7] A parity-check matrix H for a code C is a generator matrix for C ⊥ .

2.4 Binary Extended Golay [24, 12, 8] Code

Let G = (I12 |A), where I12 is a 12 × 12 identity matrix and A is a 12 × 12 matrix

given by
716 N. S. Darkunde et al.
⎡ ⎤
100011111001
⎢ 111010101010 ⎥
⎢ ⎥
⎢ ⎥
⎢ 101111100100 ⎥
⎢ ⎥
⎢ 101110010011 ⎥
⎢ ⎥
⎢ 011111001001 ⎥
⎢ ⎥
⎢ 011100100111 ⎥
⎢
A=⎢ ⎥
⎥
⎢ 110101010101 ⎥
⎢ 010010011111 ⎥
⎢ ⎥
⎢ ⎥
⎢ 011001111100 ⎥
⎢ ⎥
⎢ 100100111110 ⎥
⎢ ⎥
⎣ 010111110010 ⎦
101001001111

The binary linear code generated by matrix G is called extended binary Golay
[24, 12, 8] code.

2.5 Syndrome Decoding [8]

An efficiency of decoding technique works well, when length n of a given code is

small, but it can take a more time when, n is very large, so this time can be saved
by using the syndrome to identify the coset from which the word is taken. In [8], the
procedure of syndrome decoding have been demonstrated.
Step 1: Let w be received word in the transmission and for this received word w,
first compute the syndrome of w denoted by Syn(w) which is given by, syn(w) =
w H T , where H , is parity check matrix of a given code.
Step 2: After constructing Syndrome look up table, we will find the coset leader
u next to the syn (w) = syn(u).
Step 3: Finally decode the received word w as v = w − u.
Now, let us study the properties of extended binary Golay [24,12,8] code using
MATLAB [9].

2.6 MATLAB Program for Syndrome Decoding Using

Extended Binary Golay [24, 12, 8] Code

%Extended Golay Code

n = 24; k = 12; %Length and Dimension
clc % Clearscreen
I = eye(12); % Identity matrix of order 12
On Some Properties of Extended Binary Golay … 717

A = [1 0 0 0 1 1 1 1 1 0 0 1;
1 1 1 0 1 0 1 0 1 0 1 0;
1 0 1 1 1 1 1 0 0 1 0 0;
1 0 1 1 1 0 0 1 0 0 1 1;
0 1 1 1 1 1 0 0 1 0 0 1;
0 1 1 1 0 0 1 0 0 1 1 1;
1 1 0 1 0 1 0 1 0 1 0 1;
0 1 0 0 1 0 0 1 1 1 1 1;
0 1 1 0 0 1 1 1 1 1 0 0;
1 0 0 1 0 0 1 1 1 1 1 0;
0 1 0 1 1 1 1 1 0 0 1 0;
1 0 1 0 0 1 0 0 1 1 1 1];
G = [I A] % Generator matrix of extended Golay code
H = mod([−A’ eye(12)],2) % Denote H as a parity Check Matrix for extended Golay
Code
d = gfweight(G) % Distance of Code
slt = syndtable(H); % Produce Syndrome look up table.
w = [1 1 0 1 0 1 0 1 0 1 0 1 1 1 0 0 0 0 0 1 1 0 1 0] % Message received
syndrome = rem(w * H’, 2);
syndrome_de = bi2de(syndrome,’left-msb’); % Convert to decimal.
disp([’Syndrome = ’,num2str(syndrome_de),’ (decimal),’,num2str(syndrome),’
(binary)’])
corrvect = slt(1 + syndrome_de,:) % Correction vector
% Now compute the corrected codeword.
correctedcode = rem(corrvect + w,2)
G=
Columns 1 through 13
1000000000001
0100000000001
0010000000001
0001000000001
718 N. S. Darkunde et al.

0000100000000
0000010000000
0000001000001
0000000100000
0000000010000
0000000001001
0000000000100
0000000000011
Columns 14 through 24
00011111001
11010101010
01111100100
01110010011
11111001001
11100100111
10101010101
10010011111
11001111100
00100111110
10111110010
01001001111
H=
Columns 1 through 13
1111001001011
0100111110100
0111110010010
0011111001100
1111100100100
1010101010110
1110010011100
On Some Properties of Extended Binary Golay … 719

1001001111100
1100100111010
0010011111010
0101010101110
1001111100010
Columns 14 through 24
00000000000
10000000000
01000000000
00100000000
00010000000
00001000000
00000100000
00000010000
00000001000
00000000100
00000000010
00000000001
d=8
Single-error patterns loaded in decoding table. 4071 rows remaining.
2-error patterns loaded. 3795 rows remaining.
3-error patterns loaded. 1771 rows remaining.
4-error patterns loaded. 0 rows remaining.
w=
Columns 1 through 13
1101010101011
Columns 14 through 24
10000011010
Syndrome = 275 (decimal), 0 0 0 1 0 0 0 1 0 0 1 1 (binary)
corrvect =
720 N. S. Darkunde et al.

Columns 1 through 13
1000001100000
Columns 14 through 24
00000100000
Corrected code =
Columns 1 through 13
0101011001011
Columns 14 through 24
10000111010

2.7 MATLAB Program for Encoding Message Using

Extended Binary Golay [24, 12, 8] Code

% Matlab Programme for Encoding of the information message

n = 24; k = 12; %Length and Dimension
clc % Clearscreen
I = eye(12); % Identity matrix of order 12
A = [1 0 0 0 1 1 1 1 1 0 0 1;
1 1 1 0 1 0 1 0 1 0 1 0;
1 0 1 1 1 1 1 0 0 1 0 0;
1 0 1 1 1 0 0 1 0 0 1 1;
0 1 1 1 1 1 0 0 1 0 0 1;
0 1 1 1 0 0 1 0 0 1 1 1;
1 1 0 1 0 1 0 1 0 1 0 1;
0 1 0 0 1 0 0 1 1 1 1 1;
0 1 1 0 0 1 1 1 1 1 0 0;
1 0 0 1 0 0 1 1 1 1 1 0;
0 1 0 1 1 1 1 1 0 0 1 0;
1 0 1 0 0 1 0 0 1 1 1 1];
G = [I A] % Generator matrix of extended Golay code
On Some Properties of Extended Binary Golay … 721

H = mod([−A’ eye(12)],2) % Denote H as a parity Check Matrix for extended Golay

Code
u = input(’Enter the message bit of length 12 for which you want to encode using
extended Golay Code = ’)% input message of length 12
v = mod(u*G,2) % Encoding of message u
Synv = mod(v*H’,2) % Syndrome of encoded message
G=
Columns 1 through 13
1000000000001
0100000000001
0010000000001
0001000000001
0000100000000
0000010000000
0000001000001
0000000100000
0000000010000
0000000001001
0000000000100
0000000000011
Columns 14 through 24
00011111001
11010101010
01111100100
01110010011
11111001001
11100100111
10101010101
10010011111
11001111100
00100111110
722 N. S. Darkunde et al.

10111110010
01001001111
H=
Columns 1 through 13
1111001001011
0100111110100
0111110010010
0011111001100
1111100100100
1010101010110
1110010011100
1001001111100
1100100111010
0010011111010
0101010101110
1001111100010
Columns 14 through 24
00000000000
10000000000
01000000000
00100000000
00010000000
00001000000
00000100000
00000010000
00000001000
00000000100
00000000010
00000000001
Enter the message bit of length 12 for which you want to encode using
On Some Properties of Extended Binary Golay … 723

Golay Code = [1 0 1 1 0 0 1 0 1 0 1 0]
u=101100101010
v=
Columns 1 through 13
1011001010100
Columns 14 through 24
11001010101
Synv = 0 0 0 0 0 0 0 0 0 0 0 0

3 Conclusion

Once we know the generator matrix of any linear code, using MATLAB we can have
many of the things that can be discussed. In this paper, we have studied extended
binary Golay [24, 12, 8] code with the help of MATLAB. The parity check matrix
in standard form for extended binary Golay [24, 12, 8] code is calculated and using
it, syndrome of a particular codeword is calculated. We have encoded the message
and decoded it correctly by using MATLAB. Using MATLAB, it has been verified
that, the distance of an extended binary Golay [24, 12, 8] code is 8 and it is 3-error-
correcting code.

References

1. Golay MJ (1949) Notes on digital coding. Proc IEEE 37:657

2. Lee HP, Chang CH, Chu SI (2013) High-speed decoding of the binary golay code. J Appl Res
Technol 11(3):331–337
3. Reed IS, Yin X, Truong TK, Holmes JK (1990) Decoding the (24, 12, 8) Golay code. IEE Proc
E (Comput Digit Tech) 137(3):202–206
4. Patil AR, Darkunde NS (2019) Algorithmic approach for error-correcting capability and
decoding of linear codes arising from algebraic geometry. Springer, Information and Commu-
nication Technology for Competitive Strategies, pp 509–517
5. Patil AR, Darkunde NS (2017) On some error-correcting linear codes. Asian J Math Comput
Res 56–62
6. Huffman WC, Pless V (2010) Fundamentals of error-correcting codes. Cambridge University
Press
7. MacWilliams FJ, Sloane NJA (1977) The theory of error-correcting codes. Elsevier
8. Ling S, Xing C (2004) Coding theory-a first course, 1st edn. Cambridge University Press
9. Pratap R (2017) Getting started with MATLAB: a quick introduction for scientists and engineers,
7e. Oxford University Press
Encoding Using the Binary Schubert
Code [43, 7] Using MATLAB

M. S. Wavare, N. S. Darkunde, and S. P. Basude

Abstract The main aim of this paper is to study the encoding process of binary
Schubert code [43,7] using MATLAB, when one knows its generator matrix. Using
MATLAB, we can study syndrome decoding, weight of a codeword, error correction
and error detection of the code.

Keywords Linear code · Generator matrix · Parity check matrix · Schubert code ·
Encoding

1 Introduction

In the late 1940’s due to Shannon, Hamming and Golay, the approach to error correc-
tion coding was taken by modern digital communications. Error is the most afflicted
part in any type of communication. In this receiver receives the original message with
error in it. If the system knows how much error has come with original message then
that error can be removed. So much research work has been done in the field of error
correction and detection of incoming messages. In this paper we examine the family
of Schubert unions, in particular, the binary Schubert code which were defined in [1]
and how they are used in practice with the help of MATLAB. We have designed some
representatives of generator matrix and the encoding as well as decoding process have
been discussed. Using MATLAB, the error correcting capability of defined codes in
paper [2] have been verified. Linear error correcting codes associated to Schubert
varieties which is also known as sub-varieties of Grassmannian were introduced by
Ghorpade and Lachaud [3]. These Grassmannian codes were studied by Ryan [4,
5] and Nogin [2, 6]. The upper bound for minimum distance of Schubert code was
studied in [3].

M. S. Wavare (B)
Department of Mathematics, Rajarshi Shahu Mahavidyalaya, Latur, MS, India
N. S. Darkunde · S. P. Basude
School of Mathematical Sciences, SRTM University, Nanded, MS 431606, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 725
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_72
726 M. S. Wavare et al.

2 Justification and Need for the System

The transmission process, transmitted message(signal) passes through some noisy

channel. Due to noise in this channel, some errors are introduced in received informa-
tion. We have to detect errors and then correct it using some encoding and decoding
techniques. There are two types of error control methods:
I. Error Detection with Retransmission
II. Forward acting error correction.
In the first method, retransmission request is done when some error has been
occurred in received data, whereas in the second method, the error in the received
message is detected and proper decoding technique at receiver end is applied. This
forward acting error correction technique is used when a single source transmits
signals to number of receivers. In this situation retransmission is impossible.
Error coding techniques play the important role in the digital communication. In
the simulation tool of MATLAB, we have many error controls techniques like cyclic
code, Constitutional code, linear block code, Reed Muller code and Hamming code
but there is no space for Schubert code. So main aim of this paper is to generate a
programme in MATLAB format and encode and detect the error in transmitting data
with the help of MATLAB.

3 Preliminaries

3.1 Linear Codes [7, 8]

A linear code of length n with dimension k is a linear subspace C with dimension k

of the vector space Fqn , where Fq is the finite field with q elements. Such a code is
called a q-ary code. If q = 2 or q = 3, the code is described as a binary code, or a
ternary code respectively. The vectors in C are called codewords. The size of a code
is the number of codewords in it and it is q k .

3.1.1 Basic Definitions

Let Fq denotes the finite field having q elements, where q = p h , p a prime and
h a natural number. We denote Fqn as the n-dimensional vector space over Fq . For
any x ∈ Fqn , the support of (x), is the nonzero entries in x = (x1 , x2 , . . . , xn ). The
support weight (or Hamming norm) of x is defined by, x = |sup p(x)|.
More generally, if W is a subspace of Fqn , the support of W, Supp (W) is the set of
positions where not all the vectors in W are zero and the support weight (or Hamming
norm) of W is defined by,
Encoding Using the Binary Schubert Code … 727

W = |supp (W )|

A linear [n, k]q -code is a k-dimensional subspace of Fqn . The parameters n and k
are referred to as the length and dimension of the corresponding code. The minimum
distance d = d(C) of C is defined by,

d = d(C) = min{x : x ∈ C, x = 0}

More generally, given any positive integer r, the rth higher weight dr = dr(C) is
defined by

dr = dr (C) = min{D : D is a subspace o f C with dim D = r }

Note that d1 (C) = d(C). It also follows that di ≤ d j when i ≤ j and that
dk = |sup p(C)|, where k is dimension of code C. Thus, we have 1 ≤ d1 = d <
d2 < . . . < dk−1 < dk = n. The first weight d1 is equal to the minimum distance
and the last weight is equal to the length of the code.
An [n, k]q -code is said to be nondegenerate if it is not contained in a coordinate
hyperplane of Fqn . Two [n, k]q -codes are said to be equivalent if one can be obtained
from another by permuting coordinates and multiplying them by nonzero elements of
Fq. It is clear that this gives a natural equivalence relation on the set of [n, k]q -codes.
The matrix whose rows forms basis for linear code known as generator matrix
where as the generator matrix of dual of any linear code C is known as parity check
matrix for linear code C.

3.1.2 U-Error Detecting Code [9]

Let u be any positive integer. A code C is u-error-detecting if, whenever a codeword

incurs at least one but at most u errors, the resulting word does not belong to a
code C. A code is exactly u-error-detecting if it is u-error-detecting but it can not
(u+1)-error-detecting code.
Theorem 3.1.1 [9] A code C is u-error-detecting if and only if d(C) ≥ u + 1, that
is, a code with distance d is an exactly (d −1)-error-detecting code.

3.1.3 V-Error Correcting Code [10]

Let v be a positive integer. A code C is v-error-correcting if minimum distance

decoding is able to correct v or fewer errors, assuming that the incomplete decoding
rule is used. A code C is exactly v-error-correcting if it is v-error-correcting but not
(v+1)-error-correcting.
Theorem 3.1.2 [9] A code C is v-error-correcting if and only if d(C) ≥ 2v + 1.
728 M. S. Wavare et al.

3.1.4 Binary Schubert Code [1]

Let binary matrix G is the 5 × 18 matrix given by,

⎡ ⎤
100000000011111111
⎢ 010000011111110000 ⎥
⎢ ⎥
⎢ ⎥
G = ⎢ 001000101111001000 ⎥
⎢ ⎥
⎣ 000101010101100101 ⎦
000011100101001110

Then the linear code generated by this matrix G is known as binary Schubert code
of length 18. This code is a binary [18,5,7] linear code.
Now in the next section let us consider the binary Schubert code of length 43,
which will be constructed in the similar way of binary Schubert code of length 18
[1], So for the next section we have used a binary Schubert code of length 43 and
dimension 7 to discuss the encoding of the message.

4 MATLAB Programme for Encoding of the Information

Message

clc
I=eye (7); % Identity matrix of order 7
X= [ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1;
0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0;
0 0 0 0 0 1 1 1 0 0 1 1 0 1 1 1 1 1 1 0 0 0 0 1 0 0 0 0 1 0 1 0 0 0 0 1;
0 0 0 0 1 0 1 1 0 1 0 1 1 0 1 1 0 0 1 1 0 1 0 1 0 1 0 0 0 0 1 0 0 0 1 0;
0 1 1 1 0 0 0 0 1 0 0 0 1 1 1 1 0 1 0 1 1 0 0 0 1 0 0 1 0 1 0 0 0 1 0 0;
1 0 1 1 0 1 0 1 0 0 0 0 0 1 1 1 0 1 0 0 0 0 0 1 1 0 1 0 1 1 0 0 1 0 0 0;
1 1 0 1 1 0 0 1 0 0 0 0 1 0 1 1 0 0 0 1 0 0 0 1 1 1 1 1 0 0 0 1 0 0 0 0];% Matrix X
G=[I X]; % The generator matrix for binary Schubert Code of length 43
H=mod ([-X’ eye(36)],2); % Parity Check Matrix for the defined Code
u=input (‘Enter the message bit of length 7 for which you want to encode using the
defined v=mod(u*G,2) % Encoding of message u
Synv=mod(v*H’,2) % Syndrome of encoded message
Encoding Using the Binary Schubert Code … 729

u=input (‘Enter the message bit of length 7 for which you want to encode using the
defined Enter the message bit of length 7 for which you want to encode using the
defined Code
u = [1 1 0 0 1 0 0 1]
v=
Columns 1 through 13
1001001110100
Columns 14 through 26
1001010001110
Columns 27 through 39
1101101001100
Columns 40 through 43
1101
Synv =
Columns 1 through 13
0000000000000
Columns 14 through 26
0000000000000
Columns 27 through 36
0000000000
% Use a Binary length 43
n = 18; k = 5; %Length and Dimension
clc % Clearscreen
I=eye(7); % Identity matrix of order 7
X= [ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1;
0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0;
0 0 0 0 0 1 1 1 0 0 1 1 0 1 1 1 1 1 1 0 0 0 0 1 0 0 0 0 1 0 1 0 0 0 0 1;
0 0 0 0 1 0 1 1 0 1 0 1 1 0 1 1 0 0 1 1 0 1 0 1 0 1 0 0 0 0 1 0 0 0 1 0;
0 1 1 1 0 0 0 0 1 0 0 0 1 1 1 1 0 1 0 1 1 0 0 0 1 0 0 1 0 1 0 0 0 1 0 0;
1 0 1 1 0 1 0 1 0 0 0 0 0 1 1 1 0 1 0 0 0 0 0 1 1 0 1 0 1 1 0 0 1 0 0 0;
730 M. S. Wavare et al.

1 1 0 1 1 0 0 1 0 0 0 0 1 0 1 1 0 0 0 1 0 0 0 1 1 1 1 1 0 0 0 1 0 0 0 0]; % Matrix X
G=[I X]; % The generator matrix for extended binary Schubert Code of length 43
H=mod([-X’ eye(36)],2); % Denote H as a parity Check Matrix for defined Code
slt= syndtable(H); % Produce Syndrome look up table.
w = [1 0 0 1 1 1 1 1 1 0 0 1 1 1 1 0 0 1 0 1 1 1 0 0 0 0 0 0 1 1 1 1 1 0 1 0 1 0 1
syndrome = rem(w * H’,2);
syndrome_de = bi2de(syndrome,’left-msb’); % Convert to decimal.
disp([‘Syndrome = ‘,num2str(syndrome_de), ‘ (decimal),’,num2str(syndrome),’
(binary)’])
corrvect = slt(1+syndrome_de,:) % Correction vector
% Now compute the corrected codeword.
correctedcode = rem(corrvect+w,2).

5 Conclusions

Thus, using the MATLAB, we have encoded the message information of length 43
with the help of generator and parity check matrix of given Schubert code [1] of
length 43, But due to limitation of size of matrix one cannot construct syndrome
table and can’t decode it, which we can observe from the decoding program at the
end in above section.

References

1. Wavare Mahesh S (2019) Codes associated to Schubert varieties G(2,5) over F 2 . New Trends
Math Sci 7(1):71–78
2. Yu D (2017) Nogin (1997), Spectrum of codes associated with the grassmannian G(3; 6). Probl
Inf Transm 33(2):26–36
3. Ghorpade SR, Lachaud G (2000) Higher weights of grassmann codes, coding theory, cryp-
tography and related areas. In: Buchmann J, Hoeholdt T, Stictenoth H, Tapia-Resillas H (eds)
Springer, Heidelberg, Germany, pp 120–131
4. Ryan CT (1987) An application of grassmannian varieties to coding theory. Congr Numer
57:257–271
5. Ryan CT (1987) Projective codes based on grassmann varieties. Congr Numer 57:273–279
6. Yu D (1996) Nogin, codes associated to grassmannians, arithmetic, geometry and coding theory.
In: Pellikan R, Perret M (eds) S. G. Vl_adut_, Walter de Gruyter, Berlin, New York, pp 145–154
7. Hu_man WC, Pless V (2010) Fundamentals of error-correcting codes. Cambridge University
Press
8. MacWilliams FJ, Sloane NJA (1977) The theory of error-correcting codes. Elsevier
9. Ling S, Xing C (2004) Coding theory-a first course, 1st edn. Cambridge University Press
Encoding Using the Binary Schubert Code … 731

10. Patil AR, Darkunde NS (2019) Algorithmic approach for error-correcting capability and
decoding of linear codes arising from algebraic geometry. Springer, Information and Commu-
nication Technology for Competitive Strategies, pp 509–517
11. Pratap R, Getting started with MATLAB: a quick introduction for scientists and engineers, 7e.
Oxford University Press
Data Mining Techniques for Privacy
Preservation in Social Network Sites
Using SVM

Vishvas Kalunge and S. Deepika

Abstract The rising numbers of users over the network have also raised the privacy
risk and incidents of various types of theft and attacks. Hence, the social networks
have been the major victims. The users over these networking sites share the infor-
mation under various attributes like gender, location, contact information etc. This
personal information can get compromised due to malicious act that severely violates
the integrity of the data and privacy protection policy. As a result, it has become
mandatory for a service provider to offer privacy protection before publishing any
kind of data over the network. In this research, we have proposed data mining tech-
niques for user’s privacy preservation in social network sites using Support Vector
Machines (SVMs). Social media datasets ARNET and SDFB are used for the anal-
ysis of privacy preservation models by calculating Average Path Length parameter.
Finally, the proposed model shows 1.72% and 1.46% less information loss and 1.42%
to 5.09% reduction in APL with these datasets as compared to previous works.

Keywords Privacy · Data mining · Privacy violations · SVM · Social network sites

1 Introduction

In today’s digital scenario, every day to day activity leads to large amount of data. Data
mining is used to analyse this data and with the help of various types of methods,
knowledge is discovered which is further used for many applications in different
fields. Big amount of data about different individuals can lead to privacy breach if
accessed by unauthorised person. Sensitive data may include health data, financial
data, academic data, location based servicing data, social media data etc. [1]. To
protect this data, data mining methods have to be combined with privacy preservation
techniques so that knowledge can be taken out without exploitation of personal
information related to user. Privacy Preserving Data Mining (PPDM) is extraction
of useful patterns from large set of data while preventing the release of sensitive
data of an individual [2]. Privacy of data can be viewed from 4 different roles [2,

V. Kalunge (B) · S. Deepika

Research Scholar, University of Technology, Jaipur, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 733
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_73
734 V. Kalunge and S. Deepika

Fig. 1 Roles in PPDM

3]. According to provider of data, data should be provided to the collector taking
into consideration both requirement and sensitivity of data. Data is collected by data
collectors from different data providers and it is stored in various warehouses and
databases. Modification of data is done in such a way that protection is ensured as
well as relevance of data is not lost. Data Miner uses this stored data and applies
mining algorithms to find various patterns, rules and relationships. These mined
results should also remain hidden from unauthorised users. Decision makers use
these results to take the decisions for various purposes after analysis (Fig. 1).
Data from various data sources is extracted, transformed and then loaded into the
integrated source called data warehouse. Processing is done on this warehouse using
data mining algorithms. To preserve the privacy, privacy preservation techniques
have to be applied. With this, sensitive data of user is protected as well as required
results can also be found. Privacy preserving techniques provide protection during
process of mining as well as secure the results produced by data mining [4]. Scheme
of PPDM is shown in the figure below: This study introduces concept of Privacy
Preserving Data Mining, various roles in PPDM and basic foundation of IST.

2 Literature Survey

In this section, various existing techniques related to Privacy Preserving Data Mining
are discussed on the basis of work done by different authors in this area. Literature
study had shown that have been a constant research going on for the improvement
of k-anonymity based privacy protection. Some of the important studies have been
summarized in this section. Lin and Chen [5] focussed the issues related to violation
of privacy preservation measures and postulated a PPSVC (privacy-preserving SVM
Classifier) that could transform traditional SVM into a privacy protecting classifier.
The effectiveness of the proposed design was demonstrated by the offered defence
against adversarial attacks without compromising the classification accuracy of the
classifier [5]. Okada et al. [6] proposed a k-anonymity based technique that appends
the noise to edges of social network being studied via k-neighbourhood subgraphs.
The design was implemented such that it suppressed any change in the magnitude
of edge length between nodes. Experimental evaluation had demonstrated that the
proposed design could successfully maintain the edge length of the k-anonymity
based graphs [6]. Tsai et al. [7] studied the privacy preservation issues of anonymizing
the shortest path. The privacy preservation is addressed with the concept utilizing
Data Mining Techniques for Privacy Preservation in Social … 735

the k-anonymity based identification of k-shortest privacy paths. This is achieved

with the modification of various class of edges. Experimental results demonstrated
that proposed modifications successfully achieve the shortest path privacy with vivid
levels of information loss [7]. Maheshwarkar et al. [8] analysed the security problem
that had risen when a single attribute of the datasets exhibits correlation with multiple
attributes. In the process, they proposed K-AMOA model that could successfully
solve the issues reacted to overlapping data fields [8]. Tsai et al. [9] combined the
association rule hiding techniques to k-anonymity concept to hide out the sensi-
tive information present in the data mining outcomes. The proposed design offers
an extension of the k- anonymity to achieve higher level of privacy preservation.
Experimental evaluation showed that k-anonymity application to association rule
effectively resulted in higher privacy protection with reduces utility loss [9]. Man
et al. established that the k-anonymity is not enough to offer complete security
of various sensitive data fields. To deal with the issue, they had also evaluated p-
sensitive k-anonymity privacy preservation model. The proposed solution signifi-
cantly decreased the privacy leakage issues and information loss instances thereby
improving the data quality [10]. Song et al. [11] proposed a new random k-anonymous
method to deal with the privacy issues of the network. The authors had first worked
on small data sets and analysed the computational cost. Further k- anonymity is
analysed with the addition of noise to the data that offered reduction of information
loss along with defending against background, homogeneity and exhaustive attacks
[11]. Shailaja et al. [12] developed a Privacy-Preservation Data Mining approach that
involved both data restoration and refinement. In both the processes, optimal selec-
tion is done using modified cuckoo based opposition intensity method. It has been
established that the modified cuckoo search could effectively reduce false rule gener-
ation, failure rate while offering higher degree of information preservation rate [12].
Namdarzadegan and Khafaei [13] designed a model to address the privacy preserva-
tion that combines the strengths of k-anonymity with the Cuckoo based optimization
approach. The proposed model was evaluated based on transitivity, clustering coef-
ficient and average path length. Experiments demonstrated that the proposed model
is found to be one-unit superior in comparison to k-anonymity method [13].

3 Proposed Methodology

The proposed design involves optimization of the sensitive data that is to be dissem-
inated over the web. Enhanced k-anonymity based privacy protection is achieved
by passing information to Cuckoo Search (CS) for optimization followed by clas-
sification carried on with Support Vector Machine (SVM). The privacy protection
offered by the proposed design is evaluated with experimental calculation of APL
and Information loss against larger data sets. The steps of proposed methodology are
explained with block diagram in Fig. 2.
In the proposed design, two databases are used representing larger data sets of
social networking sites. ARNET database offers comprehensive data mining covering
736 V. Kalunge and S. Deepika

Fig. 2 Flow diagram of proposed design

social academic networks. The initial citation network was demonstrated by 629,814
papers covering 632,752 citations [14]. SDFB is Six Degrees of Francis Bacon that
corresponds to digital information of relationships of early modern social network.
The dataset summarizes the information covering 171,419 relationships established
between 15,824 EMSN. Further, data set is divided into 109 different labelled groups
whose relationships are summarized under 64 categories [15].
Cuckoo Search is a nature inspired algorithm that is inspired with the reproduction
and egg laying behavior of cuckoo birds. To reach an optimal solution CS observes
following assumptions:
• Cuckoo randomly searchers for the nest to lay eggs in the nests.
• Only the best qualities of eggs are preserved and the nest containing them is further
carried to next generation.
• Moreover there exists the probability that the host cuckoo could identify the
foreign egg.
SVM is used as a classifier to the output of the Cuckoo Search Algorithm. As
SVM is a binary classifier, hence in order to provide training to the classifier, the
data is being divided into two classes namely C1 and C2. C1 has values which were
placed in noisy nodes 1 and C2 has values which were placed in noisy nodes 2. The
average APL of every node is passed as the training data.
Data Mining Techniques for Privacy Preservation in Social … 737

Algorithm for SVM:

4 Results

Thee evaluation of the proposed design in terms of experimental calculation of infor-

mation loss and APL was performed. Here, the mathematical calculations and results
are explained. It defines the average value of the number of steps that are required
to follow the shortest possible path between a pair of points representing network
nodes. Mathematically it can be calculated as follows:
  
i=j d pi , pj
APL = (1)
(n − 1)

where (pi, pj) represents the shortest path length existing between two points that is
divided by the total number of possible paths. The observed average path length of the
proposed design using ARNET dataset against two existing studies is summarized
in Table 1. In 2019 Namdarzadegan and Khafaei had implemented Cuckoo based
optimization [13] while Macwan et al. in [16] demonstrated the effectiveness of
k-anonymity model [16] for privacy preservation of social networks.
Comparative analysis of APL with the existing studies for ARNET dataset is
shown in Fig. 3 where node count is plotted on X-axis ranging from 10 to 100
nodes against the APL values on Y-axis. It is observed that the line graph corre-
sponding to the proposed work exhibited lowest APL as compared to the existing
work of Namdarzadegan and Khafaei [13] and Macwan et al. [16]. The graph shows
that average APL of 0.397 is observed for Namdarzadegan and Khafaei, 0.434 for
Macwan et al. with lowest APL of 0.3865 for the proposed design. This means that
738 V. Kalunge and S. Deepika

Table 1 APL comparison for ARNET dataset

Node count Proposed-APL APL-Namdarzadegan and APL-Macwan et al. [16]
Khafaei [13]
10 0.356 0.373 0.416
20 0.374 0.382 0.417
30 0.375 0.384 0.422
40 0.379 0.387 0.421
50 0.381 0.389 0.435
60 0.384 0.401 0.436
70 0.389 0.407 0.439
80 0.402 0.409 0.445
90 0.412 0.415 0.449
100 0.413 0.423 0.457

Fig. 3 APL comparison for ARNET dataset

the proposed design is 1.05% better than Namdarzadegan and Khafaei work and
4.72% better than Macwan et al.’s work.
Next, APL comparison is also carried on with SDFB dataset and the corresponding
observed APL values are listed in Table 2. In column 2 lists APL values of the
proposed work against node count varying from 10 to 100. Also, the APL values
corresponding to Namdarzadegan and Khafaei and Macwan et al. work are listed in
column 3 and column 4 respectively. The table shows that magnitude of the APL
values is lowest for the proposed work in comparison to the existing works.
Figure 4 shows the APL comparison for SDFB dataset. It is observed that for SDFB
dataset our proposed design achieved an average APL value of 0.382 in comparison
to the average APL of 0.397 demonstrated by Namdarzadegan and Khafaei [13] and
Data Mining Techniques for Privacy Preservation in Social … 739

Table 2 APL comparison for SDFB dataset

Node count Proposed-APL APL-Namdarzadegan and APL-Macwan et al. [16]
Khafaei [13]
10 0.368 0.373 0.416
20 0.379 0.382 0.417
30 0.375 0.384 0.422
40 0.381 0.387 0.421
50 0.379 0.389 0.435
60 0.383 0.401 0.436
70 0.385 0.407 0.439
80 0.386 0.409 0.445
90 0.391 0.415 0.449
100 0.401 0.423 0.457

Fig. 4 APL comparison for SDFB dataset

0.434 by Macwan et al. [16]. Overall, our design exhibited 1.42% and 5.09% lower
APL as compared to Namdarzadegan and Khafaei and Macwan et al., respectively.
Information Loss
It is another parameter that evaluates the effectiveness of the proposed design in
achieving privacy preservation. It represents the percentage of information that get
compromised and lost as a result of data transfer. A better privacy protection model
always results in lower information loss. In the current study, it is calculated using
both ARNET and SDFB datasets.
Table 3 compares the information loss for ARNET dataset of the proposed design
740 V. Kalunge and S. Deepika

Table 3 Information loss

Node count Proposed-information loss Song et al. [11]
comparison for ARNET
dataset 10 30.8 32
20 30.9 32.4
30 31.3 32.7
40 31.1 32.9
50 31.5 33.1
60 31.4 33.6
70 31.9 33.4
80 32.1 34.1
90 32.2 34.1
100 32.4 34.5

with information loss observed by Song et al. who had implemented random k-
anonymous method to achieve privacy protection [11]. Table shows that compara-
tively lower information loss is observed for the proposed work over all the node
counts.
Information loss for ARNET over 10 to 100 nodes is compared in Fig. 5. The
line graph plots node count on X-axis against the information loss values plotted on
Y-axis. It is observed that information loss of the proposed design varied from 30.8
to 32.4% with an average information loss of 31.56%. Information loss observed for
privacy preservation method of Song et al. exhibited an average value of 33.28%.
In other words, our proposed design demonstrated 1.72% lower information loss for
ARNET dataset.
Further, information loss is also calculated for SDFB dataset and the resultant
parametric values of information loss are summarized in Table 4. Information loss
of the proposed design is given in column 2 and Song et al. is given in column 3.

Fig. 5 Information loss for ARNET dataset

Data Mining Techniques for Privacy Preservation in Social … 741

Table 4 Information loss for

Node count Proposed-information loss Song et al. [11]
SDFB dataset
10 31.2 32
20 31.5 32.4
30 31.4 32.7
40 31.5 32.9
50 31.6 33.1
60 31.7 33.6
70 31.9 33.4
80 32.3 34.1
90 32.5 34.1
100 32.6 34.5

It is also observed that information loss of the proposed design for SDFB dataset
lies between 31.2 and 32.6% which is little higher than information loss that was
observed for ARNET dataset.
Figure 6 compares the information loss of SDFB dataset for our proposed work
and Song et al.’s work. The graph shows that the proposed work demonstrated an
average information loss of 31.82% while Song et al.’s work demonstrated an average
information loss 33.28%. It means that the proposed work exhibited 1.46% lower
information loss corresponding to the effectiveness of our proposed design.

Fig. 6 Information loss for SDFB dataset

742 V. Kalunge and S. Deepika

5 Conclusion

Today’s time is the time of Big Data. Useful information extracted from the large
amount of data helps to take very important decisions in various fields and orga-
nizations but they can lead to privacy breach of an individual. The need of hour is
to protect sensitive data of an individual and hence, data mining algorithms have to
be combined with privacy preservation techniques. The present work had proposed
an enhanced k-anonymity based privacy protection where cuckoo search optimiza-
tion is used. The fitness function based improved privacy protection model is further
cross validated using SVM. QoS parameters namely, information loss and APL were
evaluated for ARNET and SDFB datasets over the node count ranging from 10 to
100 to show the effectiveness of the implemented techniques. It is observed that APL
of the proposed work for ARNET dataset is 1.05 and 4.72% while for SDFB dataset
is 1.42 and 5.09% lower when compared with Namdarzadegan and Khafaei’s and
Macwan et al.’s work, respectively. Similarly, information loss of the proposed work
also showed decrease of 1.72 and 1.46% for ARNET and SDFB dataset as compared
to information loss observed with Song et al.’s work. Hence, the proposed design
proved to be highly effective enhanced k-anonymity based privacy preservation.

References

1. Mendes R, Vilela JP (2017) Privacy-preserving data mining: methods, metrics, and applications.
IEEE Access 5:10562–10582
2. Bhandari N, Pahwa P (2019)Comparative analysis of privacy-preserving data mining tech-
niques. In: International conference on innovative computing and communications. Springer,
Singapore, pp 535–541
3. Xu L, Jiang C, Wang J, Yuan J, Ren Y (2014) Information security in big data: privacy and
data mining. IEEE Access 2:1149–1176
4. Siraj MM, Rahmat NA, Din MM (Feb 2019) A survey on privacy preserving data mining
approaches and techniques. In: Proceedings of the 2019 8th international conference on
software and computer applications. ACM, pp 65–69
5. Lin KP, Chen MS (2010) On the design and analysis of the privacy-preserving SVM classifier.
IEEE Trans Knowl Data Eng 23(11):1704–1717
6. Okada R, Watanabe C, Kitagawa H (2014) “A k-anonymization algorithm on social network
data that reduces distances between node”. In: 2014 IEEE 33rd international symposium on
reliable distributed systems workshops. IEEE, pp 76–81
7. Tsai YC, Wang SL, Kao HY, Hong TP (2015) Edge types versus privacy in K-anonymization
of shortest paths. Appl Soft Comput 31:348–359
8. Maheshwarkar B, Patidar P, Rawat MK, Maheshwarkar N (2016) “K-AMOA: K-anonymity
model for multiple overlapped attributes”. In: Proceedings of the second international confer-
ence on information and communication technology for competitive strategies. ACM, p
83
9. Tsai YC, Wang SL, Song CY, Ting IH (2016) Privacy and utility effects of k-anonymity on
association rule hiding
10. Casino F, Domingo-Ferrer J, Patsakis C, Puig D, Solanas A (Sep 2013) Privacy preserving
collaborative filtering with k-anonymity through microaggregation. In: 2013 IEEE 10th
International conference on e-business engineering. IEEE, pp 490–497
Data Mining Techniques for Privacy Preservation in Social … 743

11. Song F, Ma T, Tian Y, Al-Rodhaan M (2019) “A new method of privacy protection: random
k-anonymous”. IEEE Access
12. Shailaja GK, Rao CG (2019) “Opposition intensity-based cuckoo search algorithm for data
privacy preservation”. J Intell Syst
13. Namdarzadegan M, Khafaei T (2019) “Privacy preserving in social networks using combining
cuckoo optimization algorithm and graph clustering for anonymization”. Asian J Res Comp
Sci 1–12
14. Tang J, Zhang J, Yao L, Li J, Zhang L, Su Z (2008) “Arnetminer: extraction and mining of
academic social networks”. In: Proceedings of the 14th ACM SIGKDD international conference
on knowledge discovery and data mining. ACM, pp 990–998
15. SDFB Team, six degrees of Francis Bacon: reassembling the early modern social network.
www.sixdegreesoffrancisbacon.com. Accessed URL https://www.kaggle.com/rtatman/six-deg
rees-of-francis-bacon
16. Macwan KR, Patel SJ (2017) k-Degree anonymity model for social network data publishing.
Adv Electr Comp Eng 17(4):117–125
Near Field Communication (NFC)
Technology and Its Application

R. D. Kulkarni

Abstract Near Field Communication (NFC) as a form of wireless technology has

seen many improvements in recent years due to the increasing availability of NFC
enabled devices. NFC is a recently emerging technology for short range commu-
nications aimed to enhance existing near field technologies such as RFID (Radio
Frequency Identification). NFC is a standards-based, short-range (a few centimeters)
wireless connectivity technology that enables simple and safe two-way interactions
between electronic devices, allowing consumers to perform contactless transactions,
access digital content, and connect electronic devices with a single touch. In this
review paper, NFC technology is put forward with respect to its implementation,
operating modes, its application in the form of tags as well as payments and its
standards and protocols.

Keywords NFC · NFC architecture · NFC security · NFC application

1 Introduction

NFC stands for Near Field Communication, which is a short distance, low data rate
and low cost protocol. It is complementary to WIFI and Bluetooth and is a subset
of 13.56 MHz RFID. NFC is deployed to enable contactless transactions for data
exchange and simplified setup of more complex communications such as Bluetooth
and WIFI. For instance, version 4.0 and 4.1 of Bluetooth low energy has no secure
way of exchanging the temporary key for encryption without the help of NFC [1].
Hence, NFC is poised to play an important role in the connectivity for the “Internet
of Things.” In addition, NFC is widespread in smart phone markets, as in 2015,
50% of smartphones were compatible with NFC. NFC is gaining presence in more
and more applications, from POS terminals for tap & pay applications to consumer
environments with tap & pair application to home appliances and healthcare for tap
& exchange applications.

R. D. Kulkarni (B)
SVERI’s COE, Pandharpur, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 745
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_74
746 R. D. Kulkarni

The difference between NFC and RFID can sometimes raise confusion, as the
terms are sometimes used synonymously. However, NFC is actually an extension
of RFID. NFC is designed to build on RFID by enabling more complex exchanges
between devices [1]. The root cause of the confusion is the fact that it’s still possible
to read passive RFID tags with an NFC reader and to write to a limited amount of
memory. RFID was designed for identification. RFID tags can hold a small amount
of data, and you can read and write to them from RFID readers, but the amount of
data is limited to a few thousand bytes [2]. An RFID is passive when it receives
power from the reader, but it is active when it has its own power source.
An RFID exchange involves two factors: a target and an initiator. The initiator,
a tag reader or reader/writer, starts the exchange by generating a radio field and
listening for responses from any target in the field. The target, a tag, responds when
it picks up a transmission from an initiator. It will respond with a unique identifier
number (UID) [2].
RFID has various standards defined by ISO (International Standards Organiza-
tion) [2]. A standard defines radio frequency, data transfer rates, data formats, and
(optionally) layers of the protocol. It can be rather specific—for instance the ISO-
11784 standard was developed for animal tracking! The most common one is the
ISO-14443 standard that was developed for use with payment systems and smart
cards. Many NFC applications were developed based on this standard.
As said previously, NFC can be thought of as an extension of RFID and hence
shares many properties (active or passive communication modes, how to initiate
communication and hardware layer). But, instead of just delivering static data from
memory, an NFC target can generate unique content for each exchange and deliver
it back to the initiator. For example, if you’re using NFC to exchange address data
between two phones, the NFC target device can be programmed to only provide
limited information if it has never seen this particular initiator before.
NFC is becoming ubiquitous, but due to many terminologies and standards, taking
the first steps can be overwhelming. This article aims to clarify and simplify those
steps for newcomers.

2 Communication Ways

Communication in NFC is either in active mode or passive mode. Active device is

the one that generates RF and has its own power supply [3]. The passive device is
powered by another active device [4]. Following are the two communication ways.
(1) Two-way communication: Devices that are capable of reading and writing to
each other. For example, using NFC, you can touch both Android devices
together to transfer data like contacts, links, or photos.
(2) One-way communication: One-way communication: Reading and writing to
an NFC chip is done by a powered device (like a phone, credit card reader,
Near Field Communication (NFC) Technology and Its Application 747

or commuter card terminal).Hence, when a commuter card is tapped on the

terminal, money is subtracted from the balance by the NFC powered terminal.

NFC Modes
The NFC forum defines three communication modes, as in Fig. 1 and Table 1:

Fig. 1 Forum specification architecture

Table. 1 NFC terminologies

NFC RFID Bluetooth V2.1 IrDA
Information Coupling of Magnetic field Electro magnetic Infrared light
Transmission magnetic field radiation
Operating 13.56 MHz 13.56 MHz 2.54 GHz ~2 MHz
frequency
Modes Active-active, Active-passive Active-active Active-active
active-passive
Transmission range 0.04−0.1 m Up to 1 m 10 − 100 0−2 m
Network type Point-to-point Point-to-point Point-to- multiple Point-to-point
Communication Two way One Way Two way One Way
Maximum data rate 424 kbps 128 kbps 2.1 mbps 16 mbps
Setup time <0.1 s <0.1 s ~6 s ~0.5 s
Maximum current <15 mA <15 mA <30 mA <5 mA
consumption
Authentication and Yes Yes Yes No
encryption
Standards ECMA 340 ISO/IE C IEEE 802.15.1 IEEES 02. II
748 R. D. Kulkarni

1. Peer-to-Peer mode is defined for device to device link-level communication.

Note that this mode is not supported by the Contactless Communication API
[5, 6].
2. Read/Write mode allows applications for the transmission of NFC Forum-
defined messages. Note that this mode is not secure [7]. This mode is supported
the Contactless Communication API.
3. NFC Card Emulation mode allows the NFC-handset behave as a standard
Smartcard. This mode is secure. This mode is supported by the Contactless
Communication API.
NFC Terminology
1. NDEF—NFC Data Exchange Format—standard exchange formats for URI,
Smart Posters, other
2. RTD—Record Type Definition—An NFC-specific record type and type name
which may be carried in an NDEF record
3. NDEF message—Basic message construct defined by this specification. An
NDEF message contains one or more NDEF records
4. NDEF record—Contains a payload described by a type, a length, and an
optional identifier
5. NDEF payload—The application data carried within an NDEF record.
NFC compares to the other short-range communication technologies.

3 Security Concerns

As mentioned, most NFC scenarios are required to deal with sensitive data like
credit card numbers, bank account details, account balances, personalized tickets or
other personal data. For data storage and wireless data transfer security is therefore an
essential issue. NFC thereby provides several mechanisms for security and immunity.
First of all, there is obviously a certain physical security due to the touch-to-connect
principle. As a matter of fact, the NFC technology only provides data transfer between
two devices or between a device and a tagged object when the distance between the
two items falls below a certain range. The usual transmit power of radio frequency
and the receive sensitivity of NFC readers that fulfil the previously described ISO
specifications are only strong enough to operate up to a range of a few centimetres [3].
That means that e.g. at a supermarket checkout the customer needs to place his phone
directly over the reading device. Data skimming, that is capturing and intercepting
transferred data by a distant attacker, is hence not possible that easily. Misuse is
however possible in the form of relay attacks. Such attack is basically accomplished
via an attacker serving as a man-in-the-middle who is forwarding transmitted data
between a reader, e.g. a reading device for NFC payments, and a target transponder,
e.g. a NFC device serving as a credit card, that is actually out of the reading range. The
following example illustrates the procedure: An attacker places his modified NFC
phone on top of the NFC reader at the POS, e.g. at a checkout in the supermarket.
Near Field Communication (NFC) Technology and Its Application 749

The phone however forwards the data, e.g. via Bluetooth, to his distant accomplice
holding another modified NFC phone. He in turn holds his phone—without attracting
attention—next to a contactless NFC smart card, e.g. NFC credit card, of the attacked
target person. The accomplice’s phone operates as NFC reader simulating the actual
reading device at the POS. The reader at the POS assumes the target person’s card
is close, and unknowingly debits his account. The additional data exchange and
forwarding via the attacking person-in-the-middle however naturally takes some
extra time. A possible countermeasure for such relay attack could for example be
built upon a quite short timeout threshold that avoids transactions if the response
time is too slow. The concept of Google Wallet is however also secured against such
attacks as thereby a PIN has to be entered in order to activate the phone’s NFC
broadcast hardware and to activate the Secure Element that is storing all the sensitive
data. The Wallet PIN also prevents unauthorized usage of the payment card in case
the NFC phone is stolen.
For the wireless channel communication itself the NFC specifications do not
ensure any secure encrypting mechanisms. On higher layers however, NFC appli-
cations can of course use industry-standard cryptographic protocols like SSL/TLS
based methods to encrypt the data that is to be transferred over the air and that is
stored in the Secure Element. Other possible NFC vulnerabilities involve the manip-
ulation of NFC tags [3]. Existing passive NFC tags, e.g. on a smart poster, could
be replaced by spoofed tags such that a modified NDEF message is read by a NFC
reading device. Possible attacks could for example replace the URI record with a
malicious URL, e.g. in order to make the user load a phishing website that steals
the users credentials. Experiments show that one can modify the NDEF message of
a smart poster such that the visualized content on the phone, e.g. the website URL,
is hardly distinguishable from the original one. Furthermore, it is even possible to
create a malformed NDEF message that causes the some applications to crash. This
form of manipulation could be used by attackers to debase the relationship between
a user and the pretended service provider. The Signature Record Type that was previ-
ously addressed in this paper and that signs a whole NDEF message can however
serve as a countermeasure against URL spoofing and similar attacks. By manipu-
lating the signed tag payload a signed NDEF message will lose its integrity and will
be recognized as not-trusted.

4 Application of NFC

Applications as shown in Fig. 2 are explained below:

1. Smart Cards: Payment using NFC integrated smart cards offers easier payment
compared to conventional multiple step payment process. Top payment services
like Visa and MasterCard are offering NFC embedded smart cards to customers.
2. E-Wallet: Cashless payment system using mobile devices became popular in
the beginning of this decade and more services are offering cashless payments
750 R. D. Kulkarni

Fig. 2 Applications of NFC

for customer’s convenience. Using smart phone applications, payments can be

made using a simple tap or waving the card within the proximity [7].
3. Smart Ticketing: Integrated smart chips can be used to replace traditional tick-
eting systems with smart tickets for airlines, train and bus tickets etc.… NFC tags
can be used for Smart posters, movie tickets, ticket to concerts, advertisements,
flyers and information links.
4. Medicine and healthcare: NFC integrated system can be used in medicine
and healthcare activities. NFC offers greater accuracy and convenience in
prescribing medicine, easier check-in, payments, checking status of patients,
tracking records by embedding NFC tags to patient’s charts.
5. Keyless Access: Keyless access is one of the familiar applications of near field
communications today. NFC’s convenience and easy to implement feature make
it a popular choice. NFC and RFID tags can be used for access to doors and
restricted areas with an auto detect feature. It can be used to replace access keys,
identification badges and for easier access to cars and other vehicles [7].
6. Logistics and Shipping: NFC and RFID tags can be conveniently used in logis-
tics and shipping industry. Tracking and scanning of goods using smart tags
make the system smart, errorless and efficient.
7. Smart Inventory Management: Retail shops and large scale super markets
can make use of smart RFID tags for better management of inventories in their
system. Smart inventory management software can give a real-time update on
product details for customers, items in their inventory stock and it could trigger
automatic order if a particular item has low quantity [8].
Near Field Communication (NFC) Technology and Its Application 751

5 Conclusion

NFC is a wireless technology for short range data transmission. It operates in three
modes via two types of communication. RFID and magnetic induction are the key
factors in implementing NFC. Few forms of implementing this technology is through
NFC tags and NFC payment [7]. Through NFC Payments transactions are secure and
carried out with just a tap. The SE is responsible for security, authenticity and data
confidentiality with respect to payments. NFC tags will play inevitable role in future
smart devices for more integrated functions, smart transportation, aviation industry,
shipping, manufacturing industry for automation of particular tasks. Integrating NFC
technology with our modern data communication and transaction process ensures
convenience, time saving, energy efficiency and most importantly improved security.

References

1. NFC-enabled cellphone shipments to soar fourfold in next five years | IHS Online Newsroom,
press.ihs.com, (2014) [Online]. Available: https://press.ihs.com/press-release/design-supply-
chain/nfcenabled-cellphone-shipments-soar-fourfold-next-five-years
2. Klaus F (2003) In RFID handbook aplications, technology
3. Haselsteiner E, Breitfuß K (2006) Security in near field communication (NFC). In: Workshop
on RFID security
4. https://www.cnet.com/how-to/how-nfc-works-and-mobile-payments
5. Langer J, Roland M (2010) Anwendungen und technik von near field communication (NFC),
Anwendungen und Tech. von Near F. Commun. NFC, pp 205–241
6. https://nfc-forum.org/resources/what-aretheoperating-modes-of-nfc-devices/
7. EMVCo (Nov 2011) https://www.emvco.com/
8. El Madhoun N, Bertin E (2017) Magic always comes with a price: utility versus security for
bank cards. In: IEEE cyber security in networking conference (CSNet’17)
9. Sherringham K, Unhelkar B (2009) In handbook of research in mobile business. 2nd edn
10. Agbinya JI, Masihpour M (2012) Power equations and capacity performance of magnetic
induction communication systems. J Wirel Pers Commun 64(4):831–845
Performance Analysis of Convolution
Neural Network with Different
Architecture for Lung Segmentation

Swati P. Pawar and S. N. Talbar

Abstract As lung cancer is one of the significant causes of death, there is a need for
the development of algorithms for early detection of these cancers. Early detection
of lung cancer helps to provide appropriate treatment and reduce morbidity. Accu-
rate segmentation of the lung is an essential step in every computer-aided diagnosis
(CAD) system to provide an accurate lung CT image analysis. This study is focused
on the design of the appropriate architecture of the convolution neural network (CNN)
using suitable combinations of CNN blocks to improve lung segmentation efficiency.
Based on the scientific intuition, three CNN architectures are proposed for effective
segmentation of lung parts from CT images. These CNN architectures are varied by
the depth of down sampling of images as 32 × 32, 16 × 16 and 8 × 8. The perfor-
mances of these CNN are obtained as under segmentation or over-segmentation by
comparing the segmented lung part with ground truth lung images. This performance
analysis shows the segmentation efficiency greatly affected by appropriate selection
of downsampling of these images.

Keywords CNN · Lung-segmentation · Depth of downsampling · Segmentation

performance

1 Introduction

With the advancement in computational efficiency and improved image processing

algorithms, manual, labour-intensive image tasks can be automated through comput-
erized algorithms. One of such essential problems which involve the handling of
several images per patient is CT scan based lung cancer diagnosis. As this has been
one of the leading causes of death globally, it becomes more appropriate to develop
advanced image processing algorithms for lung cancer early detection process. The

S. P. Pawar (B) · S. N. Talbar

Sveri’s College of Engineering, Pandharpur, India
e-mail: [email protected]
S. N. Talbar
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 753
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_75
754 S. P. Pawar and S. N. Talbar

early detection of lung cancer helps to minimize the severity of cancer by reducing
its morbidity [1]. As Computer Tomography is the best method to detect the presence
of lung cancer, the proposed algorithm is demonstrated using CT scan images. Lung
segmentation is one of the important steps in the process of computer-aided diagnosis
(CAD) multi-stage process. The proposed study is focused on the development of
the suitable architecture of CNN for lung segmentation.
ILD (Interstitial lung diseases) is a group of diseases that contain 150 histo-
logical diagnoses associated with disorders of the lung parenchyma [2] that even-
tually affect breathing and cause progressive scarring of lung tissues. ILD lung
tissue patterns intensity is similar to the chest region which becomes a hurdle in
lung segmentation. Thus, most existing methods fail [3] to include such ILD lung
tissue patterns in segmentation which leads over-segmentation or under segmenta-
tion problem. To overcome this limitation, we proposed a robust method for accurate
lung segmentation.
The lung segmentation task is crucial to find out the accurate boundary of lung
parenchyma so that we cannot mislead lung nodules. There are various approaches
that have been proposed for lung segmentation. Initially, [1, 4, 5] grey-level thresh-
olding based approach for segmentation of lung from the non-lung region. Brown
et al.[1] presented an automated knowledge-based method to segment lung regions
from CT image based on volume, shape, relative position of the organ. Sun et al.[5]
developed a 3D region-growing algorithm to segment lung volume based on a fuzzy
logic approach and 3D morphological closing operation. Threshold-based techniques
ignore lung regions that have more attenuation, such as in the presence of interstitial
lung disease. However, these conventional methods failed in the presence of dense
abnormalities such as lung nodule, fibrosis etc.
In literature, methods proposed in [2, 6] developed to include dense abnormalities
to get accurate lung segmentation. Pu et al. [6] presented an adaptive border marching
algorithm that includes juxtapleural nodules, but this algorithm failed to include
abnormal lung tissue patterns. Further, Prasad et al. [7] proposed fixed threshold-
based techniques that included lung boundary curvature features which can help to
include abnormal lung tissue patterns. However, the performance of this method
depended on accurate rib segmentation.
Another shape-based approach developed for lung segmentation by Ye et al. [3]
for the inclusion of abnormal lung tissue patterns. They proposed a two-step segmen-
tation method to extract lung region in the first step they used to extract initial lung
mask using fuzzy thresholding method in the second step lung contour chain code
was used to obtain a complete mask. However, this method led over-segmentation
in the inclusion of large abnormal lung tissue patterns.
Further, Shen et al. [8] have proposed A bidirectional chain coding method
combined with a support vector machine (SVM) classifier is used for inclusion of
juxtapleural nodules. However cascade nature of the method, system performance
depends on scan quality of CT image, classifier and threshold.
In the existing literature, intensity-based approaches have been proposed [9, 10]
intensity-based method methods that rely on robust feature extraction. Hosseini-Asl
et al. [9] have proposed incremental constrained nonnegative matrix factorization to
Performance Analysis of Convolution Neural Network … 755

extract robust features. Further, in [10] the author proposed a 3D lung segmentation
method based on the new Nonnegative Matrix. Factorization approach to impose
incremental constraints on it. Method [9, 10] includes juxtapleural nodules but they
failed in the case of dense abnormalities.
Above lung segmentation methods failed in case of inclusion of lung tissue
patterns. Therefore, there is a need to develop a strong learning mechanism for
lung segmentation.
This paper is focused on the design of c-GAN based approach Convolution neural
network with different architecture for lung segmentation. In this study, three different
CNN networks are proposed with architectures varied by the depth of downsampling
of images as 32 × 32, 16 × 16 and 8 × 8. The lung segmentation performances of
these networks are studied to form the segmentation map of the CT slice without any
post-processing of these images. The performance of these networks is studied for
segmentation of lung in the presence of lung nodule abnormalities.

2 Design of Lung Segmentation CNN Network

Convolution Neural Networks (CNN) is one of the advanced algorithms used in

complex image processing. Lung segmentation method becomes complicated due to
a large difference in attenuation between the lung parenchyma and the surrounding
tissue. Therefore, the threshold-based lung segmentation is the simplest way to obtain
the lung field mask. However, in the presence of various abnormalities in the lung
tissue patterns viz. fibrosis, ground-glass, reticulation, consolidation, emphysema,
lung nodule, etc., these algorithms fail in giving accurate lung segmentation. These
abnormal lung tissue patterns often cause under-segmentation and thus decrease the
performance of CAD systems. Therefore, a suitable design of CNN becomes the
most appropriate algorithm for the lung segmentation problem.
The proposed CNN for lung segmentation consists of two significant operations
as generator and discriminator. In the generator operation, the network learns to map
between the input lung CT image and the reference ground truth lung segmentation
map. On the other hand, in the discriminator process, the network learns to discrim-
inate between the generator output and the reference lung segmentation map. This
study utilizes the discriminator network proposed [11] to discriminate between the
generator output and the reference lung segmentation map. The generator network
consists of (1) encoder block (2) multi-scale dense feature extraction (MSDFE)
module and (3) decoder block. In this study, the three architectures of CNN are
proposed based on the depth of downsampling of images by varying the number of
blocks of MSDFE blocks.
The Encoder and decoder blocks are formed by convolution/deconvolution layer
followed by nonlinear activation function i.e. ReLU. Network feature maps are
normalisedusing Instance normalization [12]. Each block of these encoders/decoders
are formed by 3 × 3 sized convolution/deconvolution layer filters. The encoder block
756 S. P. Pawar and S. N. Talbar

encoded the input lung CT slice into feature maps and followed by downsampling
the encoded feature maps.
A multi-scale dense feature extraction (MSDFE) module [11] which consists of
four inception blocks connected serially through the dense connection between is
utilized for robust feature extraction.
The MSDFE plays an important role in dense feature extraction based on which
the depth of downsampling varies. To understand the influence of these downsam-
pling on the segmentation performance, three CNN architectures with the depth of
downsampling of images as 32 × 32, 16 × 16 and 8 × 8 are studied. Figure 1 shows
the CNN network which comprises four MSDFE which downsamples the images to
32 × 32 whereas Fig. 2 shows the network with six MSDFE and Fig. 3 shows eight
MSDFE which downsamples the images to 16 × 16 and 8 × 8, respectively. In these
three networks half of the MSDFE modules obtain multi-scale features processed
through the simple convolution layer (filter size of 3 × 3) having stride factor =
2 whereas the remaining half MSDFE modules, to maintain the symmetry in the
network, to obtain multi-scale feature maps processed through the simple decon-
volution layer (filter size of 3 × 3) having up-sampling factor = 2. Therefore, the
network proposed in Fig. 1 downsamples the images of 128 X 128coming out of

Fig. 1 CNN network with downsampling to 32 × 32 size (CNN 32 × 32)

Fig. 2 CNN network with downsampling to 16 × 16 size (CNN 16 × 16)

Performance Analysis of Convolution Neural Network … 757

Fig. 3 CNN CNN network with downsampling to 8 × 8 size (CNN 8 × 8)

convolution layer as 128 × 128 → 64 × 64 → 32 × 32 in the downsampling phase

whereas it transforms as 32 × 32 → 64 × 64 → 128 × 128 in upsampling phase.
Similarly, the network in the Fig. 2, in the downsampling phase the images coming
from convolution layer downsamples as 128 × 128 → 64 × 64 → 32 × 32 → 16 ×
16 whereas upsampling phase the images transform as 16 × 16 → 32 × 32 → 64 ×
64 → 128 × 128 and the network in the Fig. 3 downsamplesas 128 × 128 → 64 ×
64 → 32 × 32 → 16 × 16 → 8 × 8 whereas in the upsampling phase it transforms
as 8 × 8 → 16 × 16 → 32 × 32 → 64 × 64 → 128 × 128.
The proposed networks consist of two parts of viz. generator and discriminator. In
the generator part of the network, mapping is formed between input lung CT images
and reference lung segmentation map. The network weight parameters are obtained
using adversarial loss approach [11]. Objective of conditional GAN is to minimize
adversarial loss against discriminator, which tries to maximize it.

3 Performance Analysis of the Proposed Networks

Lung segmentation performance of the proposed three networks is studied using

the benchmark interstitial lung disease (ILD) dataset [2]. The performance of the
lung segmentation is studied in the presence of Lung Nodule abnormalities in the
CT scans. In total, 4000 lung CT slices with lung nodule lung disease are created,
and their corresponding lung field segmentation maps to train the proposed CNN
networks. An adversarial loss and L1 loss parameters are considered as training
criteria. Other parameters of the model are similar to [54]. A PC with 4.20 GHz Intel
Core i7 processor and NVIDIA GTX 1080 8 GB GPU is used to train the proposed
network for accurate lung field segmentation.
After training of the network qualitative performance analysis of the proposed
three networks has been carried out using the CT scans from the ILD database. For the
758 S. P. Pawar and S. N. Talbar

(a) (b) (c) (d) (e)

Fig. 4 Lung segmentation results for under segmentation cases a Input lung CT slice b CNN (8 ×
8) c CNN (32 × 32) d CNN (16 × 16) e Ground truth

qualitative performance analysis, these cases from the database are presented in the
form of under-segmentation, over-segmentation and complex cases. Figure 4 shows
the under segmentation cases for three different networks proposed and compared
with the ground truth. The red colour indicates the under-segmented part whereas
blue colour represents the over segmented part. For the cases shown in the image,
the network with downsampling 16 × 16 shows good performance of segmentation
showing close to the ground truth images whereas other two networks fail to outline
the complex areas of the lung and do not include some of the parts of the lungs. Even
in some cases, the network with 8 × 8 downsampling performs poorly as compared
to 32 × 32 downsampling network which may be due to densely mapping of the
features resulting in the exclusion of some of the complex parts of lungs.
Figure 5 shows the over-segmentation cases from the testing of segmentation
efficiency of three networks proposed. For these cases also the network with down-
sampling to 16 × 16 gives fairly good results as compared to the other two networks.
However, these two networks give better segmentation than under-segmentation
cases. The performance variation can be greatly seen when qualitative analysis of
critical cases is carried out, as shown in Fig. 6. In this section, it can be clearly
seen that the network with downsampling to 16 × 16 outperformed the other two
networks. In these cases, the other two networks overestimated the complete region,
which is not part of the lung.
Performance Analysis of Convolution Neural Network … 759

(a) (b) (c) (d) (e)

Fig. 5 Lung segmentation results for Over Segmentation cases a Input lung CT slice b CNN (8 ×
8) c CNN (32 × 32) d CNN (16 × 16) e Ground truth

(a) (b) (c) (d) (e)

Fig. 6 Lung segmentation results for complex segmentation cases a Input lung CT slice b CNN
(8 × 8) c CNN (32 × 32) d CNN (16 × 16) e Ground truth

4 Conclusion

In this study the effect of downsampling of images in the CNN network on the
lung segmentation is studied using three CNN architectures with varied depth of
downsampling of images as 32 × 32, 16 × 16 and 8 × 8. The qualitative performance
of these three networks is studied for the segmentation cases of under-segmentation,
760 S. P. Pawar and S. N. Talbar

over-segmentation and complex cases. The qualitative analysis shows that the CNN
network with six inception layers which downsamples images to 16 × 16 gives far
better performance as compared to the other two networks.

References

1. Brown MS, Mcnitt-Gray MF, Mankovich NJ, Goldin JG, Hiller J, Wilson LS, Aberie D (1997)
Method for segmenting chest CT image data using an anatomical model: preliminary results.
IEEE Trans Med Imaging 16:828–839
2. van Rikxoort EM, de Hoop B, Viergever MA, Prokop M, van Ginneken B (2009) Automatic
lung segmentation from thoracic computed tomography scans using a hybrid approach with
error detection. Med Phys 36:2934–2947
3. Ye X, Lin X, Dehmeshki J, Slabaugh G, Beddoe G (2009) Shape-based computer-aided
detection of lung nodules in thoracic ct images. IEEE Trans Biomed Eng 56:1810–1820
4. Brown MS, Goldin JG, McNitt-Gray MF, Greaser LE, Sapra A, Li K-T, Sayre JW, Martin K,
Aberle DR (2000) Knowledge-based segmentation of thoracic computed tomography images
for assessment of split lung function. Med Phys 27:592–598
5. Sun X, Zhang H, Duan H (2006) 3d computerized segmentation of lung volume with computed
tomography. Acad Radiol 13:670–677
6. Pu J, Roos J, Chin AY, Napel S, Rubin GD, Paik DS (2008) Adaptive border marching algorithm:
automatic lung segmentation on chest ct images. Comput Med Imaging Graph 32:452–462
7. Prasad MN, Brown MS, Ahmad S, Abtin F, Allen J, da Costa I, Kim HJ, McNitt-Gray MF,
Goldin JG (2008) Automatic segmentation of lung parenchyma in the presence of diseases
based on curvature of ribs. Acad Radiol 15:1173–1180
8. Shen S, Bui AA, Cong J, Hsu W (2015) An automated lung segmentation approach using
bidirectional chain codes to improve nodule detection accuracy. Comput Biol Med 57:139–149
9. Hosseini-Asl E, Zurada JM, Gimel’farb G, El-Baz A (2015) 3-d lung segmentation by
incremental constrained nonnegative matrix factorization. IEEE Trans Biomed Eng 63:952–963
10. Hosseini-Asl E, Zurada JM, El-Baz A (2014) Lung segmentation based on nonnegative matrix
factorization. In: 2014 IEEE international conference on image processing 415 (ICIP). IEEE,
pp 877–881
11. Isola P, Zhu J-Y, Zhou T, Efros AA (2017) Image-to-image translation with conditional adver-
sarial networks, In 2017 IEEE conference on computer vision and pattern recognition (CVPR).
IEEE, pp 5967–5976
12. Ulyanov D, Vedaldi A, Lempitsky V (2016) Instance normalization: the missing ingredient for
fast stylization, arXiv preprint arXiv:1607.08022
A Secure Data Sharing Platform Using
Blockchain and Fine-Grained Access
Shamsundar Bhimade, Prashant Bhandare, Amey Bhatlavande,
and Bharati Deokar

Abstract In a research community, data sharing is an essential step to gain maximum

knowledge from the prior work. Existing data sharing platforms depend on trusted
third party (TTP). Due to the involvement of TTP, such systems lack trust, trans-
parency, security, and immutability. To overcome these issues, this paper proposed
a blockchain-based secure data sharing platform by leveraging the benefits of inter-
planetary file system (IPFS). Group data sharing in block chain technology and cloud
computing has become a hot topic in recent. With the popularity of cloud computing,
how to achieve secure data sharing in cloud environments is an urgent problem to
be solved. Although encryption techniques have been used to provide data confi-
dentiality and data security in cloud computing, current technique cannot enforce
privacy concerns over encrypted data associated with multiple data owners, which
makes co-owners unable to appropriately control whether data distributor can actu-
ally distribute their data. Data Sharing in Cloud Computing, in which data owner
can share private data with a group of users via the cloud in a secure way, and data
distributor can distribute the data to a new group of users if the attributes satisfy
the access policies in the encrypted data. Further present a multiparty access control
mechanism over the distributed encrypted data, in which the data co-owners can
append new access policies to the encrypted data due to their privacy preferences.

Keywords Block chain · Data sharing · Cloud computing · Data auditing ·

Encryption · Privacy conflict

1 Introduction

However, it is very crucial to know the three W’s for sharing purpose, such as what,
when, and where. These questions need to be very much clear before initiating the
data sharing process. There is still some scope to work on, how the data set owner
should be given incentives or reward. This research provides secure sharing and

S. Bhimade (B) · P. Bhandare · A. Bhatlavande · B. Deokar

SVERI’s College of Engineering (Polytechnic), Pandharpur, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 761
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_76
762 S. Bhimade et al.

selling of data by leveraging the benefits of blockchain [1]. Blockchain; a distributed

ledger is a new trend in the world of information technology. A lot of financial and
non-financial applications have made use of blockchain. The consensus mechanism
is considered as the most fundamental and significant invention as declared by one
of the specialists of Silicon Valley. The current paper currency relies on third party,
which means that there is a threat to security, trust, and privacy. The significant
feature of blockchain is trust, which can be achieved by eliminating third party. A
peer to peer currency; bit coin is introduced by Nakamoto in 2008 [2]. This allows
the payment to be send directly from one party to another. A paper by Nakamoto
was later came up with the implementation protocol of genesis block with 50 coins.
Currently, blockchain is used in every domain, like cloud, internet of things (IoT),
data trading, information security, health care, and many more because of its striking
features [3]. Cloud computing provides on demand service and processing resources
to the Users or devices. It is dynamic computing style in which dynamically scalable
and usually virtualization resources are provided as a service over the internet. Funda-
mental service offered by cloud providers is data storage. Cloud servers managed by
cloud providers which are not fully trusted. Users may store data files on cloud which
may be sensitive and confidential, like business plans. To preserve data privacy, a
basic solution is to encrypt data files, and then upload the encrypted data into the
cloud. One of the most significant difficulties is identity privacy for the wide deploy-
ment of cloud computing. Several security schemes for data sharing untrusted servers
have been proposed. In these approaches, data owners store the encrypted data files
in entrusted storage and distribute the corresponding decryption. Users may not be
willing to join in cloud computing systems without the guarantee of identity privacy,
because their real identities could be easily disclosed to cloud providers and attackers.
Identity privacy may incur the sabotage of privacy. For example, a misbehaved staff
can deceive others in the company by sharing false files without being traceable.
Therefore, traceability, which enables the group manager to track over the real iden-
tity of a user, is also highly desirable. Highly recommended for any member in a
group should be able to fully access stored data and sharing services provided by
the cloud, which could be defined as the multiple-owner manner. More broadly, each
user in the group is able to not only read data, but also modify their part of data
in the entire data file. Finally, groups are normally dynamic in practice. Changes in
membership makes secure data sharing extremely difficult. On the other side, the
various system challenges granted from new users to learn the content of data files
stored before their participation, because it is impossible for new approved users to
contact with anonymous data owners, and obtain the corresponding decryption keys.
An appropriate membership revocation mechanism without updating the secret keys
of the remaining users is also desired to minimize the complexity of key management.
A Secure Data Sharing Platform Using Blockchain … 763

2 Literature Survey

The major underlying problem in research data sharing is the fear of researchers
regarding misuse and misinterpretation of data. This is because data sharing
approaches are still immature in the context of trust, which is slowly going to be
established among research community. To tackle this issue, various solutions are
proposed, for example, protection of identities of every individual and controlled
access to the data rather than making all the data open access. Still these solu-
tions cannot provide trust, immutability to digital data, and traceability regarding
data usage. Cloud servers store the excessive amount of data, which is a centralized
authority. There are various type of risks associated with a central authority, such as
single point failure. To avoid such failure, third parties are involved to provide data
backups. To eliminate third party for developing a trust based model, a blockchain
is introduced to provide trust and transparency. Decentralized storage is a solution,
which allows storage of data independently on multiple nodes of the network in the
form of a distributed ledger. The problem is the storage and processing limitation of
network nodes. For this purpose, interplanetary file system (IPFS) is adapted, which
is a peer to peer architecture [4]. There is no risk of single point failure. It is similar
to web3, but with different features.
On the security of public key protocol Proposes public key encryption protocol,
describes the various techniques to encrypt the public key [1]. First complete group
key management scheme which can supports all these functions yet preserves effi-
ciency. The proposed scheme is based on the new concept of access control poly-
nomial (ACP) that efficiently and effectively support full dynamics, flexible access
control with fine-tuned granularity, and concealment.New scheme is protected from
various attacks from both external and internal malicious parties [2]. Achieving
secure role based control on encrypted data in cloud achieved through RBAC. RBE
scheme allows RBAC policies to be apply for the encrypted data stored in public
clouds. RBE-based hybrid cloud storage architecture provides facility of an orga-
nization to store data securely in a public cloud, while maintaining the sensitive
information related to the organization’s structure in a private cloud [3].
One approach to encrypt documents satisfying different policies with different
keys using a public key cryptosystem such as attribute-based encryption, and proxy
re-encryption is called broadcast group key management (BGKM), and then give a
secure construction of a BGKM scheme called ACVBGKM. Major advantage of the
BGKM scheme is that adding users/revoking users can be performed efficiently by
updating only some public information. BGKM used for an efficient approach for
fine-grained encryption-based access control for documents stored in an untrusted
cloud file storage [4]. MONA proposed a new secure multi-owner data sharing
scheme, for multiple groups in the cloud. They applied the group signature. and
dynamic broadcast encryption techniques, any cloud user can secretly share data
with others. The storage overhead and encryption computation cost of our scheme
are independent with the number of revoked users. Also they analyze the security of
scheme with difficult proofs, and demonstrate the efficiency of scheme in experiments
764 S. Bhimade et al.

[5]. Data distribution in cloud infrastructure provides an effective approach called

Secure-Split-Merge (SSM) is introduced for the security of data. The proposed SSM
scheme was it uses unique mechanism for performing splitting of data using AES 128
bit encryption key [6]. The chunks of encrypted splits are being maintained on various
group servers of different types of cloud zones. The comparative analysis shows that
the proposed system gives effective outcomes as compared to various existing and
traditional security standards [7]. Security achieves against chosen-plaintext attacks
using the k-multi linear Decisional Diffie-Hellman assumption [8]. Fine-grained two-
factor authentication (2FA) access control system for cloud services. Proposed 2FA
access control system, it was an attribute based access control mechanism imple-
mented with the necessity of both a user secret key and a lightweight security device
[9]. Efficient and secure re-encryption scheme has been proposed for data sharing
in unreliable cloud environment. This scheme is built on top of Cipher text-Policy
Attribute-Based Encryption (CPABE), fine-grained access control to share data. That
scheme can achieve user revocation without whole cipher texts re-encryption and
key re-distributions also, re-encryption is not performed until a user requests for that
data, which reduces overheads. Further, it does not need any clock synchronization
[10]. Sun et al. [11] present a privacy-preserving multi-keyword text search (MTS)
scheme with similarity-based ranking to address this problem. To support multi-
keyword search and search result ranking, we propose to build the search index
based on term frequency and the vector space model with cosine similarity measure
to achieve higher search result accuracy. Peters et al. [12] presents a work which
give a diagram of the idea of block-chain innovation and its capacity to disturb the
universe of managing an account through encouraging worldwide cash settlement,
shrewd contracts, mechanized keeping money records and advanced resources. In
such manner, they first give a concise outline of the center parts of this innovation, and
in addition the second-age contract-based improvements. Luu et al. [13] presents a
work which gives another circulated understanding convention for authorization less
block-chains called ELASTICO. ELASTICO is productive in its system messages
and permit complex foes of up to one-fourth of the aggregate computational power.

3 Proposed System

The data owner define an access policy to enforce dissemination conditions i.e.
read–write conditions. Then he encrypts data for a set of receivers, and outsources
the ciphertext to CSP for sharing and and spreading the conditions. The data co-
owners tagged by data owner can add some access policies to the encrypted data
with CSP and generate the renewed ciphertext. The data distributor can access the
data and also generate the re-encryption key to disseminate data owner’s data to
others if he satisfies enough access policies in the ciphertext. The data accessor can
decrypt the initial, renewed and re-encrypted ciphertext with her or his private key.
Figure 1 indicates different part of system &connection between them.
A Secure Data Sharing Platform Using Blockchain … 765

Fig. 1 Proposed system architecture

3.1 Group Member

In the proposed scheme, members are people with the same interests and want to
share data in the cloud using encryption technique. The most worrying problem when
users store data in the cloud server is the confidentiality of the outsourced data. In the
system, users of the same group conduct a key agreement. Subsequently, a common
conference key can be used to encrypt the data that will be uploaded to the cloud
to ensure the confidentiality of the outsourced data. Attackers or the semi-trusted
cloud server cannot learn any content of the outsourced data without the common
conference key. This system uses a technique called group signatures, which allows
users in the same group to anonymously share data in the cloud.

3.2 Group Manager

Group Manager is responsible for generating system parameters, managing group

members (i.e., uploading member’s encrypted data, authorizing group members).
The group manager in our system is a fully trusted third party to both the cloud
and group members. If an external user tries to access files from a different group
more than three times then the manager will remove that particular user from the
applications.
766 S. Bhimade et al.

Fig. 2 Working flow

3.3 Cloud Service Provider (CSP)

Cloud server provider i.e. CSP provides users with seemingly unlimited storage
services. In addition to providing efficient and convenient storage services for users,
the cloud can also provide data sharing services. However, the cloud has the charac-
teristic of honest but curious. In other words, the cloud will not deliberately delete or
modify the uploaded data of users, but it will be curious to understand the contents
of the stored data and the user’s identity. Figure 2 indicates how data flows between
different parts of the system.

4 Algorithms

4.1 AES Algorithm for Encryption

AES (advanced encryption standard).It is symmetric algorithm. It used to convert

plain text into cipher text.The need for coming with this algo is weakness in DES.
The 56 bit key of des is no longer safe against attacks based on exhaustive key
searches and 64-bit block also consider asweak.AES was to be used 128-bit block
A Secure Data Sharing Platform Using Blockchain … 767

Fig. 3 Performance analysis

with128-bit keys. Rijendeal was founder. In this drop we are using it to encrypt the
data owner file. Figure 3 Shows analysis of existing system & proposed system in
terms of performance.

Input:

128_bit /192 bit/256 bit input (0, 1)

Secret key (128_bit) +plain text (128_bit).
Process:
10/12/14-rounds for-128_bit /192 bit/256 bit input
Xor state block (i/p)
Final round:10,12,14
Each round consists: sub byte, shift byte, mix columns, add round key.
Output:
cipher text(128 bit)

Table 1 gives performance analysis between existing system & proposed system.

Table 1 Performance
Existing system Proposed system
analysis
(ms) (ms)
Key generation 500 450
time
Encryption time 1500 1200
Decryption time 10 9
768 S. Bhimade et al.

5 Conclusion

In this paper, a blockchain-based secure data sharing and delivery of digital assets
framework is presented. The main aim of this proposed scenario is to provide data
authenticity and quality of data to customer as well as a stable business platform for
owner This system is design for secure data sharing scheme, for dynamic groups
in an untruth cloud. A new type authentication system, which is highly secure, has
been proposed in this system. User is able to share data with others in the group
without disclose identity privacy to the cloud. It also supports efficient user revocation
and new user joining. User revocation can be done through a public revocation list
without updating the private keys of the remaining users, and new users can directly
decrypt files stored in the cloud before their participation. System also provides the
new double encryption technique for data security. New re-encryption provides tight
authentication.

References

1. Dole D, Yao AC (1983) On the security of public key protocols. IEEE Trans Inf Theory
29(2):198–208
2. Zoo X, Dai Y-S, Bettino E (2008) A practical and flexible key management mechanism for
trusted collaborative computing. In: Proceedings IEEE conference computer common. pp
1211–1219
3. Zhou L, Varadharajan V, Hitchens M (2013) Achieving secure role-based access control on
encrypted data in cloud storage. IEEE Trans Inf Forensics Secur 8(12):1947–1960
4. Nabeel M, Shang N, Bertino E (2013) Privacy preserving policy based content sharing in public
clouds. IEEE Trans Know Data Eng 25(11):2602–2614
5. Liu X, Zhang Y, Wang B, Yang J (2013) Mona: Secure multi owner data sharing for dynamic
groups in the cloud. IEEE Trans Parallel Distrib Syst 24(6):1182–1191
6. Zhu Z, Jiang Z, Jiang R (2013) The attack on Mona: secure multi owner data sharing for
dynamic groups in the cloud. İn: Processing ınternational conference ınformation science cloud
computer. pp 185–189
7. Khan BUI, Olanrewaju RF (2015) SSM: secure-split-merge data distribution in cloud ınfras-
tructure. İn: 2015 IEEE conference on open systems (ICOS). August 24–26, Melaka,
Malaysia
8. Xu J, Wen Q, Li W, Ji Z (2013) Circuit ciphertext-policy attribute-based hybrid encryption with
verifiable delegation in cloud computing. IEEE Trans Parallel Distrib Syst 24(6):1182–1191
9. Liu JK, Au MH, Huang X, Lu R, Li J (2013) Fine-grained two-factor access control for
web-based cloud computing services. IEEE Trans Parallel Distrib Syst 24(6):1182–1191
10. Sultan NH, Barbhuiya FA (2016) A secure re-encryption scheme for data sharing in unreliable
cloud environment 978-1-5090-2616-6/16 IEEE https://doi.org/10.1109/SERVICES.2016.16
11. Pasupuleti S, Ramalingam S, Buyya R (2016) An efficient and secure privacy-preserving
approach for outsourced data of resource constrained mobile devices in cloud computing.
J Netw Comput Appl 64:12–22
12. Sun W, Wang B, Cao N, Li H, Lou W, Hou Y, Li H (2014) Privacy preserving multi-keyword text
search in the cloud supporting similarity based ranking. IEEE T ParallDistr 25(11):3025–3035
13. Peters GW, Panayi E (2016) Understanding modern banking ledgers through blockchain tech-
nologies: future of transaction processing and smart contracts on the internet of money. İn:
Banking beyond banks and money. Springer, New York, NY, USA pp 239–278
Efficient and Interactive Fuzzy Type
Ahead Search in XML Data

Laxman Dethe, Geeta Khare, Avdhut Bhise, and Somnath Zambare

Abstract This paper we are written for XML document to store XML data in XML
formats for Security purpose. Here we compute the problem of efficiently creating
ranked result for keyword search query in XML document. In old methods there are
Xlink, Xpath and Xquery are query methods available to search data in XML file
of XML DB. Here the method new users are not able to understand syntax of query
when accessing the query. In this steps first write query, put forward to the system
and retrieve relevant results. In case of keyword search there is the fuzzy type ahead
search over XML data that user write a keyword search on fly way and access a new
information pattern, This method are optional to old methods. The users didn’t need
to know knowledge of XML query languages and its syntax. We are also adding a user
study confirming that keyword-based search in SQL for a range of DB retrieval task.
The query time, the text index carry keyword-based searches with giving interactive
answer. Successful keyword search is valuable for top-k in XML document, these
are user simply manage, semantic and steer into documents.

Keyword Fuzzy search · Keyword search · MCT · LCA · ELCA

1 Introduction

In old keyword based search system in XML data, A user write a keyword query
and submit it to system, Retrieves information. In fact particular person know about
language that what is Xpath and Xquery languages, What are the syntax, notation
etc., of them Because without syntax, no one able to retrieve data. We are study of
effective search in XML data. The system is searching XML data on the users type
in doubt keyword. It is allows to user discover data as they have type, If even in
available minor errors in their keyword. We are proposing effective index structure
and top-k algorithms to achieve a more interactive rate. We observe effective ranking

L. Dethe (B) · G. Khare · A. Bhise · S. Zambare

SVERI’s College of Engineering (Poly.), Pandharpur, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 769
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_77
770 L. Dethe et al.

functions and early massacre techniques to progressively know the top-k relevant
results [1, 2].
The today’s day most of the transactions on the internet XML are used to storing
and retrieving reason. Lots of leading products developed companies are use XML
metadata structure. This paper progress with a goal to manage XML information. It
is helps in storing and retrieving relevant answers. In this case consumer has limited
knowledge of the data, frequently the users be aware of left in dark when issuing
the query, and has to use a try and see draw near for publishing, finding, managing,
retrieving data from DB in XML formats and updating storing data in Document of
XML. There is special modules of this paper.
The one modules is a SQL manager, which help to retrieve and manage data from
XML data in XML DB and we implement keyword search in XML data in XML
database. The user security and management are another modules. DB servers is
Client–Server based DB. It is the more easy to retrieve, user-friendly and easy to
access the DB for both programmer and the client of understand. Actually It is used
to create database, report, query and the table [3].

2 Literature Survey

The method used frequently is Auto complete which is predicts expression that the
user may have type in standard on the unfinished string the user has been type. Here
the trouble with Auto complete. The system delicacy a query as a single string, If it
consist several keywords. There is one answer to this problem set by Bast and Weber
is that Complete Search in textual documents. It can find approximate answers by
allowing keywords search in query, that come out at any seats in the answer. Fuzzy
search can gives user instant replay as users type in keyword. It doesn’t need users to
type in complete keywords. The Fuzzy search can assist users browse the information
that users save typing effort and efficiently search the results.
We also considered fuzzy search in relational DB. XML data in a Fuzzy search
mode and it is not negligible to swell existing method to support fuzzy search in
XML data, Because XML has parent–child bonds. We need to recognize absolute
XML sub trees that confine such structural bonds from XML data to result the queries
with keyword. TASX find the XML data on the fly as user’s type in query keywords
still in the occurrence of the errors of their input keywords. Every query with many
keywords needs to be answered strongly. The leading challenge is search-efficiency.
This short running-time requirement is mainly difficult when the backend repos-
itory has a large amount of data. We suggest successful index structures and algo-
rithms to answer keyword queries in XML data. Efficient ranking method and timely
extinction techniques to gradually find out top-k answers.
XML stands for Extensible Markup language. The word “Extensible” imply that
a developer can expand his ability to describe a document, and describe meaningful
tags for his purpose XML is used to generate vibrant content. Databases are study of
SQL-SERVER, ORACLE, My SQL, XML are done in the portion of manipulating the
Efficient and Interactive Fuzzy Type Ahead Search in XML Data 771

stored facts by their respective query language. XML database assists professionals
and the corporate to trace and maintain the data into the database. For using the
over specified database corporate has to pay valued amount as per the company rules
and regulations for receiving the registration from the authorized database compa-
nies. Installation cost, maintenance cost and the accomplishment cost can affect the
company’s production price. The XML database is a platform self-governing server
database and can be used at no cost provided by the Sun Microsystems [4].
In XML nearby are two types Xpath and Xquery. Xpath is declarative language
for XML that give a simple syntax for addressing a fraction of on Xml file. Xpath set
of element can be retrieved by specifying a index like path with zero or more state
place on the lane. Xpath delight an a XML document as a rational tree with nodes
for each part, attribute text, processing instruction, comment, namespace and root
[1, 5]. The essential of the addressing method is the context node (begin node) and
location path which show a path from one point in an XML document to a further.
Xpointer can be used state on complete location or relative location. Location of path
is composed of a chain of steps joined with “/” each travel down the previous step.
Xquery is fit in feature from query language for relational scheme (SQL) and Object
oriented scheme (OQL) [4].
Xquery holds up operations on document sort and can negative, take out and
reform documents. W3c query running group has planned a query language for
XML called Xquery. Values always state a series node can be a document, element,
attribute, text, namespace. peak level path state are ordered according to their location
in the unique hierarchy, top-down, left–right array [6]. The significant parts are Data-
Centric document and Document-Centric file. Data-centric file Xpath are compound
for known. It can invent both in the database and outer surface the database. These
files are used for communicate data among companies. These are mainly processing
by engine; they have quite regular arrangement, fine-gained data and no combine
content.
Document-Centric are document typically designed for human utilization, they
are usually composed openly in XML or some other plan (RTF, PDF, SGML) which is
then transformed to XML. Document-Centric require not have regular arrangement,
bigger gained data and plenty of mixed substance [3, 7]. In this paper to study
of previous technology that they are working LCA (lowest common ancestor) [8],
ELCA (Exclusive Lowest common Ancestor) [8], MCT (minimum cost tree) [6, 9]
and begin new technology Top-k algorithm [1, 5, 10] recognize fairly accurate best
ranking solution in system in XML document more effectively and professionally.

3 XML Query Techniques Based Fuzzy Methods

File server is a client–server based database. It is additional user-friendly, easy to

regain and easy to right to use the database for both the programmer and the customer
or end user. It is used to make database, table, query and the reports. User can sight
the database, make table and examine the query and a behind all he can make report
772 L. Dethe et al.

on the foundation of tables and with esteem to their queries. For accessing, creating
and maintaining the DB. Users should have the agreement from the server. Server
allowance the permission and later that client (user) can perform what he desires to
perform. Client can analysis only the encrypted from of data, Because of the entire
data are preserved in the XML DB in decrypted from what a client cannot able
perceive that. For the safety point of view. It has exacting user with their passwords,
who are the approved persons who can access the DB.
This is the query analyzer DB to which several users can access the DB at the
same time with no limits. It is a platform independent DB and more cost-effective
than any other DB. We recommend the index to improve search routine. We can use
“arbitrary access” depends on the index to do an early killing in the algorithm. That
is, set an XML element and a key keyword. We can got the relative score of the
keyword and the element uses the index, Except retrieving inverted lists. Fagin et al
algorithm have proven that the threshold-based algorithm using arbitrary access is
best overall algorithm that properly find the top-k results.
It is very costly to build the union lists of each input keywords as there may be
several predicted words and many reversed lists as an alternative, we can produce
a partial virtual list on the fly of user. We only utilize the element in the partial
effective list to calculate the top-k results. The partial virtual record can stay away
from accessing all the element of inverted lists of forecasted words. It is only needs
to retrieve those with large scores, and if we have calculated the top-k results using
the partial retrieved elements, we can do an early killing and do not require to visit
other element on the inverted element lists.
This system huge number of safety options give to data and users. Administrator
has the majority of job to form user and allows agreement to maintain DB. When
the users are entering username and private key into the system to login for use, He
will do work on the particular data to store data into document format, access data
from document file present in DB, all together he departs to temp directory where
document is stored only on his data observe into temp folder not extra person data,
Because when users log out them file delete from folder preserve DB safety of every
users. One more is data traveling form one user account to document DB or other
person, Data is encrypted and preserve consistency into particular network.

3.1 Database Design

The XML server consists of different modules. The GUI Clients build the connection
with server with help own username and password. The SQL manager establishing
the connection with respective DB, parse the input query in proper syntax to display
answers Grid or in File format. The entire syntax’s for DDL and DML queries are
amassed in “syntax.xml” file.
The Example syntax: Data storage space with the popularity of XML of the server
require to effort with and lay up XML data. The Example XML file format is:
Efficient and Interactive Fuzzy Type Ahead Search in XML Data 773

Fig. 1 System model

<content>
<author>Laxman</author>
<title>The fuzzy search</title>
</content>
Association builder of xml server establish relation among two XML documents,
applying the concept of primary key. The user executive helps for managing users
by editing users, Creating users, Deleting users, Editing/Assigning their password.
The Server will give its own verification features to validate users using facility
of XML Encryption/Decryption, unauthorized persons or intruders cannot retrieve
important document. In this segment three significant XML query and keyword look
for methodologies are explained. The major problem connected to Xpath and Xquery
are their difficulty concerned in the syntax for query. The compared to Xpath and
Xquery, LCA-based interactive search [11] and MCT [6] are superior and excellent.
below part give detailed information on the above discussed methods as shown in
Fig. 1.

3.1.1 Minimum Cost Tree

To find applicable results, to a keyword query in an XML data. To each node, we

need to describe its relative results to the query as its all sub tree with paths to nodes
that contain the input query keywords. This sub tree called the “minimal cost tree
(MCT)” for that nodes. Special nodes related to different results to the input query
and we have to study how to compute the relevance of every results to the query
for the ranking. The Given an XML documents D, a node N in document D, and a
774 L. Dethe et al.

keyword query Q = {k1, k2, k3, k4,…, kl}, a MCT of query Q and node N is the sub
tree available at N, and for every keyword ki e Q, If node N is a qussi-content node
of the ki , that sub tree contain the pivotal path for ki and node particular node N.
We have to first understand the predicated words for every input keyword. After,
we will design the MCT for each node in the XML document tree depends on the
predicated words and revisit the good ones with the larger score. The important
advantage of this is even if a node doesn’t have descendent nodes that contain all the
keyword in the query input, This node could immobile be considered as a potential
result [12, 13].

3.1.2 LCA Based Interactive Search Method

We suggest a lowest common ancestor (LCA) based interactive search methods. We

will use here the semantics of exclusive LCA to find relevant results for predicated
terms. We use here trie based index the make tokenized words in XML document.
The First for a single keyword word, identify related tree node. After we set the leaf
descendants of that node and access. The related predicated word and the predicated
XML element on that inverted lists. To a query string transfer into keyword k1, k2,
k3, k4……, kl. To every keywords ki (1 < i < l), there is several predicated words
[4, 7].

3.1.3 ELCA Method

To Reduce the limitation of LCA based method exclusive LCA (ELCA) [12, 13] is
designed. It tells that an LCA is ELCA if it is immobile an LCA after without its
LCA descendants. Finding the separate LCA of every contain node is called ELCA.
XU and papakonstantion [13] designed a binary-search based method to competently
find ELCAs.

3.2 Efficient and Effective TOP-K Algorithm for XML Data

Search

This paper we are first checking it that how top-k search based algorithm are benefi-
cial. However ranking the results of keyword it needed to LCA and MCT with them
picky score [11, 14]. The parameterized top-k algorithm separated in two different
stages. The first one is a structure of algorithm that on a problem occurrence construct
a structure of possible size and the another stage is an enumerating the algorithm
that gives the k best answers to the instance depends on the structure. We are devel-
oped new methods that bear efficient enumerating algorithms. We design the relation
among fixed-parameter tractability and parameterized top-k algorithm [1, 4].
Efficient and Interactive Fuzzy Type Ahead Search in XML Data 775

3.2.1 Ranking Query Results

Now we converse how to design the MCT for a node N as result to the query.
Naturally, we first estimate the significance among node N and every input keywords
and then join this relevance score as total score of MCT. We need to target on different
methods to measure the relevance of node N to a query keywords and join relevance
scores [3, 7, 15].
A. Ranking the Sub Tree
There are have only two different ranking function to calculate the rank and score
between node N and keyword ki .
Case 1: N has keyword ki .
The score of that node N and keyword ki is computed by

SCORE1(N, ki ) = ln1 + tf ki ,n*lnf() (idf (ki ))1 − s*s*ntl(n) (1)

where,
1. tf (ki , N)—Number of incidence of ki in sub tree rooted N
2. idf (ki )—Ratio among no. of nodes in XML to no. of nodes that have keyword
ki
3. itl(n)—Inverse term length of |N/Nmax| = node with max terms s-set to 0.2
Case 2: Node N doesn’t have keyword ki, but its descendent got ki . The ranking
depends on ancestor and descendent relation. The another ranking function to find
the score between N and kj is:

SCORE (N,kj ) = p ∈ Pn, p*SCOREp, kj (2)

where

p set for pivotal nodes.

α set to 0.8
n,p Distance between n and p.

B. Ranking Fuzzy Search Result

Assume a keyword query Q = {k1, k1, k2, k3,…., kl} in words of fuzzy search, The
MCT may not have predicated list of words for each keywords, but having predicted
words for each keywords. Let predicated words are {w1, w2, w3, w4,…., wl} the best
comparable prefix of wi would be measured to be most similar to ki . The function to
count the similarity between ki and wi Where ed—edit distance ai —is prefix, wi —is
predicted word, y—is constant.

Sim(ki , wi ) = y*11 + ed(ki , ai ) + 1 − y*|ai /wi | (3)

776 L. Dethe et al.

where value of γ is between 0 and 1, As the former is more necessary, γ is close to

1. The experiment proposed that a best value for γ is 0.95. We highly structured the
ranking function by integrateing this similarity function to bear fuzzy search is:

SCORE(n, Q) = i = 1simki , wi *SCORE(n, wi ) (4)

4 Conclusion

We evaluate the effectiveness of the calculating the prefixes on the trees. This are like
to a query keywords. The proposed efficient incremental algorithm to react to single-
keyword queries that are declared as prefix states. considered special algorithms for
computing the results to a query with several keywords. Well-planned algorithms are
developed for incrementally calculating results to queries by using cached answers
of previous queries in sort to get a high interactive rate on large data sets. The
LCA-based methods to interactively find out the predicted results and the designed
a minimum cost trees based search methods to capably know the nearly all related
results. We devise a fuzzy search to additional improve search routine.

References

1. Feng J, Li G (May 2012) Efficient fuzzy type-ahead search in XML data. Proc IEEE Trans
Knowl Data Eng 24(5)
2. Sivapuja S, Mohiddin SK, Srikanth Babu S, Srikar Babu SV (March–April 2013) Efficient
searching on data using forward search. J Emerg Trends Technol Comput Sci 2(2)
3. Bast H, Weber I (2006) Type less, find more: fast autocompletion search with a succinct index.
In: Processing Asnn. int’l ACM SIGIR conference research and development in information
retrieval (SIGIR) pp 364–371
4. Andrew Eisenberg IBM (2002) Advancement in SQL/XML jim meton oracle corp
5. Willimson H (2009) The complete reference of XML. The McGrew-Hill Companies Inc., New
York
6. Ding B, Yu JX, Wang S, Qin L, Zhang X, Lin X Finding top-k min-cost—connected tree in
database”.The Chinese university of Hong Kong China
7. Li G, Feng JH, Zhou L, Interactive search in XML data department of computer science and
technology, Tshinghua national laboratory for information science and technology, Tsinghua
university. Beijing, China, 100084
8. Teradat YX (2003) Yannis papakonstantion university of California. Efficient LCA based
keyword search in XML data. ACM copyright
9. Bourret R (2005) XML and database, independent consultant, felton, A 18 woodwardia Ave.
Felton CA, USA Spring, 95018
10. Li D, Li C, Feng J, Zhou L SAIL: structure-aware indexing for effective and progressive
top-k keyword search over XML document. Department of computer science, University of
California, Irvine, CA USA pp 92697–3435
11. Liu Z, Chen Y ( 2007) Identifying meaningful return information for Xml keyword search.
Proc ACM SIGMOD In Conf Manage Data 329–340
Efficient and Interactive Fuzzy Type Ahead Search in XML Data 777

12. Bast H, Weber I (2007) The complete search engine: interactive, efficient, and towards Ir&db
integration. In: Biennial conference innovative data systems research (CIDR) pp 88–95
13. Ji S, Li G, Li C, Feng J (2009) Efficient interactive fuzzy keyword search. In: Processing Int’l
conference world wide web
14. Chen L, Kanj LA, Meng J, Xia G, Zhange F (2012) Parameterized top-k algorithm,
communicated by D-Z DU
15. Harel D, Tarjan RE (1984) Fast algorithms for finding nearest common ancestors. J SIAM
Comput 13(2):338–355
IOT Based Interactive Motion Detection
Security System Using Raspberry Pi

Geeta Khare, Subhash Pingale, Avdhut Bhise, and Sharad Kawale

Abstract The Internet of Things (IoT) is the network of physical devices with elec-
tronics, software, sensors and other objects which consists of an embedded system
that enables to collect and exchange data. It is mainly intended for transferring
user data in real time security monitoring system, e.g. to monitor and control traffic
and road condition. Unquestionably, everyday day-life and behavior of potential
subscribers have adhered to IOT. From this perspective domestic and work areas will
have effects of IoT. It is anticipated that there will about 50 billion internet-enabled
devices by 2020. The aim of this paper is to introduce security alarm system for
detecting motion and get an image or notification using low processing power when
motion is detected. The user is alerted by sending snapshots through mail or notifi-
cation via text message. In case of unavailability of the network service, Raspberry
Pi will store the data locally and send that data when the internet is available. Rasp-
berry Pi is a low-cost device as compared to other available present systems. It is a
small sized computer used to process a captured image or data as and when motion
is detected. Passive Infrared (PIR) sensors are used to detect the motion; the image
is captured through the camera and provisionally stored in the raspberry pi module.
This system is suitable for small personal area i.e. personal office cabin, bank locker
room, parking entrance.

Keywords Internet of Things (IoT) · Motion detection · PIR sensor · Raspberry Pi

G. Khare (B) · S. Pingale

Computer Science and Engineering, SKN Sinhgad College of Engineering, Solapur University,
Pandharpur, Maharashtra, India
e-mail: [email protected]
A. Bhise · S. Kawale
Sveri’s College of Engineering (Poly.), Pandharpur, Maharashtra, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 779
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_78
780 G. Khare et al.

1 Introduction

A British technology pioneer, Kevin Ashton, through his work of Auto-ID Center
at the Massachusetts Institute of Technology (MIT), introduced a term Internet of
Things in 2002 [1]. Nowadays the Internet of Things is used in numerous research
areas like in the fields of home automation, transportation, agriculture, automobiles,
and healthcare etc. Also, the IoT will change everyday life and help in situations like
managing airports’ passenger flows, smart homes, heating buildings, caring for the
elderly, military, defense. Home automation brings up the application of computer
and information technology for controlling home appliances and domestic features
[2]. This results from convenience, energy efficiency and safety benefits along with
improved quality of life. The vision of IoT is “from anytime, anywhere, connectivity
for anything” [3]. In the existing system, CCTV cameras are used for surveillance but
it incurs the high cost due to more hardware and storage for continuous recording of
activities. Also, human intervention is required to detect any unauthorized activity.
The IoT enables physical objects to see, hear, think and perform jobs to share
information and to coordinate decisions. The type of communication that we
experience today is either human–human or human–machine but the IoT assures
machine-machine communication. The overall idea of the IoT is illustrated in
Fig. 1, in which every internet enabled device communicates with other indepen-
dent services, wherein each domain sensors interact directly with each other.The
basic idea behind IoT is to permit an autonomous exchange of useful information
using internet-enabled devices.
The IoT uses low-cost computing devices shown in Fig. 1.where there are less
energy consumption and limited impact on the environment. Through IoT, billions of

Fig. 1 Internet of things

IOT Based Interactive Motion Detection Security System … 781

devices will be connected to each other including cars, phones, jet planes, appliances,
wearable gear etc.
There are many challenges to the use of IoT. Since every device needs IPv4 to
address for communication. Currently, the IPv4 has only 4.3 billion unique addresses,
which will be exhausted soon and we need to adapt to IPv6 [4]. Therefore one of
the challenges to be met is limited availability of IPv4 address. Another challenge in
IoT would be data storage as billions of devices are collecting and transferring the
data and that needs to be stored for which massive storage space is required. While
dealing with IoT, more and more authenticate information is being collected from
the devices. Hence, we have to make sure that the security policies are in place to
protect the data from the hands of hackers [5]. Now a day the need of people is to
monitor and control their business activities or house while moving across the globe.
To achieve an economical and safe way of monitoring and the controlling Internet
of things can be utilized effectively. The focus of IoT is to allow things to become
smarter, more reliable and more autonomous. Sensors have to play a key role in the
field of security systems. The use of sensors depends upon the type of application
like ultrasonic sensor, photoelectric sensor, temperature sensor, and passive infrared
(PIR) sensor.

2 Proposed Algorithm

The existing system has many disadvantages. It is difficult to implement and very
complex. It requires more human effort and hardware. So, the entire system is expen-
sive. These can be overcome by the use of interactive motion detection security system
using Raspberry Pi in IOT.
This system consists of three layers—the first layer is for motion detection. In
the second layer actions are processed according to Python script and in third layer
recorded snapshot or video is sent to the user.
Raspberry Pi (shown in Fig. 2) is a credit card sized computer and also known as
Model B + have the following features:
• 700 MHz ARM CPU
• 512 MB SDRAM
• Ethernet RJ45
• 2 × USB 2.0
• HDMI and Composite RCA.
Raspberry Pi works same as that of a standard PC. It needs a keyboard for entering
the command, a monitor for the display unit, an operating system, and a power supply.
Internet connectivity is provided via Ethernet or Wi-Fi. The Raspberry Pi board
consists of an inbuilt processor and a graphics chip, main memory (RAM), different
interfaces and connectors. It runs on a LINUX based operating system Raspbian
OS [6]. Raspbian has a desktop environment similar to windows and Mac called
Lightweight X11 Desktop Environment (LXDE), so it provides an easy transition
782 G. Khare et al.

Fig. 2 Raspberry Pi board

for those not familiar with LINUX command line. External storage device can’t be
used to boot the Raspberry Pi. Some of the pin connections are essential, others are
optional but all Raspberry Pi models have the same CPU named BCM2835 which
is cheap, powerful, and it does not consume a lot of power [7]. Raspberry Pi has a
wide range of usage.
In this security system, PIR sensor (Fig. 3) is connected to the MCP 3208 ADC
via Serial Peripheral Interface (SPI) protocol and detecting a range of PIR sensor is
6 m. It covers the 180° angle.
Python is chosen as the main programming language as it is suitable for real-world
applications. Also, it is fully fledged and easy to learn [8]. Python script compares the
last frame and the current frame of the live video. If there is any mismatch then the
motion flag is set and further events are processed. Upon execution of python script,
snapshots are captured from the camera until the motion is detected. For storing the

Fig. 3 PIR sensor

IOT Based Interactive Motion Detection Security System … 783

Fig. 4 GSM module

statement temporarily an external SD card is used. As there could be a chance that the
person would attempt to damage the device, captured images uploaded immediately
to the external server. Figure 4 shows the GSM Module.
GSM Module Needs.
1. Raspberry Pi 3
2. GSM/GPRS TTL UART Modem-SIM800L
3. A 5 V battery or power supply
4. Connecting wires
5. A mobile phone
6. An extra SIM card.
Initially, the user has to insert a SIM card into GSM modem with suitable connec-
tion with Raspberry Pi. The SIM used in this module should have enough balance
to send the text message and care should be taken that it must be placed in such a
way that it having an appropriate range of a particular network. Give power supply
to GSM module and wait for few seconds for initialization of SIM.
Figure 5 shows a block diagram of the given proposed system. A Passive Infrared
(PIR) sensor is connected to the general I/O pins of the raspberry pi [9].
When the motion is detected, the camera module records the snapshot or video for
the assigned time or until motion is detected. An external SD card is used for storing
the snapshot temporarily in the raspberry pi. An internet connection is required for
sending email and text message notifications. If the server is not available then the
snapshots are stored in the raspberry pi [10].
784 G. Khare et al.

Fig. 5 Block diagram

3 Experiment and Result

When the experimental set up is initiated and motion is detected through the PIR
sensor, the web camera has captured a series of images and these snapshots are sent to
the Gmail account of the registered user. Also, the notification through text message
has been given to the same user on mobile. Hence, it can be inferred that the security
system is successfully implemented with more reliability and efficiency.

4 Conclusion

Raspberry Pi provides a smart, efficient and economic platform for implementing the
security system and home automation. This security system mainly focuses on motion
detection and it enables the user to monitor their homes and buildings even from
remote locations. This system could be an alternative for expensive security systems
being used in the present days. This system does not need any special modifications
in the infrastructure at the place where it is installed. It can be implemented without
much difficulty.

References

1. Schoenberger CR (March 18, 2002) The internet of things. Forbes magazine

2. Jain S et al (Feb 6–8, 2014) IEEE, Raspberry Pi based interactive home automation system
through E-mail, In: International conference on reliability, optimization and information
technology—ICROIT 2014, India
3. ITU Internet Reports (2005) The internet of things
IOT Based Interactive Motion Detection Security System … 785

4. Atzori L, Iera A, Morabito G (2010) The internet of things: a survey. Comput Netw
54(15):2787–2805
5. Nilssen A (2009) Security and privacy standardization in the internet of things, in eMatch’09—
future internet workshop, Oslo, Norway
6. Patil N, Anand K Ward: IOT and raspberry PI based environmental monitoring application
7. Pawar AA, Rangole JS (Feb 2, 2016) Review paper on raspberry Pi and wi-fi based home
automation server 3
8. Ajeeth, Sandhyaraani M (June 3, 2016) Security system for industries using raspberry Pi and
IOT. 2
9. Soundhar Ganesh S et al (2015) Raspberry Pi-based interactive home automation system
through internet of things. 3(3)
10. Anitha G et al (2015) An internet of things approach for motion detection and controlling home
appliances without cloud server. 3(8)
Commercially Successful Rural
and Agricultural Technologies
Cognitive Intelligence of Internet
of Things in Precision Agriculture

Rahul Keru Patil and Suhas Shivlal Patil

Abstract In the recent scenario, the technology of the Internet of Things (IoT)
acting a vital part in Precision agriculture, Military, Engineering applications. The
main resource of our country is the agriculture field. IoT is widely adopted in the
Precision Agriculture field to count the dissimilar environmental constraints such as
soil moisture, humidity, temperature, PH value of soil, etc. for increasing the yield
of the crop. While using the IoT in Precision Agriculture it aided to decrease the
consumption of the natural assets (freshwater, clean air, healthy soils, etc.) used
in farming. Therefore, the purpose of this work is to implement the several IoT
technologies accepted for Precision agriculture. This work has also points to the
different communication technologies and wireless sensors available for Precision
Agriculture. The proposed system will very helpful to our farmers because these
technologies applied for utilize the limited resources to increase the yield of crop.
These technologies applied cognitively at exact location and at exact time so quantity
and quality of crop will definitely increase.

Keywords Precision agriculture · Internet of things · Sensors · Wireless sensor

network · Cloud technologies · Irrigation management system

1 Introduction

The important field of any nation is farming. The cardinal source of economy is
agriculture. As concerned to the Indian economy, agriculture is the main sector. The
Indian agriculture sector employs almost 50% of the countries. India is acknowledged
to the world’s major creator of pulses, rice, wheat, and spice yields that they produce,
which trusts on plant growth and the yield they acquire.

R. K. Patil (B)
Department of Technology, Shivaji University, Kolhapur, India
S. S. Patil
College of Engineering, Karmveer Bhaurao Patil, Satara, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 789
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_79
790 R. K. Patil and S. S. Patil

Plant growth and yield depend on three parameters one is monitoring the land,
second is continuous alert for identify the diseases and third is drip irrigation manage-
ment. The Indian farmers manually worked for these three parameters. Due to this
human errors will occur and yield will be reduced.
Cognitive intelligence of Internet of Things (IoT) in agriculture applications acting
a vital role to meet the above parameters for increasing yield through resulting of
improving the efficiency by effective use of inputs like soil, fertilizers, and pesticides,
monitoring the livestock, predict the pest and diseases, scanning storage capacities
like water tank level, monitor the crops are fed and watered well.
In the current scenario as concern to our Indian farmers whose agriculture land has
limited and financially weaker, the agriculture field wants such tools and expertise that
increases the effectiveness of quality of production and decreases the environmental
effect on the yield.
The Internet of things along with the Wireless Sensor Network (WSN) in agricul-
ture may place forward the central solution role to precision agriculture. Precision
agriculture is the technique of applying the correct amount of input (water, fertilizer,
pesticides, etc.) at the accurate location and at the right time to improve produc-
tion and increase quality while protecting the environment [1]. WSN is defined as a
gathering of nodes structured into a supportive network.
Internet of Things (IoT) stand-in a important role in Precision agriculture. IoT
is an internetworking of physical kits, automobiles, constructions, and other items
embedded with electronics, software, sensors, actuators, and network connectivity
that assists these devices to gather and exchange the information.

2 Literature Review

Agriculture techniques. We used three research database as follows for finding the
existing related work in the domain of precision agriculture using WSNs.
• IEEE (www.ieee.org)
• Springer (www.springerlink.com)
• Science Direct (www.sciencedirect.com).
Yuthika Shekhar et al. [2] proposes irrigation management system, the moisture
and temperature data are took using KNN (K Nearest Neighbour) cataloging machine
learning algorithm for getting the sensor data for irrigating the soil with water.
P Rajalakshmi and S. Devi Mahalakshmi introduced the Sensors are the essential
components for precise agricultural applications. Depending on the content of the
moisture, obtained from soil moisture sensor, the motor automatically switches on.
If the moisture content is very low in soil then automatically sprinkles on to the roots
of plants [3].
Karan Kansara et al. [4] presents the irrigations practiced during winter, was far
overdid the irrigation water want. While in the summer, the irrigation accepted have
Cognitive Intelligence of Internet of Things in Precision … 791

covered 44% of irrigation water requests. The revision reveals a high variety in the
practice of irrigation between different principles and different farms.
Avatade and Dhanure [5]. This paper defines the system construction on hardware
and software design. Jia Uddin et al. [6] introduce the uses ARMs and RF modules
and the system will be associated using these modules. The main essential aspect
of this system is the radio frequency module which is used to send and receive the
data to the controller. They are mentioned that this system uses three nodes that
communicate with each other and irrigate the field involuntarily without man power.
The goal of this project is to modernizing agriculture technology.
Bai and Liang [7] presents finest consumptin of water is the major idea of this
irrigation system to lessen water utilization. This project uses soil moisture sensors
to find the amount of water in their cultivation and also uses sensor to notice water
level in the tank.

3 Methodology

3.1 Components Used

• Adruino Uno
• Soil moisture sensor
• ESP8266 wifi module
• Motor driver PCB for Arduino.

Adruino Uno: The Arduino Uno is a prototype platform (Open Source) based on
easy to use hardware and software. Arduino Uno is microcontroller board based on
the ATmega328. It covers everything required to backing the microcontroller.
Soil moisture sensor: It is used to count the moisture level in the soil. It can
calculate the ideal soil moisture content for different species of plants. We can place
the points of the sensor into soil and check the moisture level, but the position of the
prongs should be right.
ESP8266 wifi module: It is a self- contained SOC with integrated TCP/IP protocol
stack that can provide any microcontroller admittance to your WiFi network. The
ESP8266 is skilful of either hosting an application or offloading all Wi-Fi networking
purposes from alternative application processor.
Motor driver PCB for Arduino: The shield and its consistent Arduino library
create it easy to regulator two bidirectional, high power, brushed DC motors with an
Arduino or Arduino clone.
792 R. K. Patil and S. S. Patil

Fig. 1 Block diagram of precision agriculture using internet of things

3.2 System Requirements

(a) Database Requirements MySQL

(b) Software Requirements
(c) Arduino UNO IDE
(d) Blynk Server/Thingspeak Server
(e) Dropbox
(f) Microsoft
(g) MySQL
(h) Soil moisture sensor
(i) Motor driver
(j) ESP8266 wifi module.
A suggested system is a mixture approach to addressing the problem of data drop.
This can be observed as an application of case selection and feature selection in
bug sources [8]. We construct a binary classifier to expect the order of applying
instance selection and feature choice [9]. Sorting is a data mining technique that
allots categories to a gathering data in order to supporter in more accurate prediction
and scrutiny. Classifier is a administered function where the learned attribute is
categorical or nominal. It is executed after the learning process to label new records
Cognitive Intelligence of Internet of Things in Precision … 793

or data by giving them the best prediction. In my project, we can use naïve Bayes
classifier.
The suggested system can be reflected of having different levels, to gather sensed
data, and actuate the end devices based on sensed data. The end devices are interfaced
with the Arduino.
A. DHT 11: It has both temperature and humidity sensor. DHT 11 has the best
structures are its simple connection, better quality, less costly, and interference
capability and precise calibration. It can shield the transmission area up to 25 m.
B. Soil Moisture Sensor: It sense the amount of water content in the soil can be
counted.
C. Light Dependent Resistor (LDR): It sense the light inside the greenhouse area.
If the light intensity changes then LDR responds with variable resistance. The
value of resistance of LDR reduces as the light intensity growths.
D. Inlet and Exhaust Fans: Inlet and Exhaust fans are used to push the heat outside
from the greenhouse and pull fresh air from outside into the greenhouse. This
maintains the suitable airflow essential for the in greenhouse crops.
E. Water Pump: The water pump reacts when the soil moisture level below or
above the specified threshold. The water pump starts and switches on or off
respectively according to conditions.
F. ThingSpeak: It is an Internet of Thing analytics platform service that deals you
to collective, visualize, and analyze live data streams in the cloud. You can also
drive data to ThingSpeak from our Arduino, generate instant visualizations of
live data. It is possible without setting up servers or developing web software,
ThingSpeak allows engineers and scientists to make a prototype and form IoT
systems.
G. Dropbox Capture a web page screenshot on web page Save Record.
H. For identification of disease, image processing is used. Following algorithm is
used for the same.
• The picture of camera is taken.
• The median filter is performed by taking the size of majority of the vectors
inside a veil and arranging the extents.
• The photo is given as a input for the system.
• The framework will utilize the picture handling strategies.
• After enhancement, division is finished then compare with thresholding and
take appropriate decision for increase yield of crops through IoT.

4 Results and Discussion

The projected system works fine and carried effective outcomes. The finale devices
like light source, water pump, and fans inside the greenhouse activated according
to the threshold condition of parameters like temperature, humidity, soil moisture
values. The evidence cached from the Arduino pass to the Thingspeak server and
794 R. K. Patil and S. S. Patil

Fig. 2 Display moisture on

LCD

data pictured in the server. The values collected by the sensor DHT11 temperature
and humidity and the soil moisture sensor. In field, the humidity values are different.
The discrepancy is in the form of increasing and decreasing and stable sometimes. So
whenever the humidity value less than or equal to the threshold value the fan switched
on automatically. In Field, the temperature values are different. The discrepancy
is in the form of increasing and decreasing and stable sometimes. So whenever
the humidity value greater than or equal to the threshold value the fan activated
on automatically In field, the soil moisture values are varied. The discrepancy is
in the form of increasing and decreasing and stable sometimes. So whenever the
humidity value larger than or equal to the threshold value the water pump switched
on automatically. In the same way, field visualized in the Thingspeak. The light will
be on and off according to the LDR sensor module threshold value condition.

5 Conclusion

This work is the applications of IoT in Precision Agriculture applications. It decreases

the farmer contribution. The threshold standards have took in the existence of farmer
because the conditions are changed. By using this system, we can sense the param-
eters regarding to the ventilation, lighting and irrigation. The sensed information
directed to the cloud platform i.e. ThingSpeak server or Blynk server. Current data
saved on webpage and save information as photo record in Dropbox.
Cognitive Intelligence of Internet of Things in Precision … 795

The data collected in the server is visualized. The collected data is to be used for
formulating the good crop yield and appropriate crops. The forthcoming scope of
this project is to add extra sensors for better maintenance.

References

1. https://www.hypertech.co.il/product/smart-agriculture/-SmartAgriculture-HyperTech
2. Shekhar Y, Dagur E, Mishra S (2017) Intelligent IoT based automated irrigation system. Int J
Appl Eng Res ISSN 0973-4562 12(18):7306–7320
3. Rajalakshmi P, Devi Mahalakshmi S (2016) IOT based crop-field monitoring and irrigation
automation. Intell Syst Control (ISCO), 2016 10th Int Conf IEEE https://doi.org/10.1109/ISCO.
2016.7726900
4. Kansara K, Zaveri V, Shah S, Delwadkar S, Jani K (2016) Sensor based automated irrigation
system with IOT: a technical review, (IJCSIT). Int J Comput Sci Inform Technol 6(6):5331–5333
5. Avatade SS, Dhanure SP (May 5, 2015) Irrigation system using a wireless sensor network and
GPRS. Int J Advan Res Comput Commun Eng 4(5)
6. Uddin J, Taslim Reza SM, Newaz Q, Uddin J, Islam T, MyonKim J (2012) Automated irrigation
system using solar power, ©2012 IEEE
7. Bai D, Liang W (Oct 2012) Optimal planning model of the regional water saving irrigation and
its application. In: 2012 International Symposium on Geomatics for Integrated Water Resources
Management (GIWRM). IEEE, pp 1–4
8. Case study on software data reduction techniques use for effective bug triage published on Dec
13, 2017 https://www.irjet.net/archives/V3/i11/IRJET-V3I11255.pdf
9. A part of bug report for bug 284541 in Eclipse. www.researchgate.net › figure › A-part-of-bug-
report
Flooring: A Risk Factor for Fall-Related
Injuries in Elderly People Housing

Unesha Fareq Rupanagudi

Abstract Flooring is one of the built features which is playing a crucial role in
elderly housing and also reported as one of the major cause for fall-related injuries
among elderly people. A house with in appropriate flooring leads to falls which
is a major serious accident to people in old age (Designing supportive spaces for
the elderly with the right floors, https://professionals.tarkett.com/en_EU/node/des
igning-supportive-spaces-for-the-elderly-with-the-right-floors-1092, [1]). The study
aimed to ascertain the gap between existing flooring design and necessities of the
elderly people. The elderly women aged sixty and above were the subjects of the
study. The sample was drawn from Kurnool district of Andhra Pradesh. Most of the
elderly people houses were not provided with flooring that has minimum chances
for falls. The elderly people preferred to have non-slippery flooring, provision of
sound-absorbing materials and floor with different colors at various levels to avoid
risk of falls in houses.

Keywords Flooring · Elderly housing · Standard guidelines for flooring design

1 Introduction

A fall is defined as an event which results in a person coming to rest accidentally on the
ground or other lower level or floor level. Now a day’s falls are the second major cause
for injuries and death among the elderly people worldwide and reported as major
public health problem [2, 3]. Flooring in houses is also one of the environmental
risk factors which include home threats. Home hazards such as cluttered flooring
and floor furnishings predict falls among elderly people [4]. Carrying heavy objects,
walking on slippery floors and poor lighting may increase the risk of falls [5]. Kitchen
is the place where most non-fall injuries (31%) occurs and in contrast, fall-related
injuries happen most frequently (20%) in the bedroom [6].

U. F. Rupanagudi (B)
Department of Family Resource Management, APGC, Acharya N. G. Ranga Agricultural
University, Guntur, Andhra Pradesh, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 797
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_80
798 U. F. Rupanagudi

The risk level of fall-related injuries among the elderly may be influenced by
physical, sensory and cognitive levels related to combination of environments and
housing requirements that are not designed for an ageing population [3]. Caring for
the housing of elderly people is becoming one of the greatest challenges in society.
Elderly people feel safer when the house has provided with physical features that
are specifically designed for elderly people keeping in view their needs [7]. Shiny
textured and dark-colored flooring will influence elderly people perception of the
environment as unsafe [8]. Physical features in interiors of the house that support
safety and independence need to be carefully considered while designing the house
for the elderly [9]. The present study aimed to identify the gap between existing
flooring conditions and requirements of the elderly regarding flooring design that
assists the elderly people to live safely and comfortably without any risk in the
houses.

2 Materials and Methods

The study was conducted at Kurnool district of Andhra Pradesh state in India. A
total sample of 60 elderly women in the age of sixty and above who were living in
a separate residence without children with or without spouse was the criteria for the
selection of sample for the study. The study was conducted during the year 2019–
2020. Purposive sampling method was adopted to draw the sample. Standard design
guidelines that ease the elderly people to have a movement without any trouble
proposed by various researchers assisted as a base for identifying the features to
measure the present flooring in the houses of the elderly. The flooring was physi-
cally observed and evaluated against the recommended guidelines. An interview cum
observation schedule was developed. Depending on the presence and absence of the
design feature, scoring was given. Score 3,2,1 were given in case the existing feature
was ‘above the recommended guidelines’, ‘exactly as per the recommended guide-
lines’ and ‘below the recommended guidelines’ respectively. Design requirements
of flooring specified by respondents were quantified in terms of essential, preferred
and neutral with scores 3, 2 and 1. The probable score each respondent can get was
between 8 and 24. The results were interpreted such that the higher the score, higher
was the probability of the flooring as per the recommended guidelines. Frequencies
and percentages were calculated for the existing flooring and needs of the elderly
people. Chi-square statistical analysis was carried out to find out the gap between
existing flooring and design requirements of the elderly people.
Flooring: A Risk Factor for Fall-Related Injuries in Elderly … 799

2.1 Standard Design Guidelines for Flooring

Flooring provided had to be non-slippery both outside and inside the house to avoid
the risk of falls. Slip-resistant floor finishes should be used and shiny reflective floors
such as marble, glazed tiles and the like should be avoided [10–14].
Flooring should visually contrast with the walls, handrails and sanitary fittings.
Avoid using transparent glass materials and mats on floors. The home should be
provided with a large amount of clear floor space. Loose carpets should be avoided
because of the risk of tripping. Provision of bedside carpets in the bedroom should be
avoided [10, 14]. Guiding blocks provided at the walkway should be of red chequered
tile with smooth bricks or stone finish [11].
Floor patterns such as stripes that could be mistaken for steps should not be used for
floors in corridors. The slope of floors should not be greater than1:20 in case of unlev-
elled floor surfaces. If the greater slope is adopted, the floor should be designed as a
ramp. For people with low vision, lines of brightly coloured fluorescent tape should
be placed on the floor surface to assist mobility in poorly lighted areas. Carpeting
should be avoided in circulation areas. Carpets provided in the circulation spaces
should not be deeper than 0.4 inches and has to be securely fixed and firm cushion.
Backing and exposed edges of carpets should be fastened to the floor surface and
trimmed along the entire length of the exposed edge [12]. Flooring should be smooth
and levelled for materials such as ceramic tile or brick or stone. Door thresholds and
minor changes in the floor level should be avoided [13]. Floor surface provided has
to be of matt finish. Projections and variation in floor level that obstruct the easy
access for persons with disabilities should be avoided [10, 12, 14].

3 Results and Discussion

3.1 Existing Flooring Design

Eight standard design guidelines were identified to assess the design of flooring in
houses. Slightly less than three fourth of the houses had no non-slippery floors and
68.33% had loose carpets or rugs in the house. No house was provided with a wooden
floor to accommodate wheel chair users and no provision was made to assist people
with low vision. The majority (91%) of the houses had flooring that contrasts with the
wall. Three-fifths of the houses had avoided bedside carpets in the bedroom. Ninety-
five per cent of the houses had no sound-absorbing materials for walls and floors to
avoid echoes and 90% of houses had no differentiation of color of floor levels and
areas of different surfaces. The existing flooring features in elderly housing were
presented in the Table 1.
800 U. F. Rupanagudi

No extra effort was taken to make flooring non-slippery in the existing elderly
houses. Similar findings regarding bathroom flooring were reported in the study [15].
The provisions to make the floor safe and assist the elderly to lead a comfortable life
in the old age were not made. The flooring was one of the components in the housing
highly neglected to suit to the needs of the elderly.

Table 1 Distribution of respondents by existing flooring design features n = 60

Recommended Position of existing flooring against the guidelines Total
Design Above the Exactly as per the Below the
guidelines recommended recommended recommended
guidelines guidelines guidelines
N % N % N % N %
Non-slippery 4 6.67 12 20 44 73.33 60 100
floors inside as
well as outside
the building
Unwaxed 0 0 0 0 60 100 60 100
wood floors
for the wheel
chair users
Visually 55 91.67 1 1.67 4 6.67 60 100
differentiating
flooring from
the walls
No loose 17 28.33 2 3.33 41 68.33 60 100
carpets or rugs
Sound 3 5 0 0 57 95 60 100
absorbing
materials were
used for floors
and walls to
avoid echoes
No bedside 36 60 0 0 24 40 60 100
carpets in the
bedroom
For people 0 0 0 0 60 100 60 100
with low
vision, lines of
brightly
colored
fluorescent
tape were
placed on the
floor surface to
assist mobility
in poorly
lighted areas
(continued)
Flooring: A Risk Factor for Fall-Related Injuries in Elderly … 801

Table 1 (continued)
Recommended Position of existing flooring against the guidelines Total
Design Above the Exactly as per the Below the
guidelines recommended recommended recommended
guidelines guidelines guidelines
N % N % N % N %
Different 2 3.33 4 6.67 54 90 60 100
colors were
used for
different floor
levels, zonings
or areas of
different
functional
surfaces

3.2 Design Requirements in Flooring

The respondents were inquired to state their recommendations to design flooring to

enable them to have movement in their houses comfortably. Slightly less than three
fourth of the elderly preferred to have no non-slippery floors and also preferred to
avoid loose carpets or rugs in the house. Sixty-five per cent of the elderly felt neutral
for flooring that contrasts with the wall. Nearly half of the elderly preferred to have
sound-absorbing materials for walls and floors to avoid echoes. Sixty-one per cent
had preferred to have floors with colors that differentiate floor levels and areas of
different surfaces in the house. Design requirements in flooring were represented in
Table 2.
The elderly preferred to have non-slippery flooring, provision of sound-absorbing
materials, floor with different colors. The aged individuals felt neutral for flooring
that contrasts with the wall.

3.3 Association Between Existing Flooring Design

and Requirements of the Elderly People

Chi-square analysis was done to find out the association between existing flooring
design and needs of the elderly. The null hypothesis formulated was.
H0 There exists no significant association between existing flooring design
and requirements of elderly people.
More than half (66.67%) of the elderly preferred to have flooring as per the recom-
mended guidelines. Provision of non-slippery flooring in the house, wooden flooring
to accommodate wheelchair users, floor color that visually contrasts with handrails,
walls to enable the elderly to have easy movement were recommended by various
802 U. F. Rupanagudi

Table 2 Distribution of respondents by their requirements in flooring design n = 60

Flooring Adequacy of flooring requirements Total
requirements of Essential Preferred Neutral
the elderly
N % N % N % N %
Floors should be 21 35 39 65 0 0 60 100
non-slippery
inside as well as
outside the
building
Visually 1 1.67 20 33.33 39 65 60 100
differentiate the
flooring from the
walls
Avoid using 23 38.33 36 60 1 1.67 60 100
loose carpets or
rugs
Sound absorbing 1 1.67 31 51.67 28 46.67 60 100
materials should
be used for floors
and walls to
avoid echoes
Different colors 5 8.33 37 61.67 18 30 60 100
should be used
for different floor
levels, zonings or
areas of different
functional
surfaces

researchers. Association between the existing flooring design and requirements of

the elderly people were presented in Table 3.
The Chi-square value was found to be non-significant. There exists no association
between the existing flooring design and requirements of the respondents to age in
place regarding flooring design.
Hence, the null hypothesis was accepted.

4 Conclusion

The elderly felt flooring as one of the important feature that should be considered
while designing the house for elderly people to reduce the risk of fall-related injuries.
The design of flooring was not found as per the recommended guidelines in most of
the houses. The recommended guidelines can be adopted for housing while planning
a house for elderly people, wheelchair users, people with various abilities to avoid
risk of falls in houses.
Flooring: A Risk Factor for Fall-Related Injuries in Elderly … 803

Table 3 Association between the existing flooring design and requirements of the elderly people
n = 60
Existing The design requirements regarding flooring
flooring design Neutral Preferred Essential Total
features
N % N % N % N %
Below the 1 1.67 17 28.33 1 1.67 19 31.67
recommended
guidelines
Exactly as per 1 1.67 40 66.67 0 0.00 41 68.33
the
recommended
guidelines
Total 2 3.33 57 95.00 1 1.67 60 100.00
χ2 value 2.5579
Probability 0.2332
value

References

1. Designing supportive spaces for the elderly with the right floors. https://professionals.tarkett.
com/en_EU/node/designing-supportive-spaces-for-the-elderly-with-the-right-floors-1092
2. Sibley KM, Voth J, Munce SE, Straus SE, Jaglal SB (2014) Chronic disease and falls
in community-dwelling Canadians over 65 years old: a population-based study exploring
associations with number and pattern of chronic conditions. BMC Geriatr 14:22
3. World Health Organization. https://www.who.int/news-room/factsheets/detail/falls#:~:text=
Falls%20are%20the%20second%20leading,greatest%20number%20of%20fatal%20falls
4. Northridge ME, Nevitt MC, Kelsey JL, Link B (1995) Home hazards and falls in the elderly:
the role of health and functional status. Am J Public Health 85(4):509–515
5. Stevens M, Holman J, Bennett N (2001) Preventing falls in older people: impact of an inter-
vention to reduce environmental hazards in the home. J Am Geriatrics Soc 49(11):1442–1447
6. Carter SE, Campbell EM, Sanson-Fisher RW, Gillespie WJ (2000) Accidents in older people
living at home: a community-based study assessing prevalence, type, location and injuries.
Aust N Z J Public Health 24(6):633–636
7. Zamora T, Alcantara E, Artacho MA, Cloquell V (2008) Influence of pavement design
parameters in safety perception in the elderly. Int J Ind Ergon 38(11):992–998
8. Bamzar R (2018) Assessing the quality of the indoor environment of senior housing for better
mobility: a Swedish case study. J Housing Built Environ 34:23–60
9. Engineer A, Esther MS, Bijan N (2018) Designing interiors to mitigate physical and cognitive
deficits related to aging and to promote longevity in older adults: a review. Gerontology 64:612–
622
10. Design guidelines for the elderly and elderly with frailty, https://www.lwb.gov.hk/en/consult_p
aper/BFA_ch6.pdf
11. Guidelines and space standards for barrier-free built environment for disabled and elderly
persons. https://cpwd.gov.in/Publication/aged&disabled.PDF
12. National building code (2016) Requirements for accessibility in the built environment for elders
and persons with disabilities. Bureau of Indian standards, New Delhi, vol 1 pp 49–136. https://
standardsbis.bsbedge.com/
13. Parker WR (1987) Housing for the elderly. In: DeChaira J, Callender J (eds) Time-saver
standards for building types 2nd edn. McGraw-hill international editions. pp 87–101
804 U. F. Rupanagudi

14. Guidelines for the planning of houses for senior citizens. https://www.housinglin.org.uk/_ass
ets/Resources/Housing/Support_materials/Other_reports_and_guidance/
15. Nagananda MS, Sengupta A, Santhosh J, Anand S, Rehman SMK, Khan AM, Rautray P,
Gharai D, Das LK (2010) Circumstances and consequences of falls in community-dwelling
older women. Wseas Trans Biol Biomed 7(4):287–305
Analysis of Construction Readiness
Parameters for Highway Projects

Harshvardhan R. Godbole and R. C. Charpe

Abstract Highways contribute greatly to the economy and growth of a country. For
highway building projects to complete on time and on budget, it is therefore necessary.
However, when the project begins prematurely, interruption often occurs, resulting
in delays, which have had numerous negative results on all project shareholders.
Previous origins are one of the causes of construction hold up. Before studying the
information, however, there is little understanding of whether a highway is premature
or ready to build. The purpose of this research is to find out the criteria consumed in
action to determine whether a road construction project is prepared or not. To this
end, interviews are conducted and analyzed with sixteen practitioners working on
road construction projects. Important attached (1) building readiness can be evaluate
even during the launch phase; and (2) failure to comply with construction prepara-
tion requirements can lead to job delay, wasteful operation, reworking, and labour,
equipment or material deficiencies. The study provides to the current knowledge
by defining criteria which point out whether a road project is ready to be built or
not. Learnings from this study will prevent premature initiation of road construction
projects by industry.

Keywords Highway · Construction readiness · CIDB · Delay

1 Introduction

A highway can be delineated as a public road, particularly a significant road

connecting cities and cities. For each nation, road construction is important because
it contributes to economic and social development. In order to join the needs of
zonal economic development, motorways are necessary component for the quick
delivery of persons and goods. Empirical evidence indicates that the ties between
road transport infrastructure and economic development are substantial and posi-
tive [1]. Although policymakers press for efficient road projects, highways have a

H. R. Godbole (B) · R. C. Charpe

Kalinga University, Raipur, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 805
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_81
806 H. R. Godbole and R. C. Charpe

negative impact on India’s economic growth due to hold up and cost annul [2]. Conse-
quently, it is important for economic and social growth to recognize approaches for
improving incentives to have a successful road construction project.
When finished exact time, within the budget and for stakeholder gratification, a
construction project has been commonly recognized as fruitful. Since at least one
stakeholder is benefitting from early construction, almost always project teams are
under pressure to start constructing, whether or not they are really ready to start.
Premature building begins when a determination is taken by at least one party to
start building at a cost which surpass an adequate indulgence of a party and may
lead to an interruption of construction. Over rusts of costs, sliding schedules, extra
time and poor output are among the most usual and major jolts on projects that do
not start ready [3]. The industry must therefore prevent early start-ups in a building
project which will inevitably have various undesirable impacts. The preparation for
construction is outlined as the series of actions and process to be finished or essentially
finished before construction in order to begin and carry the operations productively.
Many projects are started before they get ready, which leads to decreased capacity
and displeasing project execution (premature start). Construction-ready projects are
reduced on normal by 22% in schedule, 29% in productivity improvements, 20%
in cost savings, 7% in rework savings and 21% less in change in comparison with
projects that are not ready for construction [4]. Moreover, sufficient readiness for
construction can help to avoid many of the factors that lead to sequence works and
ultimately improve project performance [4]. Furthermore, before starting construc-
tion it is important for industry experts to assess the preparation of the sector. This
research examines the parameters used in practice to distinguish between a ready or
unprepared highway projects. To accomplish this goal, this article discusses research
questions concerning the following: What metrics can be used to disseminate between
whether or not a road project is ready for construction? By interviewing seventeen
industry professionals who are employed on road building projects in India, the
writers address the issue. The data derived from the interviews were then investi-
gates with themes. At last, a series of criteria for construction readiness is devel-
oped for road building’. The instruction from this research could prevent industry
escape premature beginning of road building projects by inspecting the construction
readiness of a project.

2 Background

2.1 Construction Readiness

Building readiness is one of the subjects that researchers have researched. A prepara-
tion evaluation identifies potential challenges for the implementation of new policies,
structures and processes in supervisory conditions. The objective of a readiness eval-
uation is to decide if latent obstacles to accomplishment exist and to allow personals
Analysis of Construction Readiness Parameters for Highway … 807

or teams, before starting a project, to overcome these barriers. Assessing the project’s
readiness to build helps the project team avoid premature commencement by quan-
titatively control whether a project is ready for construction with an improvement
in building cost performance, scheduling and development efficiency [4]. Previous
research has identified criteria for assessing building preparation in general to avoid
premature construction starts. However, the characteristics of different kinds of
construction projects vary. Of example, road buildings are horizontal, require more
territories and are physically larger than houses. It also includes working around a
wide and demanding range of current conditions, including roads, wetlands, under-
water or overland networks. The construction preparation requirements common to
highway projects are therefore worth evaluating.

2.2 Highway Construction in India

In India researchers and industry professionals have carried out numerous studies,
because of the significance of road construction for the growth of an economic and
societal region. The main five key advance aspect for Butterworth Outer Ring Road
BORR Expressway in India were identified as an efficient communication system,
excellent project governance, public and private sector accountability and commit-
ment and competitive acquisition and delegation ability [5]. Meanwhile, inadequate
planning, temperature, bad spot management, insufficient site analysis and under-
ground infrastructure are the top five circumstances that assist to the retardation of
road building projects. Five aspects of green highway jargon were pointed out as
general understandings by stakeholders: conservation and habitat protection, energy
and emission mitigation for life cycles, recycling, recycling, re-use and recycling,
wetlands based storm water management, and overall corporate benefit [6]. Different
studies have examined the common issues in road projects in India as well as the
various aspects of road projects, because this is a major subject. There is still a defi-
ciency of investigation into how the road project should be decided whether or not
it is ready to start construction. This research will therefore occupy the space by
analyzing the criteria to determine the quality of road construction projects.

3 Methods

Data collection consists of the collection of data from interviews with professionals
going on road-building projects. The analysis of the collected data is based on qual-
itative approaches. The following sections discuss how the construction readiness
parameters can be collected and analyzed. Figure 1 summarizes the methodology of
this study for examining the practical parameters of whether a road project is ready
to begin construction or not.
808 H. R. Godbole and R. C. Charpe

Fig. 1 Research methodology

3.1 Data Collection

This research gathers data by interviewing professionals in the industry. Individual

interviews were selected to help the interviewer explain, understand, and explore
views and experiences in research subjects as an approach to data collection. In
order to share their incomparable insights, functional knowledge and experiences,
business experts were consulted. The interview staff belongs to companies holding
the G7 grade of the CIDB. A Grade G7 license contracting firm recognized by the
Civil Engineering Construction and Building Construction Board of India shall be
authorized to undertake unlimited projects for civil engineering and building projects.
Therefore, this research aimed to pick G7-grade contractors because they are well
known companies with strong financial potential in the Indian building industry. The
writers are encouraged to contribute as much comprehensive information as possible
to open-ended questions [7]. Data collection from this study includes interviews with
16 valid interviewees. The interview began with an introduction explaining why the
researcher wished to converse with all and what the researcher wished to converse
with. The question for the interview is: What framework might be used to distin-
guish whether a road buildings project is prepared or not? Additional questions were
answered depending on the response of the participants. The following questions
tried to gain a deeper sense of the knowledge they provided and to ensure that their
utterance were correctly assumed. If the participator could not reply or explain the
questions posed, the interviewer attempted to put differently the question and gave
time for a reaction. The interviewer motivated the participators to proceed without
ending their explanations if they started answering. The interview was condensed
and sent to participators for approval direction after each interview.

3.2 Data Analysis

The qualitative dissection includes a topical investigation to recognize parameters

of the construction readiness of road building because the methodology will help to
make the qualitative data meaning [8]. This method was used to analyze qualitative
data on building management subjects by Rahman and Ayer [9] and Radzi et al.
[10]. Depending on the 6 step explanations in Braun and Clarke [8], the thematic
analysis was conducted. The first step familiarizes itself with the details. Interview
Analysis of Construction Readiness Parameters for Highway … 809

data were transcribed, interpreted, re-read and documented by the reporters. The
second phase is the elementary application creation. The writers coded the data for
as many possible topics and trends. The writers then discussed, spoke over, and
decided on any coding enhancements and/or changes. The third stage is to search
for subjects situated on the primary codes. The writers also updated the codes from
the second and the original data from the first stages during the process of creating
the themes. Phase four is to discuss the subjects. The authors tested, identified and
optimized sub-themes constantly in order to ensure that data is saturated, checked
whether the themes work with coded extracts and the whole data set. Phase 5 is
the identification and naming of the subjects. To order to ensure the participants are
responsive to individually coded answers, the authors continue to reverse and reverse
the thematic codes and interpretation of the interview. Finally, the performance of
the study should be recorded (phase six).

4 Results and Discussions

Figure 2 gives a description of the themes and sub-themes of construction preparation

criteria for road building projects that were defined by reviewing interviews with
seventeen professionals in the Indian industry. This analysis identified a total of 28
parameters. The parameters may be divided up into five different sections: ‘Approval’
and ‘General Requirements.’ These five sections may be divided into two themes:
“Start-up project” and “Execution.” The following sections discuss details for each
framework.

Fig. 2 Overview of the construction readiness parameter of highway projects

810 H. R. Godbole and R. C. Charpe

4.1 Parameters Related to Project Start-Up

4.1.1 Approval

In accordance with interview results, the project has received permission from various
legal bodies, such as regional authorities, utilities, the Department of Work and Safety
(DOWS) and the Construction Industry Development Board (CIDB), as part of the
criteria used to determine if development was ready or not. Project approval refers
to the acquisition of permits by the appropriate authorities to ensure that the plan
complies with the standards set in the building regulations and that local authorities
are monitoring certain phases of construction. Officers conduct inspections on the
building site during construction, in order for work to proceed in quality and compli-
ance as expected. If the inspection officers are not informed of permits, they may
stop their work until a decision has been made to minimize effectiveness and detain.

4.1.2 Common Demand

(a) Received letter of award from client. As drafted acceptance that the
entrepreneur has been fortunate and will obtain the job, the customer shall
send the letter of award to the contractor. Other problems may occur without a
letter, as the owner has given the project to the contractor, due to the absence
of a written and signed contract. There is no agreement without a contract, no
safety for the builder and no assurance that the builder is paid for the work he
does. In fact, no agreement is reached on the terms of the contract when the
project manager actually checks them; the cost for the job is often paid and
can’t persuade the customer to change the terms and conditions more appro-
priate or reasonable before starting the work. If a buyer is too expensive, their
cash flow may be affected. The staff may therefore refuse to work because they
can’t pay their wages on time, as a result of the job stopping. Buildings that
would induce loss of productivity without workers could not be achieved.
(b) Acquisition of land completed. Land accomplishment is the procurement
method for the project of the necessary or mandatory property. It is one of the
key elements in the process of soil development. The land procurement method
gives the public sector with a legal procedure to acquire land for advanced
projects to benefit the country. The construction of roads may require local
people to acquire land for the country’s infrastructure and economic devel-
opment. It is often difficult for many reasons to acquire land from the state,
mainly in conjunction with compensation or land prices, thereby lengthening
the duration of the scheme. If the project has started, the building work will
stop until the purchase of the land is completed and productivity losses result.
(c) Adequate money. Funding for a building project is important. Throughout the
building, project director must make sure that the project has sufficient money.
Failure to finance them could lead employees and caters to decline to work
Analysis of Construction Readiness Parameters for Highway … 811

and to afford equipments on the site which could have a negative impact on
building work. However, a shortage of funds will involve the cash flow of the
project and result in delays in site acquisition, resulting in project delays as a
whole [11].
(d) Acquire project insurance. Building insurance is a method of trade for the
fixed payment based premium to secure the interests of the parties involved
with building projects. In the construction sector, it is an important technique
of leading endanger. Construction is a risky and typically volatile activity that
presents a significant risk of personal injury, real property harm and economical
damage. Insurance shortage can cause workers to reject jobs because they may
not have adequate money to pay medical charges if accidents occur.
(e) Date and period of construction confirmed official commencement. The
starting date is the time when the project schedule starts to work. The construc-
tion time is defined in comparison as the time frame given to the contractor by
the owner to complete the project under normal working conditions. Project
managers need to monitor the start and length of construction and schedule
building work accordingly. Failure to plan a project may lead to the delayed
project [12]. Failure to plan means that the schedule to be followed by workers
is not established. When the workers work on projects, they won’t have a clear
picture. There are no deadlines to reach, so that the workforce’s loose environ-
ment reduces productivity. It means that the project is not finished on a late
basis.
(f) The nearest authority. Building site theft can cause different project problems.
Therefore, to prevent crime at the construction site, work with local authorities
is essential. Project managers should verify and inform local authorities of
the site. For fact, on weekends and holidays, the project managers can also
seek the authority to monitor the site if the contractor is afraid of fraud. To
addition to the direct costs of repairing the stolen goods, the cost of renting the
stolen equipment is also borne and revenue is destroyed. The project schedules
can also be delayed due to interruptions due to a shortage of equipment and
material.
(g) Meeting with consultant and client. The client is the creator of the project,
and contractors include architects, structural engineers and others who are
hired by the customer to carry out expert tasks for a project. Stakeholders will
discuss the outstanding questions on the project and provide feedback during
this meeting. Productivity may be diminished due to unresolved problems.
Furthermore, there could be no disruption to the project timeline to certain
actions on the project.
(h) Verified construction work plan. A work plan is a series of actions or tasks
that have to be carried out well for a successful project. In fact, a job plan
shows all the tasks of a project that are involved and responsible for each task
and when the tasks are completed. It can help the contractor remain organized
during the project work and help him complete the project in time and on
budget. Starting buildings without a work plan could lead to several problems,
including reduced productivity and sliding schedules.
812 H. R. Godbole and R. C. Charpe

(i) Client and consultant approved construction work plan. In order to ensure
no conflict between parties, customer and consultant should know the work plan
that has been established by the contractor Failure to understand the project
work plan could cause stakeholder conflict. If litigations are not managed
adequately, they will lead to project delays, a mentality of teams that is not
decided, project costs are higher and business relations are damaged [13].
However, there may be work stoppage due to the lack of competitiveness and
delay due to a disagreement between stakeholders. Therefore, before beginning
a job, the customer and contractor will approve the work plan.

4.1.3 Requirement for Drawing

(a) Drawings approved by consultant. Until project management can start the
project, it is important that the contractor will review and authorize the design
drawing. The approval of a consultancy can ensure compliance with designs,
specifications and standards of construction. During the construction period,
buyers often change the design, which at the same time affects the entire
project life. Until the latest drawing problem from the architect, the contractor
can do the construction work. The job cannot therefore be completed within
the contract period resulting from delays and reduced productivity. Therefore,
drawings must be accepted by the contractor prior to construction in order to
prevent work interruptions.
(b) Governments approved drawings. Authorities are accepting plans to ensure
the safety of new building for human beings and as laid down in the planning
rules. The process of approval tests if the structure is secure and structurally
sound and whether the approved building materials are safe for human life.
Starting with a drawing approved still by an authority might pose a risk to the
plan for a project. If officials at the site check and realize that the drawing is
not allowed, they can stop the works causing delays and lost productivity.
(c) Everything is provided in complete drawings. To guarantee that the project
will be done in a timeline, full sketches are essential. Errors and omissions
may result in project delays causing the drawings to be incomplete. The plan is
vague during building because significant details are not shown in the drawing;
workers cannot continue the work during construction work. The research must
therefore stop immediately until it is possible to obtain details of that particular
drawing which lead to loss of productivity and delay.
(d) Verified variations in drawings and designs. Differences between tenders and
drawings will result in project delay. The bidders (contractors) are selected at
the time of tender. For building purposes, the design drawing shall be provided
by the architect. Sometimes these sketches vary. There are variations. These
differences contribute to rework, quantity adjustments and construction delays
and defects. In addition, refurbishments result in reduced productivity [14].
Analysis of Construction Readiness Parameters for Highway … 813

4.2 Parameters Related to Project Execution

4.2.1 On-Site

(a) Site office. The building site includes office facilities to provide project
managers with housing, rooms for meetings, and the site records storage. The
site office is also important to ensure that all documentation of the project
is preserved safely. When the project begins before the site office is built,
documents including project files and sketches can be seen anywhere. Lack of
project documents such as reports or sketches may cause a project schedule to
be disrupted. Of example, before the work can restart, project managers need
to find the missing sketches which lead to lost productivity and delay.
(b) Site condition as per the contract. Project managers should ensure that the
conditions of the site are identical to those stated in the contract and that no
different conditions are observed. The site condition of the project manager
is different when the site condition varies significantly from the terms of a
contract with the owners of the project or what is normally expected in the site
area. The site condition is different. Differing conditions at the site will cause
the project to stall due to work on the site until the relevant stakeholders have
made decisions [15].
(c) Verified utility cables and relocate if necessary. Transfer of the services
means the modification, substitution or relocation as required by the road-
building project: elimination and reconstruction of the facility, procurement
of the appropriate right-of - way, displacement or rearrangement of existing
infrastructure, alteration of the form of utilities and the introduction of the
necessary safety and protection measures [16]. Due to construction, damage
to underground utilities can result in weeks of delays. For example, hitting
a water pipe unintentionally will cause the project to stop momentarily. The
flood will splash and inundate the field, making the site impossible to work.
Public issues and legal problems will also be posed which will prolong the
wait.
(d) Verified traffic diversion and control system. The diversion of traffic system
is designed to prevent traffic around the building site being disrupted by the
project, and to reduce the disruption of traffic if desired. The traffic control
system, on the other hand, relates to a traffic management system in and around
building sites. In terms of forbidding an incident at a building site, these criteria
may be related. Vehicles are often the main cause of deaths and injuries to
workers and public members when reversing, loading and unloading. The root
cause of staff and construction vehicle accidents is inadequate coordination and
monitoring. Accidents involving a building vehicle may result in an eventual
overrun of the time due to an accident. A building workers accident may also
lead to work shortages, which ultimately lead to construction delays.
(e) Safety signboard. (e) In any working environment, safety signs are key. The
primary importance of a safety sign is to ensure that workers are well aware of
814 H. R. Godbole and R. C. Charpe

the potential hazards in particular situations and environments. The safety signs
must be posted. Without signs, there would be a lack of direction from many
workers in times of crisis, so companies could face major legal problems should
accidents occur. Because of insufficient satellite safety accidents involving
construction workers can reduce labour and efficiency, resulting in disruption
of work and eventually a delay in planning [17]. Furthermore, the number of
workers involved in job disruption can be impacted by injuries. Sometimes it
takes more time than expected to find a substitute for the injured workers.
(f) Provision of CCTV and checked building engine parking space. CCTV is
a network which sends TV signals to a limited set of screens and is widely
used for crime prevention in shops and public places. Conversely, the checked
construction machinery parking space is concerned with the design of construc-
tion machinery parking space. To order to prevent the occurrence of crime
on the construction site, CCTV deployment and parking inspection for the
construction machinery can be connected. Robbery and vandalism today on
commercial construction sites.
Industry can influence productivity and income runoff. This is a concern. The
deployment of security cameras is one of the approaches used to prevent crime
at the construction site [18]. In the allocated parking space at the site, in order
to prevent crime, construction equipment should also be stored. Theft will
easily target computers without parking at the site that are left unattended in
remote areas. Companies must pay to replace or temporarily hire the damage
equipment to support their projects in due course. Nevertheless, there are
other consequences, not least the resultant downtime due to the lack of equip-
ment available. It is therefore essential to ensure that pre-construction parking
area for building machinery is provided and that CCTV was installed prior to
construction commencement in order to prevent an equipment shortage.
(g) Verified temporary utilities on site. Several services like water, electricity,
internet connection, lighting, etc., need to be temporarily established before
work on a construction site can begin. Building site can interrupt work on
a building project without electricity, causing a delay. To order for project
managers to coordinate with clients about the job, Internet and telephones are
required on the construction site. One reason behind delays in the Ethiopian
construction industry is the lack of readily available on-site services [19].
(h) Adequate work force. Building workers mean those employed in the building
industry. Workers’ shortages cause project planning interruptions. For example,
inadequate operators on the site will disorder construction, as the workforce
is insufficient for building equipment to operate. In fact, given the preparation
of material, an insufficient worker will control the length of the projects. The
lack of foreign and local workers is one of the most common factors leading
to delays in the Indian building industry [20].
Analysis of Construction Readiness Parameters for Highway … 815

4.2.2 Equipment and Materials

(a) Verified nearest material supplier and quarry. The provider of materials
means any individual providing equipment, materials, services and to be used
in connection with the building contract work. On the other side a quarry is a
place where rocks, sand and gravel are excavated from the ground, dimension
stone, slate building aggregate. The geographical location and delivery lead
time transport and managerial charges may be affected by material supplier or
quarry. The option of a source or quarry close to the site, for example, can help
ensure that the material arrives at the correct time. Providing remote suppliers
or quarries may mean longer delivery times and additional costs for carriage.
Unless a supplier delivers material on time, the processes can be interrupted and
the project conclusion postponed. One of the reasons influencing productivity
is the availability of materials on site.
(b) Equipment availability. A variety of specialized equipment is necessary for
road construction. A lack of equipment affects labour productivity as equip-
ment supports the construction process is used to support the workforce. The
digger is used, for instance, to excavate and move a large object. Building
material could not be relocated without this resulted in a delay of the project
and work halt. In the Indian construction industry, machinery supply is also
one reason for delay [21].

5 Conclusion

By reviewing interviews with sixteen industry practitioners in India, this study

has identified construction preparation criteria for road projects. Finally, readiness
parameters were identified for either the beginning or execution of the project.
Building preparation can also be tested at the outset of the road construction project.
Starting building also can lead to a working halt, inadequate working, rework, and
shortening of work, machinery or materials, without properly fulfilling the criteria.
These results emphasize the need for practitioners in industry to evaluate the readi-
ness of a project before construction begins. Although some of the framework placed
in this research can be applied to other building types, there are various parameters
for road construction. In particular, the preparation of general construction variables
are not component in criteria such as completed land acquisition, checked traffic
around the building site and traffic control systems. Such results suggest.
In highway construction, land acquisition is vital as road projects are generally
long and involve a lot of land. Meanwhile, due to the risks faced by construction
workers, the traffic control system and checks are necessary before construction is
undertaken close to the passing highways. In other words, this research adds to the
current information on road development by presenting basic criteria relevant to the
816 H. R. Godbole and R. C. Charpe

evaluation of road projects’ preparation for construction. If correctly applied, the

lesson from this study will help the industry prevent premature start in highway
construction projects by assessing the preparation for project construction.

References

1. Chandra A, Thompson E (2000) Does public infrastructure affect economic activity?: Evidence
from the rural interstate highway system. Reg Sci Urban Econ 30(4):457–490
2. Bank Negara India (2018) Bank Negara India quarterly bulletin Q4
3. Griego R, Leite F (2016) Premature construction start interruption: how awareness could
prevent disputes and litigations. J Leg Aff Dispute Resolut Eng Constr 9(2):04516016
4. Ibrahim MW (2018) Improving project performance by mitigating out-of-sequence work and
assessing construction readiness. The university of Wisconsin-Madison
5. Ghazali FM, Rashid SA, Sadullah AM (2017) The critical success factors for public-private
partnership highway construction project in India. J Eng Technol (JET) 8(1)
6. Nusa FNM, Nasir S, Endut IR (2018) Awareness of green highway concept and terminology: a
perspective of on-site personnel in Indian highway construction industry. Advan Trans Logistics
Res 1(1):475–487
7. Turner DW III (2010) Qualitative interview design: a practical guide for novice investigators.
Qual Rep 15(3):754–760
8. Braun V, Clarke V (2006) Using thematic analysis in psychology. Qual Res Psychol 3(2):77–
101
9. Rahman RA, Ayer SK (2017) Prevalent issues in BIM-based construction projects. In:
Proceedings of joint conference on computing in construction, vol 1. pp 645–652
10. Radzi AR, Bokhari HR, Rahman RA, Ayer SK (2019) Key attributes of change agents for
successful technology adoptions in construction companies: a thematic analysis. In: Computing
in civil engineering 2019: data, sensing, and analytics. American Society of Civil Engineers,
Reston, VA, pp 430–437
11. Abdul-Rahman H, Berawi MA, Berawi AR, Mohamed O, Othman M, Yahya IA (2006) Delay
mitigation in the Indian construction industry. J Constr Eng Manage 132(2):125–133
12. Marzouk MM, El-Rasas TI (2014) Analyzing delay causes in egyptian construction projects.
J Adv Res 5(1):49–55
13. Cheung SO, Suen HC (2002) A multi-attribute utility model for dispute resolution strategy
selection. Constr Manage Econ 20(7):557–568
14. Jarkas AM, Bitar CG (2011) Factors affecting construction labor productivity in Kuwait. J
Constr Eng Manage 138(7):811–820
15. Amarasekara WD, Perera BA, Rodrigo MNN (2018) Impact of differing site conditions on
construction projects. J Legal Aff Dispute Resolut Eng Constr 10(3):04518006
16. Jung YJ (2012) Evaluation of subsurface utility engineering for highway projects: benefit–cost
analysis. Tunn Undergr Space Technol 27(1):111–122
17. Durdyev S, Omarov M, Ismail S (2017) Causes of delay in residential construction projects in
Cambodia cogent. Engineering 4(1):1291117
18. Berg R, Hinze J (2005) Theft and vandalism on construction sites. J Constr Eng Manag
131(7):826–833
19. Gebrehiwet T, Luo H (2017) Procedia engineering, vol 196. p 366–374
Analysis of Construction Readiness Parameters for Highway … 817

20. Abdul Kadir MR, Lee WP, Jaafar MS, Sapuan SM, Ali AAA (2005) Factors affecting
construction labour productivity for Indian residential projects. Struct Surv 23(1):42–54
21. Sambasivan M, Soon YW (2007) Causes and effects of delays in Indian construction industry.
Int J Proj Manage 25(5):517–526
Biogas Generation by Utilizing
Agricultural Waste

Prathamesh Chaudhari and Shivangi Thakker

Abstract Biogas is produced using agricultural waste and cow dung with help of the
anaerobic co-digestion in a biogas digester. Generally cow dung is used as the raw
material for the Biogas generation. In this biogas production using different compo-
sition of the cow dung and agricultural waste as right now cow dung is indispensable
member for the biogas production with agricultural waste. Also other factors such
as retention time, temperature can also be observed while taking different compo-
sitions. Theoretical maximum biogas yield is calculated for different raw material
composition for biogas production. It is an extensive study on design of the biogas
digester and how agricultural waste can be used to produce biogas. Design of experi-
ment was performed with computer software to predict the biogas yield and to decide
important factor responsible for biogas generation.

Keywords Biogas · Anaerobic digestion · Design of experiment

1 Introduction

Energy is an important factor for the economic progress and further improving the
standard of the living of the people. Due to ever increasing demand of the energy for
industrial as well as commercial sector it has become a major issue. Due to environ-
mental impact of currently used sources of energy as the natural gas, fossil fuels, coal
and crude oil there is need for using alternate renewable sources. Renewable energy
systems are considered as a favorable key that plays an essential part in solving
the issue on sustainable resources as it provides clean and efficient. Energy sources
that are considered renewable are those that are naturally and continually refilled by

P. Chaudhari (B)
K. J. Somaiya College of Engineering, Vidyavihar, Maharashtra, India
e-mail: [email protected]
S. Thakker
Department of Mechanical Engineering, K. J. Somaiya College of Engineering, Vidyavihar,
Maharashtra, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 819
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_82
820 P. Chaudhari and S. Thakker

nature such as solar, hydro, wind and biogas and are very environment-friendly due
to the fact that these resources produce no carbon emissions.
In this paper Minitab software was used to make a design of experiment for
finding out the optimization of the biogas yield with an agricultural waste, theoretical
maximum biogas potential was also calculated to few agricultural waste. This method
of finding design of experiment model can be utilized for various other agricultural
wastes to optimize the process of biogas generation.

2 Literature Review

2.1 Agricultural Waste for Biogas Production

Biogas as an alternate energy source is growing across world and technological

advances in anaerobic digestion and feedstock optimization has further increased
its importance [1]. The biogas can be produced by the anaerobic digestion of
biodegradable material such as manure, food waste agricultural waste and energy
crops [2].
Sugarcane bagasse is found in abundance which is obtained from sugar industry
after extraction of juice and it is almost 25% of the total processed sugarcane.
Advanced technological improvement can be incorporated in the biogas production.
Biogas yield of biogas plant depends on the various factors [3–5].
The model used to predict maximum biogas potential from the composition of
the raw materials. By using various pretreatment processes it improves quality of
biomass biogas production [6].

2.2 Design of Experiment

Design of experiment is used to determine the relationship between factors affecting

a process and the response generated by the process. It is used to find cause and
affect relationships between input variables and output. Design of Experiment can
be used for deciding major factor affecting the results in given experiment. The
DOE experiment performed on the biogas yield as per different beating time and
temperatures [7–12].
Minitab software tool was used to perform the design of experiment on
experimental data.
Biogas Generation by Utilizing Agricultural Waste 821

3 Theoretical Analysis of Biogas Potential

With knowledge of the chemical composition of a biomass the quantity of methane

can be predicted from the stoichiometric formula developed by Buswell and Hatfield
in 1936 [8].

3.1 Material and Methods

The chemical composition of the raw material to be fed to biogas reactor has to be
known to apply this model. This model (refer Fig. 1) utilizes only carbon, hydrogen
oxygen and sulfur as obtained from ultimate analysis of the raw materials. This
ultimate analysis values are obtained from published values for ultimate analysis of
these agricultural wastes.
Below are the values obtained from literature review for the raw material as
bagasse, rice straw and cow dung. As it is a biological data composition mainly
depends on the condition they are harvested and are going to be different as per
conditions. The balance between simplicity and effective production of the biogas
are to be achieved with help of this model.
This model can predict biogas output if it’s considered that reaction goes into
completion. This model is simplified and is used to estimate the theoretical maximum
biogas potential with help of ultimate analysis composition of raw materials. The
composition of various biomasses is shown in Table 1.
Different existing models utilize different approaches to predict the biogas yield.
Buswell and Hatfield developed the stoichiometric formula to predict the quantity
of methane from information of the chemical composition of an agricultural waste

Fig. 1 Biogas digester

822 P. Chaudhari and S. Thakker

Table 1 Composition of
Raw material Rice straw Bagasse Cow dung
various biomasses
C 34.11 49.2 49.04
H 6.53 4.69 6.43
O 59.19 0.18 2.25
N 0.17 43 41.94
S 0.1 0.02 0.34

which is shown in equation below:

 y z x y z x y z
CxHyOz + x − − H2 O → + − CH4 + − + CO2
2 2 2 8 4 2 8 4
This reaction of Buswell and Mueller was modified by Boyle to further
consideration of ammonia and H2S within the produced Biogas.
   
b c 3d e b c 3d e
CaHbOcNdSe + a − − + + H2 O → a + − − − CH4
4 2 4 2 8 4 8 4
 
a b c 3d e
+ − + + + CO2 + dNH3 + eH2S
2 8 4 8 4

This can be written in simplified form as:

A + C1 B C2 C + C3 D + C4 D + C5 F

Where A and B are reactant and C, D, E, F are the products and constant of the
reactions are:
   
b c 3d e b c 3d e
C1 = a − − + + C2 = a + − − −
4 2 4 2 8 4 8 4
 
a b c 3d e
C3 = − + + + C4 = d C5 = e
2 8 4 8 4

It can be seen that reaction can be applied to any biomass which is used to produce
the biogas with anaerobic digestion. This model makes and assumption that these
elements are the sole components of the raw biomass. Ultimate masses of the feed-
stock are taken to calculate the maximum methane potential, which are used as the
variables.
a · mass · ultimate a · mass · ultimate
a= =
mmC 12.0107
b · mass · ultimate b · mass · ultimate
b= =
mm H 1.0079
Biogas Generation by Utilizing Agricultural Waste 823

c · mass · ultimate c · mass · ultimate

c= =
mm O 15.999
d · mass · ultimate d · mass · ultimate
d= =
mm N 14.0067
e · mass · ultimate e · mass · ultimate
e = =
mm S 32.065
So, molar mass of the Reactants and products are found out as:

mmA = a*mmC + b*mmH + c*mmO + d*mmN + e*mmS

g
= 12.017a + 1.0079b + 15.999c + 14.0067d + 32.065e in
mol
In similar method molar amss of the each reactant and product are calculated.
g
mmB = 2*mmH + 1*mmO = 18.0158
mol
g
mmC = mmC + 4*mmH = 16.04
mol
g
mmD = 1*mmC + 2*mmO = 44.02
mol
g
mmE = 3*mmH + 1*mmN = 17.03
mol
g
mmF = 2*mmH + 1*mmS = 34.08
mol

3.2 Calculations

By using Boyle modified formula we can estimate the biogas yield from the elemental
composition of biomass. The theoretical biochemical methane potential (TBMP) of
the material is calculated in ml CH gVS−1 as:
 
22.4 × a + b8 − 4c − 3d8 − 4e
TBMP = (1)
12.017a + 1.0079b + 15.999c + 14.0067d + 32.065e

3.3 Results Obtained from the Calculations

The values of TBMP obtained from calculations are presented in Table 2.

824 P. Chaudhari and S. Thakker

Table 2 Result table values

Raw material TBMP values (ml Corrected TBMP
of TBMP values
CH gVS-1) values (ml CH
gVS−1 )
Cow dung 37,805 30,244
Cow dung(70) + 35,197 28,158
Rice Husk(30)
Cow dung(60) + 34,408 27,526
Rice Husk(40)
Cow dung(50) + 33,493 26,794
Rice Husk(50)
Cow dung(70) + 36,667 29,334
Bagasse(30) 36,330 29,064
Cow dung(60) +
Bagasse (40)
Cow dung(50) + 36,000 28,800
Bagasse (50)

4 Design of Experiment

In this study Biogas yield from glass silage is optimized and compared with the
mechanical beating of the glass silage as well as varying temperatures in the
controlled conditions. An experimental data required for the DOE was taken from
the paper by FATMA A. ALFARJANI [7]. In this paper anaerobic digestion of glass
silage is performed and beating time and temperature were taken as main factors.

4.1 Response Surface Methodology (RSM).

It is often required to find the conditions, which would optimize the process of
requirement. We require the process input parameters values for which we will obtain
optimized response.

4.2 Design Matrix

Design matrix is generated by the Minitab software as per the conditions (refer
Table 3) fed to the software then experiments are performed according to the design
matrix. Glass silage was used as the biomass waste in the given paper and face
centered composite design was used for design matrix generation.
Experiment was performed for the values given by the design matrix and Table 4
shows that biogas yield in cc for collections as it was collected over span of 24 days
with taking reading every third day so it shows collections.
Biogas Generation by Utilizing Agricultural Waste 825

Table 3 Design matrix table generated by software

Exp. No. 1 2 3 4 5 6 7 8 9 10 11
Beating time 0 10 0 10 0 10 5 5 5 5 5
Temperature 35 35 39 39 37 37 35 39 37 37 37

Table 4 Experimental yield of biogas in cc

Beating time Temperapture Collection in cc
(min) (°C) First Second Third Fourth Fifth Sixth Seventh
0 35 1021 1769 2052 2225 2425 2560 2673
10 35 1049 1538 1807 2061 2272 2386 2477
0 39 1570 2128 2471 2694 2858 2950 3027
10 39 1409 1877 2252 2464 2576 2664 2728
0 37 1490 2079 2425 2674 2798 2926 3023
10 37 1252 1762 2149 2422 2521 2606 2675
5 35 1111 1651 1939 2160 2370 2497 2592
5 39 1418 1983 2400 2629 2788 2887 2973
5 37 1394 1918 2306 2551 2669 2768 2868
5 37 1491 1998 2368 2609 2707 2807 2901
5 37 1526 2072 2486 2723 2859 2969 3063
5 37 1582 2106 2518 2754 2887 2991 3088

4.3 Response Surface Regression: Seventh Versus Beating

Time, Temperature

Regression Equation in Un-Coded Units.

Seventh collection = −55305 + 105 Beating Time + 3057 Temperature

− 3.74 Beating Time ∗ Beating Time
− 40.0 Temperature ∗ Temperature
− 2.57 Beating Time ∗ Temperature.

4.4 Contour Plot of RSM

If the data obtained from the experiments is plotted on the graph then we get the
better picture of the effect of the temperature and beating time on the biogas yield.
Figure 2 illustrate contour plot for the effect of beating time and temperature on
826 P. Chaudhari and S. Thakker

Fig. 2 Contour plot for seventh collection

biogas yield for the seventh collection. Which shows that biogas yield is higher for
lower beating time and as beating time is increased biogas yield decreases. As for
temperature higher the temperature higher will be the biogas yield.

5 Conclusion

Biogas Generation is studied with Anaerobic Digestion. It was found that maximum
theoretical potential reduces as the cow dung in reduced and agricultural waste
is added. In case of Cow dung maximum TBMP was calculated. For Rice husk
maximum biogas yield affects more negatively than that of bagasse. Design of experi-
ment Response Surface Method showed that Temperature is the major factor affecting
the biogas yield than beating time. Optimal condition obtained at high tempera-
ture and low beating time, as beating time increases it affects negatively on yield.
Experiments can be performed for the different agricultural waste and for varying
compositions to find optimum composition and condition for biogas production.
Biogas Generation by Utilizing Agricultural Waste 827

References

1. Kigozi R, Muzenda E, Aboyade AO (2014) Biogas technology, current trends, opportunities and
challenges. In: Proceedings of the 6th international conference GREEN technology, renewable
energy and environmental engineering. Cape Town, South Africa, pp 27–28
2. Ilaboya IR, Asekhame FF, Ezugwu MO, Erameh AA, Omofuma FE (2010) Studies on biogas
generation from agricultural waste; analysis of the effects of alkaline on gas generation. Worl
Appl Sci J 9:537–545
3. Vats N, Khan AA, Ahmad K (2019) Observation of biogas production by sugarcane bagasse
and food waste in different composition combinations. Ener 185:1100–1105
4. Almomani F (2019) Prediction of biogas production from chemically treated co-digested
agricultural waste using artificial neural network. Fuel 280:118573–118586
5. Kucukkara B, Yaldiz O, Sozer S, Ertekin C (2011) Biogas production from agricultural wastes
in laboratory scale biogas plant. J Agri Machina Sci 7:373–378
6. Achinas S, Gerrit JWE (2016) Theoretical analysis of biogas potential prediction from
agricultural waste. Resou Effic Techno 2:143–147
7. Alfarjani F (2012) Design and optimization of process parameters in bio-gas production
systems. Dublin City University, Diss
8. Buswell AM (1939) Anaerobic fermentations. Bulletin (Illinois State Water Survey). 32
9. Mittal S, Ahlgren E, Shukla PR (2019) Future biogas resource potential in India: a bottom-up
analysis. Renew Ener 141:379–389
10. Guarino G, Carotenuto C, Cristofaro F, Papa S, Morrone B, Minale M (2016) Does the C/N ratio
really affect the bio-methane yield? a three years investigation of buffalo manure digestion.
Chem Engg Trans 49:463–468
11. Chandratre SJ, Chaudhari V, Kulkarni B, Mahajan B, Bavaskar KP (2015) Biogas production
from local agricultural waste by using laboratory scale digester. Res J Rec Sci 4:157–165
12. Mogal V (2016) A review on quality by design. Pharm Biol Eval 3:313–319
Effect of Change in the Resilient Modulus
of Bituminous Mix on the Design
of Flexible Perpetual Pavement

Saurabh Kulkarni and Mahadeo Ranadive

Abstract High modulus bituminous mix is one of attractive option to increase load-
bearing capacity of pavement structure against conventional structural distresses,
such as rutting and fatigue crack. A comparative study was carried out between five
combinations of flexible pavement mentioned in IRC guidelines with conventional
perpetual pavement design criteria. All five combinations were subjected to high
modulus bituminous mix of 3000 MPa to 8000 MPa. The perpetual pavements were
designed by using mechanistic-empirical design software IITPAVE. The results were
further validated using the design software KENPAVE and WESLEA. The study
investigates the effect of high modulus bituminous mix in conventional perpetual
pavement design with respect to life cycle cost analysis and overall pavement thick-
ness. It is found that use high modulus bituminous mix results in substantial decrease
in pavement thickness.

Keywords Perpetual pavement · Mechanistic analysis · Life cycle cost

1 Introduction

A perpetual pavement is a long lasting pavement designed to last for about 50 years.
It does not undergo major structural rehabilitation or reconstruction and needs only
periodic surface renewal in response to distresses confined to the top of the pave-
ment [1]. The mechanistic empirical approach is generally observed for analysis
and design of perpetual pavements. Various limiting values of strain for different
layers of pavement are considered while designing or analysing the performance of
perpetual pavements. As a perpetual pavement structure is subjected to only non-
structural deteriorations, only a periodic surface renewal is generally required. The
critical characteristic of perpetual pavements lies in the fact that it is never replaced
entirely and removed. The point that only the surface layer is renewed with the base
structure staying in place, there is a considerable reduction of construction materials
[2]. There were no official standard provisions for the design of perpetual pavement

S. Kulkarni (B) · M. Ranadive

Department of Civil Engineering, College of Engineering Pune, Pune, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 829
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_83
830 S. Kulkarni and M. Ranadive

in India until the year 2012. However, with the latest publication of the guidelines
by the Indian Roads Congress [1] the design to satisfy 50 year design periods can be
adopted with mechanistic pavement design methodology.

1.1 Objectives and Scope

The specific objectives of this study are to,

• Design conventional perpetual pavement structure using the design philosophy of
IRC 37:2018 by considering increasing values of resilient modulus of bituminous
mix.
• Perform life cycle cost analysis (LCCA) comparison between five different combi-
nations of perpetual pavement with different resilient modulus of bituminous
mix.
• Examine the effect of changing resilient modulus of bituminous mix on overall
thickness and cost of pavement.

2 Design and Analysis

The tensile strain acting in the horizontal direction at the bottom of asphalt layer
and the compressive strain acting in the vertical direction at the top of subgrade are
considered as the critical strains in any pavement structure. The general principle
of perpetual pavement design is to keep these strains in some particular limits, and
many researchers have adopted a maximum limit of 70 microstrains (µ) and 200
microstrains (µ) for horizontal tensile strain and vertical compressive strain respec-
tively [3]. Indian Roads Congress has proposed the strain values of 80µ and 200µ
respectively for the fatigue and rutting endurance limit [4]. The Endurance Limit
(EL) is a level of strain below which there is no cumulative damage over an infinite
number of cycles. A bituminous layer experiencing strain levels less than EL should
not fail due to fatigue. Similarly, if compressive strain acting in the vertical direction
at the top of subgrade is less than the EL, it should not fail due to rutting. Signifi-
cance of EL is that, such a limit would provide a thickness limit for the pavement and
increasing the thickness beyond this limiting thickness would provide no increased
structural resistance to fatigue and rutting damage.

2.1 Sample Design for Analysis

In IRC 37:2018, five different categories of flexible pavements are given. The five
trial combinations proposed are:
Effect of Change in the Resilient Modulus of Bituminous … 831

1. Bituminous Concrete and Dense Bituminous Macadam (BC+DBM) with

granular base and sub-base (Composition A)
2. BC+DBM with Cement Treated Base (CTB), Cement Treated sub base (CTSB)
and granular crack relief layer (CRL) (Composition B)
3. BC+DBM with CTSB, CTB and with Stress absorbing membrane interlayer
(SAMI). (Composition C)
4. BC+DBM with Granular Sub-base (GSB), CTB and granular crack relief layer
(Composition D)
5. BC+DBM with CTSB and GSB. (Composition E).
As per IS SP53:2010 [4], SAMI may consist of elastomeric modified binder like
Styrene-Butadiene Rubber (SBR) applied at the rate of minimum 1 kg/m2 . For the
pavement analysis, the SAMI layer is not considered as a structural layer [1]. As per
the IRC guidelines, Viscosity Grade 40 (VG 40) bitumen shall be used for surface
course and for the DBM and it shall have a minimum viscosity of 3600 Poise at 60 °C
temperature to safeguard against rutting.

2.2 Fatigue and Rutting Criteria

For the present analysis, the fatigue model suggested in the Indian Road Congress
(IRC 37:2018) guidelines was adopted. The model was calibrated in the research
scheme R-56 (1999) studies using the pavement performance data collected during
the research scheme R-6 (1995) studies and R-19 (1994) studies 90% (Eqs. 1 and 3)
reliability levels.

1 3.89 1 0.854
Nf = 0.561 × C × 10−4 × ( ) ×( ) (1)
εt MRm

where
C = 10 M , and
M = 4.84 × [(V be / V a + V be ) – 0.69]
Va = percent volume of air void in the mix used in the bottom bituminous layer,
Vbe = percent volume of effective bitumen in the mix used in the bottom bituminous
layer, Nf = fatigue life of bituminous layer (cumulative equivalent number of 80 kN
standard axle loads that can be served by the pavement before the critical cracked
area of 20% or more of paved surface area occurs), 1t = maximum horizontal tensile
strain at the bottom of the bottom bituminous layer (DBM) calculated using linear
elastic layered theory by applying standard axle load at the surface of the selected
pavement system, MRm = resilient modulus (MPa) of the bituminous mix used in
the bottom bituminous layer, selected as per the recommendations made in these
guidelines.
832 S. Kulkarni and M. Ranadive

The factor ‘C’ is an adjustment factor used to account for the effect of variation
in the mix volumetric parameters (effective binder volume and air void content) on
the fatigue life of bituminous mixes and was incorporated in the fatigue models
to integrate the mix design considerations in the fatigue performance model. For
this paper, the Va , Vbe and ‘C” factor values considered are 3.5%, 11.5% and 2.35
respectively as recommended in IRC guidelines. So Eq. 1 becomes

1 3.89 1 0.854
Nf = 1.32 × 10−4 × ( ) ×( ) (2)
εt MRm

For the present analysis, the rutting model suggested in the Indian Road Congress
(IRC 37 2012) guidelines was adopted as shown in Eq. 3. The model was calibrated
in the research scheme R-56 (1999) studies using the pavement performance data
collected during the research scheme R-6 (1995) studies and R-19 (1994) studies at
90% reliability levels.

1 4.5337
NR = 1.41 × 10−8 × ( ) (3)
εv

where NR is the cumulative number of repetitions for rutting failure, 1v is the vertical
strain. According to the concept of fatigue endurance limit, if the tensile strain caused
by the traffic in the bituminous layer is less than 70 microstrains, as per tests conducted
in laboratories at 20 °C in USA, the bituminous layer never cracks. Similarly, if the
vertical subgrade strain is less than 200 micro strain, rutting in subgrade will be
negligible. For a pavement temperature of 35 °C, the endurance limit is about 80µ
and 200µ [4]. Hence in this study, the design of pavement is carried out by trial
and error with IITPAVE till endurance limit of the strain values closest to 80µ and
200µ respectively for the fatigue and rutting endurance limit are obtained. Thus, it
will also avoid overly conservative design. For inputs to the IITPAVE, KENPAVE
and WESLEA software, a single axle dual wheel assembly was considered for the
analysis. The standard axle load considered was 80 KN. The contact radius was
assumed as 15.5 cm with a tire pressure of 0.56 MPa. Subgrade California Bearing
Ratio (CBR) is assumed as 10% as per IRC guidelines. The proposed pavement
combinations were designed using guidelines given in IRC 37. In this study, as shown
in Table 2, minimum thickness permissible for sub-base, base and CRL is considered
for all the trial sections and bituminous pavement layers are kept as variable ones
for perpetual pavement design. It was adopted to provide sufficient stiffness in the
upper pavement layers as per the concept of perpetual pavements. Combinations with
CTSB and CTB were checked against fatigue cracking as per the guidelines in IRC
37.
Effect of Change in the Resilient Modulus of Bituminous … 833

3 High Modulus Bituminous Mix

High Modulus Bituminous mix which was originally originated from France brought
a significant change in the selection of bituminous materials by road construction
and maintenance agencies [5]. This mix requires stiffer bitumen with a penetration
grade of 10/20 or 15/25 along with an aggregate gradation with good interlocking
properties and accommodating the high amount of bitumen, nearly 6% [6]. These
High Performance bituminous mixes are the mixes that would give value of more
than 3000 MPa of resilient modulus 35 °C. The modulus and poisons ratio of different
pavement layers as mentioned in Table 1 are considered for analysis here (Table 2).
Availability of hard binder is a primarily concern issue as most of the refineries
in India as refineries are producing VG-40 binder which is harder grade at present.
Though it is possible to produce much harder by air blowing process, these binders
will lead to premature cracks [7]. On the other hand, hard binder can be obtained
through different techniques: asphalt derived additives and polyolefin addition to
normal binder Selection of modifiers to produce hard binders is based on ability

Table 1 Recommended material properties for structural layers as per IRC guidelines
Material type Elastic/Resilient modulus (MPa) Poisson’s ratio
Bituminous layer with VG40 3000 to 8000 0.35
Cement treated base 5000 0.25
Crack relief layer 450 0.35
Cement treated sub-base 600 0.25
Unbound granular layers 0.2(h*)0.45 MR Subgrade 0.35
Subgrade 17.6 × (CBR)0.64 = 76.82 0.35
* Thickness of sub-base layer in mm

Table 2 Sample table of thickness of pavement layers for different trial combinations and
corresponding pavement responses from IITPAVE (Resilient modulus of bituminous mix considered
as 3000 Mpa)
Trial Sub-base Base CRL/SAMI DBM BC Total 1t 1v
combination (mm) (µs) (µs)
A 200 (GSB) 150 (WMM) 245 50 645 80 167.3
B 200 (CTSB) 100 100 (CRL) 140 50 590 80 172.9
(CTB)
C 200 (CTSB) 100 SAMI 135 50 485 13.64 199
(CTB)
D 200 (GSB) 100 100 (CRL) 150 50 600 80 171.4
(CTB)
E 200 (CTSB) 150 – 210 50 610 78.53 171.9
(Granular
base course)
834 S. Kulkarni and M. Ranadive

Table 3 Pavement thickness calculations for different modulus and pavement combinations
MRm (MPa) Trial combination and corresponding thickness (mm)
A B C D E
DBM BC DBM BC DBM BC DBM BC DBM BC
3000 245 50 140 50 135 50 150 50 210 50
3500 235 50 130 40 125 50 140 50 190 50
4000 220 50 120 50 120 50 130 50 175 50
4500 205 50 110 50 110 50 120 50 165 50
5000 190 50 100 50 105 55 110 50 155 50
5500 180 50 95 50 100 50 105 50 145 50
6000 170 50 90 50 95 50 95 50 135 50
6500 160 50 80 50 95 50 90 50 130 50
7000 155 55 75 50 90 50 85 50 120 50
7500 145 50 70 50 90 50 80 50 115 50
8000 140 50 65 50 85 50 75 50 110 50

of the binder that yield a binder which is less susceptible to temperature and offers
resistance to major distresses and durability. In general the modifiers used to produce
hard binders should enhance the stiffness of the binders that provides higher perma-
nent deformation resistance at higher temperatures as well as resistance to fatigue
cracking at intermediate and low-temperature domains. From the literature it is found
that Ethylene vinyl acetate (EVA), Styrene Butadiene Styrene (SBS), crumb rubber,
hydrated lime, Gilsonite etc., can improve stiffness of bituminous mix [8].The present
study as depicted in Tables 3 and 4 try to evaluate effect of increased modulus in
case of perpetual pavement design ref Fig. 1.

4 Life Cycle Cost Analysis

Life-cycle cost is considered to include the initial cost of construction, cost of reha-
bilitation or reconstruction, and cost of future maintenance regime. In the present
study, LCCA is carried out for 50 years for comparing the overall cost associated
with construction of one-kilometer length of perpetual pavements with 4 lanes i.e. of
14 m width using Net Present Value (NPV) method. The following steps are involved
for the analysis.
1. Determine the initial construction cost.
2. Develop a maintenance plan and find out maintenance cost.
3. Determine life-cycle costs using the NPV method.
Effect of Change in the Resilient Modulus of Bituminous … 835

Table 4 MRm and corresponding pavement thickness for perpetual pavement design
MRm (MPa) Trial Combination
A B C D E
Total Pavement Thickness (mm)
3000 645 590 485 600 610
3500 635 580 475 590 590
4000 620 570 470 580 575
4500 605 560 460 570 565
5000 590 550 455 560 555
5500 580 545 450 555 545
6000 570 540 445 545 535
6500 560 530 445 540 530
7000 555 525 440 535 520
7500 545 520 440 530 515
8000 540 515 435 525 510

675
650
Pavement Thickness (mm)

625
600
575
A
550
B
525
500
C

475 D
450 E
425
400
3000 3500 4000 4500 5000 5500 6000 6500 7000 7500 8000
Modulus of Bituminous Mix

Fig. 1 Pavement thickness comparison

The discount rate has been adopted as 10% whereas inflation rate has been adopted
5%. Based on inputs from field engineers dealing with road construction and officials
working with estimates in Government Departments, it was brought to light that while
estimating the future cost of road construction we should assume that the inflation
in price happen only during 50% of the tenure of design period of the pavement.[9].
According to the recommended guidelines [10], a layer of 25 mm BC is to be provided
once every five years in case of flexible pavements. The cost comparison for a 1 km
stretch of 14 m for was done using the standard rates in Indian Rupees as per the state
government’s public works department schedule of rates for year 2019–20 as shown
836 S. Kulkarni and M. Ranadive

Table 5 Schedule of rates (INR)

Material Specification clause Rate Unit
BC MoRT&H 509 7272 Cubic meter
DBM MoRT&H 507 6742 Cubic meter
SAMI MoRT&H 517 100 Square.meter
CRL MoRT&H 401 1657 Cubic meter
CTSB MoRT&H 404 1827 Cubic meter
CTB MoRT&H 404 1902 Cubic meter
WMM MoRT&H 406 1657 Cubic meter
GSB MoRT&H 401 1598 Cubic meter

in Table 5 [11]. Depending on these parameters, year-wise maintenance schedule

and cost of laying overlay for conventional perpetual pavements have been reported
in Tables 6 and 7. For illustration purpose, sample calculations of economics of

Table 6 Year-wise maintenance schedule and economics for perpetual pavements (millions INR)
Year Maintenance for proposed rigid Cost per K.M Inflation @5.0% per annum NPV
pavements
5th Overlay of 25 mm 2.54 3.24 2.01
10th Overlay of 25 mm 2.54 4.14 1.60
15th Overlay of 25 mm 2.54 5.28 1.26
20th Overlay of 25 mm 2.54 6.74 1.00
25th Overlay of 25 mm 2.54 8.60 0.79
30th Overlay of 25 mm 2.54 10.98 0.63
35th Overlay of 25 mm 2.54 14.01 0.50
40th Overlay of 25 mm 2.54 17.88 0.40
45th Overlay of 25 mm 2.54 22.82 0.31
50th Reconstruction and major repair work in 50th year as per concept of Perpetual Pavement
Total 8.50

Table 7 Sample cost calculation of pavement reconstruction (after 50 years) for perpetual pavement
with MR = 3000 MPa (millions INR)
Perpetual pavement design Initial construction cost Inflation @5.0% per annum NPV
combination
A 36.17 122.48 11.30
B 28.40 96.17 8.88
C 25.63 86.79 8.01
D 28.71 97.22 8.97
E 33.38 113.04 10.43
Effect of Change in the Resilient Modulus of Bituminous … 837

Table 8 Sample life-cycle cost analysis per KM for perpetual pavements with respect to NPV for
period of 50 years with MR =3000 MPa (millions INR)
Type of pavement Initial construction Maintenance Major maintenance Total
(Overlay) (Reconstruction)
A 36.17 8.50 11.30 55.97
B 28.40 8.50 8.88 45.78
C 25.63 8.50 8.01 42.14
D 28.71 8.50 8.97 46.18
E 33.38 8.50 10.43 52.32

pavement reconstruction and life cycle cost analysis are shown in Tables 7 and 8
respectively. Life-Cycle Cost Analysis ref Fig. 2 per KM for all the combinations is
shown in Table 9.
Total Cost in million INR

60.00
55.00
50.00 A
45.00 B
40.00 C
35.00 D
30.00 E
3000 3500 4000 4500 5000 5500 6000 6500 7000 7500 8000
Modulus of Bituminous Mix

Fig. 2 Cost comparison

Table 9 Life-cycle cost analysis per KM for perpetual pavements with respect to NPV for period
of 50 years
Type of Modulus of bituminous mix (MPa)
pavement 3000 3500 4000 4500 5000 5500 6000 6500 7000 7500 8000
Total lifetime cost in INR millions
A 55.97 55.03 53.61 52.20 50.64 49.84 48.90 47.95 47.48 46.39 45.92
B 45.78 44.84 43.89 42.95 42.00 41.53 41.06 40.12 39.64 39.17 38.70
C 42.14 40.88 40.41 39.32 38.69 38.07 37.90 37.90 37.28 37.28 36.66
D 46.18 45.23 44.29 43.35 42.40 41.93 40.99 40.52 40.04 39.50 39.10
E 52.32 50.28 48.87 47.92 46.54 45.59 45.24 44.62 43.67 43.20 42.29
838 S. Kulkarni and M. Ranadive

5 Conclusion

This paper demonstrates comparison between various combinations of conventional

perpetual pavement. This study also shows changes in pavement thickness and overall
pavement thickness occurring due to change in resilient modulus of BC. Tables 2
and 4, 5, 6, 7, 8 and 9 show that combination C provides least cost as well as least
thickness. Table 9 shows that life time cost reduction of 21.88%, 15.45%, 13%,
15.33% and 19.17% is observed in composition A to E respectively by increasing
resilient modulus of bituminous mix from 3000 to 8000 MPa. Similarly, reduction
in thickness by 16.27%, 12.71%, 10.30%, 12.5% and 16.39% is observed in compo-
sition A to E respectively. From the analysis it can be concluded that high modulus
bituminous mix does help to reduce material requirement by reducing the thickness.

References

1. Indian Road Congress (IRC 37) (2018) Guidelines for design of flexible pavement
2. Ranadive MS, Kulkarni SS (2016) Perpetual pavement for rural roads: a concept. In: National
conference on fifteen years of Pmgsy. Transportation engineering group, civil engineering
department, Indian institute of technology Roorkee, India, p 23
3. Walubita LF, Scullion T (2010) Texas perpetual pavements new design guidelines. Texas
department of transportation and the federal highway administration
4. Indian Road Congress (IRC SP 53) Guidelines on use of modified bitumen in road construction
(Second revision)
5. Serfass JP, Bense P, Pellevoisin P (1997) Properties and new developments of high modulus
asphalt concrete. Eighth international conference on asphalt pavements
6. Geng H, Clopotel CS, Bahia HU (2013) Effects of high modulus asphalt binders on performance
of typical asphalt pavement structures. Constr Build Mater 44:207–213
7. Corte JF (2001) Development and uses of hard-grade asphalt and of high-modulus asphalt
mixes in France. Transp Res Circular 503:12–31
8. Goli A (2018) Rheological investigations on high modulus asphalt binders. In: International
conference on pavement and computational approaches
9. Gupta G et al (2018) Life-cycle cost analysis of brick kiln dust stabilized perpetual pave-
ments for lowering greenhouse gas emissions in India. Urbanization challenges in emerging
economies. ASCE
10. Ministry of Road Transport and Highways Government of India (2013) Specification for road
and bridge works
11. Government of Maharashtra (2019–20) PWD schedule of rates
Analysis of Three-Stage Open Pan Heat
Exchanger Working on Dual Fuel
for Jaggery Making

Abhijeet N. Kore and Sanjay S. Lakade

Abstract From ancient times the jaggery production has been done by the same
traditional process. Jaggery industry provides good employment to the village people
in the rural area. But efficiency for the traditional plant is minimum due to the wall heat
losses in the furnace, stack losses through a chimney. The efficiency of the jaggery
plant is very much less looked upon as compared to other worldwide industries.
Different researchers work on these problems to enhance the thermal performance
of the plant and to minimize the extra usage of bagasse. The introduction of a thermic
fluid heater along with an open pan heat exchanger to the replacement of conventional
furnace in the industry, improves production and efficiency. Besides the thermic fluid
heat exchanger, multiple pans are also introduced for a better heat transfer rate. From
the new method developed the total cycle time reduces up to 40% of the total time
in the traditional method.

Keywords Jaggery · Sugarcane · Traditional plant · Modified plant · Thermic

fluid heater

1 Introduction

India is a large producer of jaggery. Jaggery is used in many real-life applications like
Ayurveda and medical fields. Jaggery contains 60–85% of sucrose, 3–10% moisture
and 10–15% of invert sugar and also jaggery consist of vitamins, fibers, protein,
glucose, minerals, content, etc. [1]. Not only jaggery but Sugarcane is also serving
as a green biofuel in a developing country [2]. India is the main merchant’s exporters
of Jaggery throughout the world. India sent out 3,41,120.09 MT of Jaggery and
dessert shop items to the world for the value of Rs. 1633.05 Crore in the year 2019–
20. So it is important to protect this industry by solving the problem associated with
it [3].

A. N. Kore (B) · S. S. Lakade

Department Mechanical Engineering, Pimpri Chinchwad College of Engineering, Pune, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 839
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_84
840 A. N. Kore and S. S. Lakade

Various researchers are carrying out research on conventional plant to enhance

the efficiency of the plant. In India concept of single pan is used in Kolhapur region
and concept of multi pan used in Lucknow region to manufacture jaggery. In current
scenario efficiency of plant ranges from 14 to 20% for a single pan jaggery plant [4–
6] to 48% for multi pan jaggery unit [1, 7]. Multi pan units give more efficiency than
the single pan unit due to effective utilization of combusted gases. Mainly bagasse
is used as fuel in furnace for combustion purpose for the evaporation of water from
the sugarcane juice [1, 5, 6, 8–10]. Most of the energy become waste in the form of
wall heat losses and stack losses. All these parameters are affecting on the efficiency
of the plant. Due to these reasons jaggery entrepreneurs demand additional bagasse
as a fuel which involve extra cost of jaggery manufacturing.
Conventionally the method by which carbohydrate is produced involves the pulp
to supply heat to the pan for evaporation of water from juice. Once the heat is provided
to pans the warmth loss is most, the flue gases are exhausted through the chimney
that may be a total waste of fine offered energy.
Many researchers worked on performance improvement of traditional jaggery
plant with various kinds of heat recovery or combustion technology, but no one has
worked on both to remove wall heat losses, stack losses and also worked on complete
combustion technology. The paper presents the methodology to minimize all these
losses and enrich the livelihood of all jaggery entrepreneurs by increasing the thermal
performance of the plant.
In jaggery processing plant according to the study carries out the maximum effi-
ciency of conventional four pan jaggery manufacturing unit is 48%. Multi pan jaggery
methodology works on the concept of effective utilization of flue gases before leaving
into the atmosphere through the chimney [1, 7].
Process of Jaggery Manufacturing.
In jaggery manufacturing consist of three process i.e., clarification, evaporation, and
crystallization.
Clarification Process: This process is also called as sensible heating process. In this
process juice is heated up to 100 °C. During this process scum present in the juice is
gets removed and clean juice proceeded for next pan.
Evaporation Process: This process as also called as latent heating process. In this
process juice gets heated up to temperature 108 °C. Most of the water present in juice
is gets evaporated during this process. Juice present in this pan at this temperature is
called as liquid jaggery.
Crystallization Process: In this process first juice is transferred from evaporation
pan to crystallization pan where the juice is heated up to striking temperature. i.e.
118–120 °C as per the requirement.
Further concentrated juice is transferred to cooling pit where its temperature is
cooled down up to 60–70 °C then filled in different shapes of mold then it’s cooled
down up to room temperature and ready for packing and dispatching.
This is the generalized process of multi pan jaggery making but the main problem
associated with this process is heat generation. In the conventional method the furnace
Analysis of Three-Stage Open Pan Heat Exchanger Working … 841

is used for the transfer of heat to the juice through indirect contact of flue gasses where
wall heat loss and stack losses are present.
So most of the researchers are working on this problem to minimize all these losses
by designing controlled fire furnace, varying the height of the chimney [11], changing
the material of bricks [12], providing dryers to bagasse but still the efficiency of the
plant is not improved by more than 48%.
Jaggery plant with modifying furnace ordinary masonry bricks by fire bricks at
bottom, by changing chimney cross-section from square to circular that optimized
height of the chimney.Which causes less bagasse consumption & more jaggery
production with a better quality of jaggery. It saves around 208 kg bagasse for 8
quintal production per day [13].

2 Project Identification

As a demand of jaggery is increasing day by day, current jaggery manufacturing

process does not fulfil the market requirement. The major losses are associated with
heat carried in flue gas and wall heat losses hence alternative method is required to
improve its performance. Looking at the aim for increasing the thermal efficiency,
“Open pan heat exchanger based on thermic fluid” will be one of the best alternative
solution that minimize the process time, heat loss and save bagasse consumption.

3 Objective

To minimize the process time of jaggery manufacturing.

4 Methodology

In modified plant furnace used in conventional plant is replaced by the well-insulated

thermic fluid heater, is used to heat a thermic fluid which further used to heat
sugarcane juice using indirect contact of thermic fluid.
Modified plant consist of Thermic Fluid Heater: Thermic fluid heater is used
to heat the fluid with the assistance of a pump. A thermal fluid system is more than
just a heater, and all the components of the system are working together in harmony
for proper performance. The heater is created from iron that is insulated from within
by boiler bricks. Boiler bricks are created up sensible insulation that doesn’t pass
the warmth. Due to the insulation within the heater low heat loss takes place. It is
a type of indirect heating in which thermic fluid in the liquid phase is heated and
circulated through pipes within closed-loop to transfer heat to the sugarcane juice.
842 A. N. Kore and S. S. Lakade

One end of TFH is connected to the pan using pipes through which hot thermic fluid
gets flowing and returns to the TFH through a pump and reservoir.
Thermic Fluid: Thermic fluid is heat transfer oils. These are high-performance
products intended for use in the closed indirect heating system. These oils have very
good heat transfer efficiency. The recent liquid goes through the containers of heat
money handler within which heat vitality is going to be changed. HPCL markets
thermal fluid under the name of HYTHERM.
HYTHERM 600—It is the finest quality oil base stocks and is formulated with
elite extra substances to boost execution at higher temperatures. Ability to provide
maximum heat transfer rate in indirect closed fluid flow system up to 320 °C operating
temperature.
Pan–Pan is one of the important elements in the modified plant. It is used to
collect and heat sugarcane juice indirect way through the thermic fluid. There are
three types of pans conferred i.e., clarification, evaporation and crystallization pans.
This is made by using stainless steel. The warmth thermic fluid handler tubes are
organized deep down of pan. The initial pan is rectangular in size and therefore the
arrangement is limpeting. The rest of the two pans are in a circular cross-section with
the spiral arrangement of tubes are available at the bottom of the pan.
Reservoir–Reservoir is used to accommodate excess volume of hot thermic fluid
when it gets heated from thermic fluid heater which is situated before the pump.
Pump: Pump is used to circulate thermic fluid in closed loop cycle. It is use to
force the thermal fluid into the thermic fluid heater. Some amount of force is needed
to pass the thermal fluid in to heater that is completed by pump as shown in Fig. 1.

5 Construction and Working

In the system the pump pushes the thermic fluid to TFH from the reservoir. In the
thermal fluid heater, the bagasse is burnt in combustion chamber and to heat the
thermic fluid up to 220 °C. The blower is used to enhance the combustion process by
providing a controlled airflow rate to the thermic fluid heater. Complete combustion
provides the utmost energy to the thermal fluid and it saves the additional bagasse
consumption. After heating the thermal fluid within the thermic fluid heater it passes
through the main supply header. This header is further connected to inlet sub supply
header through which fluid is travelled to the bottom of the pan1, pan 2 and pan 3
i.e. clarification pan, evaporation pan and crystallization pan respectively through
the pipe. In the first pan, juice temperature is increased up to 100 °C during this
process, scum present in the juice gets removed and clean juice passed to the pan
2 for further process. In pan 2 temperature of juice is increased up to 108 °C. This
juice is further transferred to pan 3 where the temperature of juice is increased up to
striking temperature i.e. 118–120 °C as per the sugarcane juice quality.
This different temperature of juice is maintained by controlling the mass flow rate
of the thermic fluid which is flowing from the bottom of the pan through the tubes.
Analysis of Three-Stage Open Pan Heat Exchanger Working … 843

Fig. 1 2D modified plant

layout

Inlet and outlet temperature of thermic fluid in the supply header and return header is
maintained constant i.e., 220 °C and 208 °C respectively to avoid thermal imbalance.
This low temperature fluid then return to TFH through main return header and
reservoir as shown in Fig. 1. In second cycle temperature rise for thermic fluid in TFH
required is relatively very less as compared to first cycle so less energy generation
in TFH. Result of this bagasse consumption is relatively very less for this plant.

6 Experimental Procedure

Trials were conducted on a 50 TCD three pan jaggery unit. Daily trials on six batches
per day are taken to measure process time and corresponding temperature and brix of
juice. 1400 lit juice was taken in to pan for every trial. As thermic fluid requires more
time for heating during start of the day, So for every first batch of the day, 120 min to
844 A. N. Kore and S. S. Lakade

130 min were required for completion of process. For later batch it took up to 90 min
to 95 min. Consecutive eleven days trials were conducted on a plant to measure the
process time of jaggery completion and to measure corresponding temperature and
brix of juice. Inlet and outlet temperature of the thermic fluid is maintained at 220–
208 °C during the trial. Digital Refractometer was used to measure the brix of juice
and digital infrared thermometer gun was used to measure the temperature of juice.
Trial was conducted on 1400 lit of sugarcane juice in a rectangular stainless steel
pan1 of 0.5 mm thickness under constant heat input. Heat is provided to sugarcane
juice through hot thermic fluid flowing through hollow pipe attached at the bottom of
the pan1, where temperature of juice is increased up to 100 °C. During this process
scum present in the juice gets separate out and clean juice get transferred in to pan2.
In pan2 evaporation of juice is carried out and maximum amount of water present
in the juice was evaporated and temperature of juice increased up to 108 °C. At the
end relatively concentrated juice gets transferred to pan3 where temperature of juice
increased up to striking temperature of jaggery. For every process in the pan, time
required for its completion and brix of sugarcane juice was measured and compared
with conventional plant.
Figures 2 and 3 depicts that time required for the process completion in conven-
tional plant was around 160 min whereas modified plant need around 95 min. and
it is measured with respect to temperature of juice and brix value at the start and
end of the process of every pan. It shows that to reach striking temperature of juice
conventional plant took more time as compared to modified plant.
For 50 TCD plant average dry bagasse developed after extraction of juice is around
15,000 kg and requirement of bagasse for conventional jaggery making process is
around 18,400 kg, which is more than that of available bagasse as shown in Fig. 4.
So it is necessary to buy additional bagasse to run the plant. Whereas in case of
modified plant it consumes only 7640 kg of bagasse per day. Saved bagasse can be
sold to earn additional revenue for jaggery manufacturer.

140
120
Temperature C

100
O

80 Clarification

60 Evaporator

40 Crystalizer
20 Conventional
0
0 30 60 90 120 150 180
Time in min

Fig. 2 Temperature versus time

Analysis of Three-Stage Open Pan Heat Exchanger Working … 845

Brix Vs Time
90
80
70
Brix in % 60
50
40
30
20
10
0
0 50 100 150 200
Time in min
Conventional Plant Modified Plant

Fig. 3 Brix no versus time

Fig. 4 Comparison of 20000 18400

Bagasse Consumption In Kg

bagasse consumption per day 18000

for modified and 16000
conventional plant 14000
Per Day

12000
10000 7640
8000
6000
4000
2000
0
Modified Conventional
Types of Plant

7 Conclusion

From the trials on 50TCD plant it analyse that:

• Time required for completion of batch in conventional plant was around 160 min
whereas modified plant need around 95 min. So process time for a single batch
of jaggery is reduced up to 40%.
• Bagasse consumption can be saved up to 10,760 kg per day.
• Reduces the pollution in thermal fluid heater plant due to closed-cycle and less
production time.

Acknowledgements Special thanks to Mr. Dhananjay Tekawade, Jaggery Consultant. The exper-
imental work is supported by Cybas Agro Neutraceuticals, Hatvalan Kangaon road, Tal:Daund,
Pune-412215, Maharashtra, India.
846 A. N. Kore and S. S. Lakade

References

1. Sardeshpande VR, Shendage DJ, Pillai IR (2010) Thermal performance evaluation of a four
pan jaggery processing furnace for improvement in energy utilization. Ener 35:4740–4747
2. Adewuyi A (2020) Challenges and prospects of renewable energy in Nigeria: a case of
bioethanol and biodiesel production. Ener Rep 6:77–88
3. APEDA: India export jaggery per year. [Online]. Available: https://agriexchange.apeda.gov.in/
IndExp/PortNew.aspx
4. Rao KSS, Sampathrajan A, Ramjani SA (2003) Efficiency of traditional jaggery making
furnace. Mad Agri 90:184–185
5. Kulkarni SN, Ronge BP (2018) Case study on heat and mass balance of single pan jaggery
plant. Int J Adv Mech Engg 8:143–151
6. Kore A (2018) Unpublished field work at Sangola report
7. Khattak S, Greenough R, Sardeshpande V, Brown N (2018) Exergy analysis of a four pan
jaggery making process. Ener Rep 4:470–477
8. Singh RD, Baboo B, Singh AK, Anwar SI (2009) Performance evaluation of two pan furnace
for jaggery making. J Inst Eng Agric Eng Div 90:27–30
9. Rao P, Das M, Das S (2007) Jaggery—a traditional Indian sweetener. Ind J Tradit Knowl
6:95–102
10. Manjare A, Hole J (2016) Exhaust heat recovery of jaggery making furnace. Int J Sci Res
5:1349–1351
11. Shiralkar KY, Kancharla SK, Shah NG, Mahajani SM (2014) Energy improvements in jaggery
making process. Ener Sust Dev 18:36–48
12. Arya P, Jaiswal UK, Kumar S (2013) Design based improvement in a three pan jaggery making
plant for rural India. Int J Eng Res 2:264–268
13. Madan KH, Jaiswal UK, Kumar JS, Khanna SK (2004) Improvement in gur (jaggery) making
plant for rural areas. J Rur Technol 1:194–196
The Performance and Emission Analysis
of Diesel Engine with Sunflower Biodiesel

Aniruddha Shivram Joshi and S. Ramesh

Abstract The world utilization of petroleum derivatives is expanding quickly and

it influences nature by green house gases causing health risks. Biodiesel is rising
as a significant promising optional energy source which can be utilized to lessen or
even substitute the petroleum utilization. As it is principally produced from vegetable
oils or animal fat that can be produced enormously. Anyway the broad use of the
biofuels can make deficiencies in the food life. This research work analyzes the
Sunflower Methyl Ester (SFME) and its mixtures as a substitute of fuel for any
diesel engine. Biodiesel can be made from sunflower oil in the lab in a little biodiesel
set-up (30Litres) by base transesterification. Four cylinder diesel engine was utilized
for testing on different mixtures of sunflower bio-diesel. The emissions of CO, HC
are lesser than diesel fuel for all mixes experimented. The NOx emission is more
because of the higher volatility and viscosity of bio-diesel.

Keywords Biodiesel · Vegetable oils · Diesel engine · Sunflower oil · Emissions

1 Introduction

Diesel engines are commonly used in many sectors like transportation, marine
engines, power generation & in agricultural due to their better thermal efficiency,
lower fuel consumption. Hence demand for the fossil diesel fuel is increasing rapidly.
The main problem associated with the CI engine is particulate matter (PM) emission
and oxides of nitrogen (NOx). The reduction of PM and NOX is a difficult task [1]
in conventional CI engines. Numbers of researchers have suggested that bio-diesel
as the replacement completely or can be used partially blended with mineral diesel,
because use of bio-diesel reduces the exhaust emissions. Bio-diesel consists more
oxygen, less carbons and sulphur by weight than the mineral diesel [2]. Different
countries such as USA, Germany, etc., were utilizing bio-diesel mixed with diesel.
The neat palm oil not used as an edible oil in Malaysia it is used as the Palm bio-diesel
as alternative to diesel [3, 4].

A. S. Joshi (B) · S. Ramesh

University of Technology, Jaipur, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 847
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_85
848 A. S. Joshi and S. Ramesh

Biodiesel have certain advantages like intimacy to diesel properties, local avail-
ability & renewable in nature, good lubricity, safe to store and lack of difficulty in
transportation, low exhaust emissions [5]. However it is having disadvantages like
cold flow & low heat content. This deficiency can be eliminated by choosing of
proper feedstock material for bio-diesel production which is done by the process of
transesterification and use of proper catalyst in the transesterification process. The
use bio-diesel noted that increase in the emission of NOx [6]. Different conclusions
were made for increase of NOx, like advancement in injection [7], high temperature
of flame [8], change in fuel properties like density, volatility and iodine number [9],
and rapid burning due to availability of oxygen [10]. It is also identified that bio-
diesel is capable to reduce soot emission [11, 12] because of sulphurs & aromatic
compounds are not present [13], and existence of fuel-bonded oxygen [14]. Thus it
is important to study performance, combustion and exhaust emissions of bio-diesel
and embrace as alternative fuel for CI engines.
Bio-diesel effectiveness as a substitute fuel is considered on the basis of three
different characteristics. Combustion is a vital process helpful for testing the viability
of the fuel as well as performance efficiency. The main performance characteristics
are thermal efficiency, specific fuel consumption, torque etc. as per as emission
characteristics are concern, Environmental clearance is a key aspect for selection of
fuel to fulfill environmental regulatory norms which are measured for exhaust gases.
The characteristics are mainly divided into three subsections;
1. Performance
2. Emission and
3. Combustion.

1.1 Performance

As discussed in previous section, the bio-diesel performance is referred to engine

parameters such as brake thermal efficiency (ηbth.), brake specific fuel consumption
(BSFC), mechanical efficiency (ηmech.), torque (τ) etc. The detailed literature review
reveals that investigators used following types of engines for experimentation, 1.
Single cylinder engine 2. Multi cylinder engine 3. Single cylinder VCR engine.
The review article by Tamilselvan et al. [15] provides comprehensive report on
performance, combustion and emissions of bio-diesel fuelled plain bio-diesel and
its blends more than 40 number of bio-diesel fuels were studied for engine type,
test conditions, effect on different parameters and concluded that high BSFC (Brake
Specific Fuel Consumption) and less thermal efficiency than diesel. Bio-diesel and
the blends bio-diesel produce low exhaust emission of HC, CO and PM and higher
emission of NOx & CO2 due to more oxygen in fuel and high Cetane number. Bio-
diesel up to 20% blend may be used as alternate fuel with the little or no engine
modification. The flow chart of various steps involved in production of bio-diesel
along with reactants, different products from vegetable oil is shown in Fig. 1. From
the work of Sharma et al. [16].
The Performance and Emission Analysis of Diesel Engine … 849

Fig.1 Flow chart for

biodiesel production [16]

Efe Şű et al. [17] prepared bio-diesels from five vegetable oils (soybean, canola,
corn, sunflower and hazelnut) by transesterification methanol is utilized as alcohol
and KOH as a catalyst; three stage water washing was did for filtering out contami-
nants from bio-diesel. Uyumaz [18] used transesterification method to prepare bio-
diesel from mustard oil; methanol is utilized as alcohol; reaction temperature was
60 °C methanol mustard oil ratio was 20%. Shrivastava et al. [19] prepared bio-diesels
from two different feedstocks Roselle and Karanja by transesterification method;
methanol is utilized as alcohol, KOH as a catalyst; bio-diesel is washed with pure
water for the removal of methanol content. At last bio-diesel was heated up to 100 °C
for one hour to remove moisture. Mishra et al. [20] produced oil from Simarouba
Glauca by two step transesterification; in first step was acid esterification methanol
is utilized as a alcohol and H2 SO4 as an catalyst; in second step of alkaline ester-
ification methanol and KOH was used as catalyst. Nair et al. [21] produced Neem
bio-diesel by two step transesterification in first step NaOH was used as catalyst; in
second step H2 SO4 is utilized as catalyst in both steps methanol is utilized as alcohol.
Raman et al. [22] derived bio-diesel from rapeseed oil by trans-esterification using
hexane solvent and alkaline catalyst method; NaOH is used as catalyst and methanol
as alcohol.
Goga et al. [23] produced oil from rice bran oil by single stage alkaline transes-
terification process; KOH was used as catalyst and methanol as alcohol. Nathagopal
et al. [24], have prepared Calophyllum inophyllum methyl ester (CIME) from Calo-
phyllum inophyllum oil by three stage transesterification process; first stage was
acid catalyzed Esterification with methanol and concentrated H2 SO4 as catalyst;
second stage esterification KOH was used as an catalyst; third stage was purification,
produced CIME washed with pure (distilled) water followed by heating to remove
moisture. Higher alcohol blends were prepared by addition of higher alcohols (n-
pentanol and n-octanol) on volume basis. Abed et al. [25], have prepared bio-diesel
by transesterification from waste cooking oil (sunflower oil), using methanol and
NaOH as catalyst. Asokan et al. [26], prepared bio-diesel from juliflora oil by two step
transesterification; in acid transesterification, juliflora oil was mixed with methanol
850 A. S. Joshi and S. Ramesh

and sulfuric acid (catalyst), followed by alkali transesterification with methanol and
NaOH as catalyst. From number of studies it is understood that bio-diesel may be
produced from different feedstocks under different conditions and by using different
catalysts.
The literature also reveals that, the different investigators have conducted exper-
iments on different engines using different bio-diesel blend ratio from 5 to 100% by
weight/mass. This makes the difficult to compare the performance. For this reason,
the literature review on engine performance is presented in tabular format for every
one of the engine separately for better understanding. Table 1 illustrate the exper-
imental investigations on the performance and emission of different engines. The
tables describes, author name, engine parameters, fuel blends and interpretation along
with the comments.

2 Methodology

2.1 Biodiesel Production

The biodiesels were derived from alkali catalyzed transesterification process of

sunflower oil. It uses methanol and the 99% pure NaOH. At first, the catalyst was
broken down into methanol and warmed up to 55 °C. The mixture was observed
steadily at 450 rpm. The response time is of about 1 hr. This was again separated
into ester and glycerol layers. The ester layer contains essentially methyl ester and
methanol and the glycerol layer contains basically glycerol and methanol. The blend
is allowing to be settled in 48 h that separates bio-diesel and glycerins. Water is used to
wash the bio-diesel so that it can remove the impurities and heating process is carried
out at 110 °C. This process removes the remnants of the catalyst and methanol. [27]
(Fig. 2).

2.2 Characterization of Biodiesel

The characterization and testing of sunflower made bio-diesel is done for the compli-
ance with Indian standards. Bio-diesel mixtures of B10 (10 bio-diesel, 90% diesel),
B20 (20% bio-diesel, 80% diesel), B30, B40, B00 etc. are set-up in volume premise.
The physio-chemical characteristics of SFME are enlisted in Table 2.
Table 1 The performance and emission analysis of different engines
S. No. Author Title Fuel blend Interpretation Comments
1 ManKee Lam et al. “Cultivation of microalgae Microalgae Microalgae requires less More economica compare to
for bio-diesel production: A land as compare to other other feedstock
review on upstream and feedstock
downstream processing”
2 M. Vijay Kumar et al. “The impacts on Antioxidant and oxygenated Combustion, performance All characteristic are
combustion, performance additives and emissions competitive
and emissions of bio-diesel characteristics are found out
by using additives in direct
injection diesel engine”
3 Babban Yadav et al. “Performance evaluation Microalgae, fossil diesel and Slightly lower power & Compatible wit petro diesel
and emission characteristics Soybean Methyl Ester torque, More SFC, lowered
of microalgae fuel in No2 & PM
combustion engine”
4 Xiaolei Zhanga et al. “The potential of microalgae Microalgae, traditional Review paper More economica compare to
in bio-diesel production” bio-diesel (vegetable oils); other feedstock
5 Rajendra Pawar et al. “A comprehensive review Vegetable oil and animal oil Increase BSFC & BTE and Within th acceptance range
The Performance and Emission Analysis of Diesel Engine …

on influence of biodiesel reduce HC & CO

and additives on
performance and emission
of diesel engine”
6 Muragesh Bellad et al. “Production of biodiesel Algae Comparable Properties ar within ASTM
from algae for alternative Range
fuel as diesel”
7 V. Naresh et al. “Performance, emissions, Algae SEC increases with Oxides of carbo increases, an
sound and combustion increase in Blends nitrogen, hydrocarbons
characteristics of algae oil decreases
biofuel”
(continued)
851
Table 1 (continued)
852

S. No. Author Title Fuel blend Interpretation Comments

8 A. K. Agarwal et al. “Experimental and Soybean, Rapeseed, Cotton All the properties are within Also studied the spray
computational studies on seed, Palm oil, Lard Fatty the acceptable range characteristics
spray, combustion, Acids
performance and emissions
characteristics of biodiesel
fueled engines”
9 J.M. Marchetti et al. “Techno-economic FAEE, FFA, Glycerol, Acidic oil using sulphuric Catalysts increases the
feasibility of producing Triglyceride acid and calcium oxide are performance of engine
biodiesel from acidic oil used as catalysts
using sulphuric acid and
calcium oxide as catalysts”
10 J.M. Marchetti et al. “Economics of biodiesel Biodiesels Review the economics of Production is economical
production: review” biodiesel production
process
A. S. Joshi and S. Ramesh
The Performance and Emission Analysis of Diesel Engine … 853

Fig. 2 Bio-diesel
production scheme

Table 2 The physical and chemical properties of blends of SFME and diesel EN14214
Property Method Specifications Reference B10 B20 B30 B40 B100
diesel
Flash point °C EN ISO 3679 min. 120 68 75 86 97 110 183
Kinematic EN ISO 3104 3.5–5 3.18 3.21 3.26 3.31 3.35 4.7
viscosity
40 °C, mm2/s
Density EN ISO 3675 860–900 832 835 839 843 848 868
15 °C, kg/m
Calorific – – 44.8 43.6 42.2 41.7 40.3 33.5”
value MJ/kg

2.3 Engine Stand

Experimentation was performed on 4-cylinder diesel engine with direct injection.

The specification details of such engine are enlisted in Table 3 and it’s schematic is
shown in Fig. 3.
The motor stand is set-up on a frame which is associated with electronic display
and fuel supply. Testing stand consists of fuel feeding installation, controller and an
equipment for measurement. Mechanical connections among motor and engine is
upheld by a shaft fitted with protection shield. Feeding framework has the part to
guarantee adequate gracefully of fuel to the engine by utilizing a fuel tank which
remains on electronic parity to gauge the fuel utilization. The electric braking process
is the 110 kW 3-phase electric motor and its controlling is done by the inverter and
requested by the cycle PC. This feeding framework comprises of a rack with fuel tank
and has the partition tap. Controller is furnished with the touch screen for controlling
the engine. The precisions of this measurement framework are mentioned in the
Table 4. “BEA350” gas analyzer was utilized to calculate the engine discharges.
854 A. S. Joshi and S. Ramesh

Table 3 Specifications of
Number of cylinders 4
Deutz F4L912 diesel engine
Bore/stroke (mm) 100/120
Displacement (l) 3.770
Compression ratio 19
Maximum rated speed (rpm) 2500
Mean piston speed (m/s) 10
Power (kW) @ 2350 rpm 51
Mean effective pressure (bar) 6.9
Maximum torque (Nm) @ 1450 rpm 238
Minimum idle speed (rpm) 650”

Fig. 3 Schematic diagram of experimental setup

Table 4 Precision of
Parameter Precision value
parameters by measuring
system Speed (rpm) 0…6000 ± 2%
Torque (Nm) 0…1000 ± 2%
Fuel consumption (kg/h) 0…50 ± 2%
Intake air temperature (°C) 0…50 ± 5%
Inlet pressure (kPa) −50…300 ± 5%
Exhaust temperature (°C) 0…800 ± 5”

3 Results

Exhaust Gas Temperature indicates the heat discharge rate of a fuel in combustion
time. The combustion with bio-diesel has been increased because of more oxygen
levels in it. As shown in Fig. 4, that EGT increases with the load for the all mixtures
of SFME. It can be correlated with the research work of Godiganur et al. [28] of
6-cylinder turbo-charged diesel engine with mahua bio-diesel mixtures showed that
EGT increased with the engine load. Datta et al. [29] experimented on 2-cylinder,
The Performance and Emission Analysis of Diesel Engine … 855

Fig. 4 EGT versus load graph

4-stroke, diesel engine with jatropha bio-diesel mixtures showed that EGT increased
with the increased flame temperature of bio-diesel.
Carbon monoxide emission is due to the incomplete fuel combustion. We have
observed that carbon monoxide emission decreases at low-load and it increases
abruptly for all fuels (Fig. 5). The similar results were carried out by Nabi et al. [30]
on a single cylinder with cotton bio-diesel that the Carbon monoxide emissions were
decreased by 24% in comparison with diesel; Çelikten et al. [31] 4-cylinder diesel
engine with soybean Bio-diesel observed a reduction by 28% of carbon monoxide
emissions.
The total combustions in a combustion chamber increases carbon di-oxide emis-
sion. We have obtained that carbon di-oxide emission increases with the loads for the
tested fuels (Fig. 6). Huzayin et al. [32] observed that the carbon di-oxide emission
increases for jojoba bio-diesel mixture of each engine load. Fontaras et al. [33] for
mixtures of soybean the carbon di-oxide emission increases by 14% in the case of
B100 and 9% in the case of B50. Aydin and İlkiliç [34] observed that it is reduced
by 16% of carbon di-oxide emission of rape-seed bio-diesel.
In general the emission of hydrocarbons (HC) depends mainly by compositions
and combustion characteristics of the fuels tested. If combustion is improved the HC
emissions decrease and vice versa. Because of the high content of oxygen in the bio-
diesel it is expected that HC emission will decrease for blends of SFME and diesel. It
is observed that the decrease of HC emissions depends of the percentage of bio-diesel
in the blend (Fig. 7). Sahoo et al. [35] found a reduction in HC emissions by 20.64,
20.73 and 6.75% using bio-diesel of karanja, jathropa and polanga. Tsolakis et al.
856 A. S. Joshi and S. Ramesh

Fig. 5 CO emissions versus load graph

Fig. 6 Carbon di-oxide versus load graph

The Performance and Emission Analysis of Diesel Engine … 857

Fig. 7 HC emissions versus load graph

[36] observed a reduction of nearly 50% for rape-seed bio-diesel compared with low
sulfur diesel.
Nitric oxide (NO) and nitrogen dioxide (NO2 ) are formed in the process of oxida-
tion of the nitrogen during combustion and depend of the combustion temperature
and oxygen content. The tests show an increase in NOx emissions with the increase
in engine load for all blends (Fig. 8). It is found highest for SFME because of
high oxygen content which results in complete combustion causing high combustion
temperature. Gumus [37] found an increase of NOx emissions for apricot seed kernel
oil methyl ester by 10% compared to diesel fuel. Aydin and İlkiliç [34] reported an
increase of NOx emissions 16.7% with B20 blend and 11.8% with B100 for bio-diesel
of rape-seed.

4 Conclusion

The 4-cylinder diesel engine worked successfully while testing with bio-diesels of
Sunflower Methyl Ester and its mixes. These mixtures are examined to obtain the
physio-chemical characteristics. It was observed that the Exhaust Gas Temperature
was more for unadulterated bio-diesel. The more oxygen level increased the temper-
ature of combustion which was resulted in the increase of temperature for the various
mixtures tested. Carbon monoxide and Hydrocarbon emissions were very high for the
858 A. S. Joshi and S. Ramesh

Fig. 8 NOX emissions versus load graph

diesel fuel and low for the mixtures of Sunflower Methyl Ester. It was observed that
the NOX was very high for pure Sunflower Methyl Ester and it’s mixtures because
of high volatility, low heat content and high viscosity as compared to the diesel fuel.

References

1. Yamaki Y et al (1994) “Application of common rail fuel injection system to a heavy duty diesel
engine”. SAE paper 942294
2. Xue J et al (2011) Effect of biodiesel on engine performances and emissions. Renew Sustain
Energy Rev 15(2):1098–116
3. Sharma YC, Singh B (2009) Development of biodiesel: current scenario. Renew Sustain Energy
Rev 13:1646–51
4. Palit S et al (2011) Environmental impact of using biodiesel as fuel in transportation: a review.
Int J Global Warming 3(3):232–256
5. Misra RD, Murthy MS (2011) Blending of additives with biodiesels to improve the cold flow
properties combustion and emission performance in a compression ignition engine—a review.
Renew Sustain Energy Rev 15:2413–2422
6. Sun J et al (2010) Oxides of nitrogen emissions from biodiesel—fuelled diesel engines. Prog
Energy Combust Sci 36:677–95
7. Thangaraja J et al (2014) Experimental investigations on increase in biodiesel—NO emission
and its mitigation. Proc Inst Mech Eng, Part D: J Automobile Eng 228:1274–1284
8. Weiss B et al (2007) A numerical investigation in to the anomalous slight NOx increase when
burning biodiesel; a new (old) theory. Fuel Process Technol 88:659–667
9. Canakci M (2009) NOx emissions of biodiesel as an alternative diesel fuel. Int J Veh Des
50:213–228
The Performance and Emission Analysis of Diesel Engine … 859

10. Xue J et al (2011) Effect of biodiesel on engine performances and emissions. Renew Sustain
Energy Rev 15:1098–116
11. Ogunkoya D et al (2016) Investigation of the effects of renewable diesel fuels on engine
performance, combustion, and emissions. Fuel 140:541–554
12. Suh HK, Lee CS (2016) A review on atomization and exhaust emissions of a biodiesel fueled
compression ignition engine. Renew Sustain Energy Rev 58:1601–20
13. Yunuskhan TM et al (2014) Recent scenario and technologies to utilize non-edible oils for
biodiesel production. Renew Sustain Energy Rev 37:840–51
14. Boehman A et al (2003) Fuel formulation effects on diesel fuel injection, combustion, emissions
and emission control. In: Proceedings of the DOE diesel engine emissions reduction conference,
newport, RI, August 24–28, 2003. US: DEER, pp 1–9
15. Ayoola AA, Anawe PAL, Ojewumi ME, Amaraibi RJ (2016) Comparison of the properties
of palm oil and palm kerneloil biodiesel in relation to the degree of unsaturation of their oil
feedstocks. Int J Appl Nat Sci 5(3):1–8
16. Ghadge SV, Raheman H (2006) Process optimization for biodiesel production from mahua
(Madhuca indica) oil using response surface methodology. Biores Technol 97:379–384
17. Saleh HE (2009) Experimental study on diesel engine nitrogen oxide reduction running with
jojoba methyl ester by exhaust gas recirculation. Fuel 88:1357–1364
18. Singh S, Prabhahar M Experimental investigation of performance and emission characteristics
using chlorella algae bio diesel as alternative fuel
19. Efe Ş et al (2018) Comparative engine characteristics of biodiesels from hazelnut, corn,
soybean, canola and sunflower oils on di diesel engine. Renewable energy, (Accepted
manuscript)
20. Uyumaz A (2018) Combustion, performance and emission characteristics of a DI diesel engine
fuelled with mustard oil biodiesel fuel blends at different engine loads. Fuel 212:256–267
21. Rathore Y et al (2019) Experimental investigation of performance characteristics of compres-
sion—ignition engine with biodiesel blends of Jatropha oil & coconut oil at fixed compression
ratio. Heliyon 5:e02717
22. Shrivastava P et al (2019) An experimental evaluation of engine performance and emission
characteristics of CI enine operated with roselle and Karanja biodiesel. Fuel 254:115652
23. Mishra SR et al (2018) Impact of simarouba glauca biodiesel blends as a fuel on the performance
and emission analysis in an unmodified DICI engine. Renew Energy Focus 26
24. Raman LA et al (2019) Experimental investigation on performance, combustion and emission
analysis of a direct injection diesel engine fuelled with rapeseed oil biodiesel. Fuel 246:69–74
25. Goga G et al (2019) Performance and emission characteristics of diesel engine fueled with rice
bran biodiesel and n-butanol. Energy Rep 5:78–83
26. Soni T, Gaikwad A Waste pyrolysis tire oil as alternative fuel for diesel engines
27. Tutunea D, Bica M, Dumitru I, Dima A (2014) Experimental researches on biodiesel prop-
erties. SMAT 2014 3rd international congress science and management of automotive and
transportation engineering 23rd–25rd of October, Craiova, Tome I, ISBN 978-606-14-0864-1,
Universitaria, pp 295–300
28. Godiganur S, Murthy CHS, Reddy RP (2009) 6BTA 5.9 G2–1 Cummins engine performance
and emission tests using methyl ester mahua (Madhuca indica ) oil/diesel blends. Renew Energy
34:2172–2177
29. Datta A, Palit S, Mandal BK (2014) An experimental study on the performance and emission
characteristics of a CI engine fuelled with Jatropha biodiesel and its blends with diesel. J Mech
Sci Technol 28(5):1961–1966
30. Nabi MN, Rahman MM, Akhter MS (2009) Biodiesel from cotton seed oil and its effect on
engine performance and exhaust emissions. Appl Therm Eng 29:2265–2270
31. Çelikten I, Koca A, Arslan MA (2010) Comparison of performance and emissions of diesel
fuel, rapeseed and soybean oil methyl esters injected at different pressures. Renew Energy
35:814–820
32. Huzayyin AS, Bawady AH, Rady MA, Dawood A (2004) Experimental evaluation of diesel
engine performance and emission using blends of jojoba oil and diesel fuel. Energy Convers
Manag 45:2093–2112
860 A. S. Joshi and S. Ramesh

33. Fontaras G, Karavalakis G, Kousoulidou M, Tzamkiozis T, Ntziachristos L, Bakeas E, Stournas

S, Samaras Z (2009) Effects of biodiesel on passenger car fuel consumption, regulated and non-
regulated pollutant emissions over legislated and real-world driving cycles. Fuel 88:1608–1617
34. Aydin H, İlkiliç C (2011) Exhaust emissions of a CI engine operated with biodiesel from
rapeseed oil. Energy Sour Part A 33:1523–31
35. Sahoo PK, Das LM, Babu MKG, Arora P, Singh VP, Kumar NR, Varyani TS (2009) Compar-
ative evaluation of performance and emission characteristics of jatropha, karanja and polanga
based biodiesel as fuel in a tractor engine. Fuel 88:1698–1707
36. Tsolakis A, Megaritis A, Wyszynski ML, Theinnoi K (2007) Engine performance and emissions
of a diesel engine operating on diesel-RME (rapeseed methyl ester) blends with EGR (exhaust
gas recirculation). Energy 32:2072–80
37. Gumus M, Kasifoglu S (2010) Performance and emission evaluation of a compression ignition
engine using a biodiesel (apricot seed kernel oil methyl ester) and its blends with diesel fuel.
Biomass Bioenergy 34:134–9
Deployable Environment or Healthcare
Technologies
Experimental Analysis of Effect
of Bio-lubricant Between Tribological
Systems of Piston Ring Under
the Jatropha Oil

Mhetre Rahul Sanjay and L. B. Abhang

Abstract In an Engineering scenario, it is necessary to increase the efficiency and

life of every component in the machine. A particular vehicle is selected based on
its working ability to increase the efficiency and life span. In this project, I have
chosen the piston ring material with the Jatropha as lubricating oil to increase the
efficiency and the life span of the engine. The life span and working ability are
governed by the friction and wear characteristics of components of the engine of the
vehicle. The life span of any vehicle is calculated by the life of its sensitive part that
is susceptible to wear. In tribological system “piston -ring-liner” observed 45–55%
of frictional mechanical losses. To reduce such losses, lubrication performs a vital
role. Lubricant is the main part of lubrication to maintain reliable functions of the
machine, provide smooth operations and chances of failure is less. Crude oil is a
vital source to manufacture lubricant. But in today’s condition, the prices of crude
oil increasing and depletion of crude oil reserves in the world, and global concerning
protecting the environment from pollution have renewed interesting developing and
using environment-friendly lubricants derived from alternative sources.

Keywords Design methodology · Experimental analysis · Result · Conclusion

1 Introduction

In any mechanical systems, friction and wear loss is one of many factors in energy
consumption. To decrease friction and wear, oil lubricant has been studied to be used
as lubricant additives that have promising effects on friction and wear reduction in
automotive. Oil lubricant of many compositions and sizes have demonstrated certain
degrees of friction modifying and anti-wear effects. We recently concluded that in
the boundary lubrication region, the addition of oil lubricant could reduce the friction
coefficient up to 65%, and wear use as high as 74%. Such lubricants combining with
a base oil and dispersed Jatropha oil lubricant emerged as a new class of lubricants,
the bottleneck for further development, however, is the aggregation of Jatropha oil

M. R. Sanjay (B) · L. B. Abhang

Vishwabharati Academy’s College of Engineering, Ahmednagar, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 863
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_86
864 M. R. Sanjay and L. B. Abhang

lubricant in a base oil. A stable suspension of Jatropha oil lubricant is essential for
a usable lubricant.
Selected Jatropha oil lubricant is found better among all added into the oil. Among
those that were added into oils, Jatropha oil lubricant has received much attention
and exhibited excellent applications for their excellent friction and wear properties
[1]. The reduction of wear depends on interfacial conditions such as normal load,
geometry, relative surface motion, sliding speed, surface roughness, lubrication, and
vibration. Chemical additives in lubrication fluid control anti-wear properties, load-
carrying capacities, and friction under the specified boundary lubrication conditions.

2 Important Considerations While Jatropha Oil Lubricant

Selection

Designing of Jatropha oil lubricant Selection lubricant additives Preparation depends

upon so many factors. The factors were analyzed to get design inputs for Jatropha
oil lubricant Selection. Following are factors for the selection of lubricant additives.
(a) Study of lubricant oil and finished quantity size and property.
(b) Capacity and of the lubrication oil, its limits of automation.
(c) Provision of lubrication preparation devices in the machine.
(d) Available Jatropha oil lubricant and their property.
(e) Accuracy.
(f) Evaluation of variability in the performance results of the preparation [2].

3 Problem Statement

1. Bio-lubricant easy available and more cooling property, good frication charac-
teristic.
2. Less surface area and good lubrication property.
3. It may cause friction due to high temp or increase the cycle time of replacement
of part.
4. Chances of slippage are more while higher applied force chance to an accident
occurs.
5. Increase life of Piston Ring.

4 Necessity of Work

The principal motivation for formulating new additives using Jatropha oil lubricant
as a promising solution for improving the Tribological behaviour is that Jatropha oil
lubricant has the potential to offer significant tribological benefits of both solid and
Experimental Analysis of Effect of Bio-lubricant Between... 865

liquid lubrication and extend the life of the mechanical components. The decline
of the friction coefficient between the rings/liner assemblies plays a critical role
in improving engine performance and fuel efficiency. Bio lubricant additives play
a significant role in the formation of a tribofilm layer on the worn surfaces via a
physical or chemical absorbed mechanism to enhance protection of the worn surfaces
and create a rolling effect between sliding surfaces.

5 Objective of Work

Experimental workout to find Friction and wear characterization.

Experimental activity to find the Coefficient of friction.
Experimental workout to find Lubricants temperature.
Experimental workout to find Viscosity Experimental work for testing weight loss
of Polymide piston ring under the Jatropha oil and SAE 15 W 40.

6 Design Methodology

6.1 Piston Ring and Its Properties

A piston ring is one of the essential parts of the Diesel/Petrol engines. It is an open-
ended ring that fits into a groove on the outer diameter of a piston in a reciprocating
engine such as an internal combustion engine or steam engine [3]. The principal
function of the piston rings is to form a seal between the combustion chamber and
the crankcase of the engine. The goal is to prevent combustion gases from passing
into the crankcase and oil from passing into the combustion chamber the three main
functions of piston rings in reciprocating engines are:
1. Sealing the combustion/expansion chamber.
2. Supporting heat transfer from the piston to the cylinder wall.
3. Regulating engine oil consumption [4].

6.2 Ring Material

The materials used to make Piston Rings are one of the most critical factors in its
Performance. Thus we have selected Cast Iron bronze and nylon material to make
piston Ring.
1. Cast Iron: Cast iron is an iron-carbon alloy with a carbon content great than
2%. As the most common ring material, cast iron has a low—friction, wear
866 M. R. Sanjay and L. B. Abhang

resistance, characteristics used in a broad range of high temperature and high-

speed engine application. It produces Consistence presence against the bore.
Cast iron retains the integrity of its original shape under heat, Load and other
dynamic forces [5].
2. Polyamide (Nylon6): Polyamide exhibit property of low friction and no lubri-
cation property. Nylon is quiet in operation, resists abrasion, wears at a low
rate, and easily moulded cast, or machined to close tolerance. Nylon is very
inexpensive [6].

6.3 Lubricant and Its Properties

The primary purposes of lubrication are

(1) To reduce wear and heat loss that results from the contact of surfaces in motion,
that is, to reduce the coefficient of friction between two contacting surfaces;
(2) To prevent rust and reduce oxidation;
(3) To act as an insulator in transformer applications;
(4) To act as a seal against dirt, dust, and water [7].

6.4 Experimental

6.4.1 The Cygnus Friction and Wear Test Machine Configuration

The Cygnus friction and wear test machine are used to evaluate the friction and wear
characteristics of lubrication in this experiment. The Cygnus friction and wear test
machine are designed to study friction and wear in dry or lubricated sliding over a
wide range of speed, load and temperature. It is a tri-pin-on-disc machine which is
conducted by using three pins on a disc as testing specimens. Specifications of the
Cygnus friction and test machine is mentioned in Table 1 [3].
The Cygnus friction and wear test machine are connected with a computer having
embedded a block diagram based application construction program, Visual Designer.
Visual Designer allows developing custom data acquisition, analysis, display in the

Table 1 Specifications of the

Parameter Value
Cygnus friction and test
machine Test disc diameter 110.0 mm
Test pin diameter 6.0 mm
Test disc speed range 25 to 3000 rpm
Motor Tuscan; (2000 rpm, 1.5 kW)
Load range 0 kg to 30 kg
Electrical input 220 V AC 50 Hz
Experimental Analysis of Effect of Bio-lubricant Between... 867

form of a graphical chart or numerical meter. The programs can be controlled simply
by drawing a program’s data structure in the form of a block diagram. The block
diagram can also contain textual comments, allowing the process being monitored or
controlled to be documented. The block diagram can also contain textual comments,
allowing the process being monitored or controlled to be documented [3].

7 Results and Discussion

7.1 Cygnus Friction and Wear Testing Machine Result

Analysis

With the help of Cygnus Friction and wear testing Machine I have taken the following
reading.

7.1.1 Friction and Wear Characterization

Figure 1 Show the curves of pins wear as a function of sliding time for various
Jatropha oil blended with lubricant SAE 40. The values of linear pin wear under
2000 rpm and 30 N loads for each pin vary from 0.02 to 0.05 mm. It was observed
that the higher or maximum wear occurred at the beginning of the experiment for
some of the test specimens. It is clear from the graph that maximum wear occurred
for JBL40 and for JBL10 wear is minimum. We also can observe from the graphs
that except JBL40, for each JBL, pin wear decreases gradually and continuously. At
the beginning of the test, we can see the wear rate was fast in the period that is called
the running-in period.

Chart Title
Wear (mm)

0.1

0.05

0
1 2 3 4 5 6 7 8 9 10 11 12
Sliding Time (m/MIn)
0% of jatropha 10% of jatropha 20% of jatropha
30% of jatropha 40% of jatropha 50% of jatropha

Fig. 1 Linear pin wear as a function of sliding time for various bio-lubricants
868 M. R. Sanjay and L. B. Abhang

Chart Title
0.35
0.3
0.25
0.2
COF

0.15
0.1
0.05
0
1 2 3 4 5 6 7 8 9 10 11 12
Sliding Time (m/MIn)
0% of jatropha 10% of jatropha
20% of jatropha 30% of jatropha
40% of jatropha 50% of jatropha

Fig. 2 Coefficient of friction as a function of sliding time for various bio-lubricants

7.1.2 Coefficient of Friction

Figure 2 shows the curves of friction coefficient plotted against the sliding time
for various Jatropha oil-based bio-lubricants. The results in Fig. 2 depict that the
lubricant regimes that occurred during the experiment were the boundary lubrication
where the value of friction coefficient (µ) for boundary lubricant is in the range of
0.001–0.2 except for JBL50. For 0% of Jatropha oil-based bio-lubricants, it can be
seen that the coefficient of friction is highest at the beginning, and then it falls rapidly
up to minimum value compared to all samples. This phenomenon can be explained
by attributing oxide layer on the aluminium surface. Lubricant additives react with an
oxide layer, which causes to form a thick tribo-film on the aluminium surface. At the
initial stage, the shear stress of the tribo-film is high. Thus, the coefficient of friction
is high early. With continuing the sliding, the aluminium pin is eroded, and a new
metal surface is exposed. Since the fresh metal surface has a lower tendency to react
with lubricant additives, the formation rate of tribo-film decreases than the beginning.
Therefore, shear stress becomes lower than beginning and constant throughout the
process.

7.1.3 Lubricants Temperature

Figure 3 Shows the relation of the averages oil temperature of varies percentage
of Jatropha oil-based bio-lubricants with the sliding time, respectively. The range
of temperature is about 20–100 °C. From the graphs, we can see that lubricant
temperature increases with increasing sliding time for each percentage of Jatropha
oil-based bio-lubricants. The maximum temperature rise occurred for JBL40, and
minimum temperature rise occurred for JBL 10. The progressive increase in temper-
ature period is known as the running-in period during the asperities of the sliding
surface are progressively cut off. From the graphs, it can be noticed that the 30 and
Experimental Analysis of Effect of Bio-lubricant Between... 869

Chart Title
60
40
Temp 20
0
1 2 3 4 5 6 7 8 9 10 11 12
Sliding Time (m/MIn)
0% of jatropha 10% of jatropha 20% of jatropha
30% of jatropha 40% of jatropha 50% of jatropha

Fig. 3 The bio-lubricant temperature as a function of sliding time

40% of Jatropha oil Based bio-lubricant are produced more heat than other while
10% of Jatropha oil-based bio-lubricants is generated lower heat.

7.1.4 Viscosity

Figure 4 show that the viscosity of Jatropha oil-based bio-lubricants decreases expo-
nentially for both 40 °C and 100 °C operation temperatures. The Table 3 shows the
viscosity grade requirement for the lubricants set by the International standard orga-
nization (ISO) while Fig. 4 shows the behaviour of the viscosity of Jatropha oil-based
bio-lubricants at 40 °C and 100 °C operation temperatures. Comparing Table 2 and
Fig. 4, it can be stated that at 40 °C operation temperature 10, 20 and 30% of Jatropha
oil-based bio-lubricants meet ISO VG100 requirement.
But 40 and 50% of Jatropha oil-based bio-lubricants do not meet the criteria.
However, it can also be noted that all bio-lubricants have a higher viscosity than the
expected value.

Chart Title
200
150
Viscosity

100
50
0
120% 100% 80% 60% 40% 20%
Axis Title
40 C 100 C

Fig. 4 The viscosity of various percentages of Mineral Oil for temperature 40 °C and 100 °C
operations
870 M. R. Sanjay and L. B. Abhang

Table 2 Experimental layout for wear test under sae15w40 Oil

Experiment no Test load Sliding velocity Quantity of oil Weight loss of piston ring (mg)
(N) (m/min) (ml) Experimental Predicted value
value
1 30 18 30 0.0053 0.0052625
2 30 18 10 0.0056 0.0055875
3 30 15 20 0.0045 0.0046333
4 30 15 20 0.0046 0.0046333
5 30 12 30 0.0040 0.0040125
6 40 12 20 0.0041 0.0041125
7 40 18 20 0.0054 0.005462
8 20 15 10 0.0050 0.005025

7.1.5 Experimental Work for Testing Weight Loss of Polymide Piston

Ring Under the Jatropha Oil and SAE 15 W 40

The accuracy of the model was developed by substituting the experimental data in
the model equation at various conditions of test load, sliding velocity and quantity of
oil and the corresponding predicted values of weight loss of piston ring both under
SAE 15W40 and Jatropha oil were obtained in the following Tables 2 and 3.
The Table 2 shows that the predicted values for the piston ring weight loss and
the actual values of weight loss of piston ring were quietly closed with each other
under SAE15W40 oil. This validated the credibility of the model developed for
establishing a correlation between the process variables and eight loss of piston ring
under SAE15W40 oil.
The Table 3 shows that the predicted values for the piston ring weight loss and
the actual values of weight loss of piston ring were close with each other under

Table 3 Experimental layout for wear test under jatropha oil

Experiment no Test load Sliding velocity Quantity of oil Weight loss of piston ring (mg)
(N) (m/min) (ml) Experimental Predicted value
value
1 30 18 30 0.0048 0.0049
2 30 18 10 0.0051 0.00505
3 30 15 20 0.0041 0.00416
4 30 15 20 0.0042 0.00416
5 30 12 30 0.0034 0.00345
6 40 12 20 0.0036 0.0037
7 40 18 20 0.0049 0.00495
8 20 15 10 0.0045 0.00465
Experimental Analysis of Effect of Bio-lubricant Between... 871

Jatropha oil, thus, validating the credibility of the model developed for establishing a
correlation between the process variables and eight loss of piston ring under Jatropha
oil.
From the experimental layout for both wears under SAE15W40 oil and Jatropha
oil, it was concluded that the combinations of process parameters such as test load
of 60 N, sliding speed of 1200 cm/min and 30 ml of oil yields a low value of weight
loss of piston ring. But Jatropha oil yields a low value of weight loss of piston ring
as 0.0034 mg when compared to the SAE15W40 oil which yields the weight loss of
the piston ring as 0.004 mg at the same process parameter combinations.

8 Conclusion

Based on the above experiment, the following conclusions can be summarized:

1. The rate of wear for a various percentage of Jatropha oil-based bio-lubricants
was different. However, the rate of wear for 10% and 20% of Jatropha oil-based
bio-lubricants are near to the pure lubricant
2. In this experiment, the temperature of lubricating oil increases with sliding
increasing time for each percentage of heat compared to the other samples.
3. In this experiment, it has been found that having lower wear resistance bio-
lubricant contains a higher coefficient of friction.
4. Since Jatropha oil-based bio-lubricants have a higher coefficient of friction
compared to pure lubricant; it can be assumed that the fatty acid molecules
available in Jatropha oil do not build a soap film on a surface test.
5. For each experiment Cast Iron, Bronze and Polymide content increase because
of wear occur in pin and disc.
6. In term of viscosity, except JBL40 and JBL50, all bio-lubricants meet the ISO
VG100 requirements.
7. In weight loss experiment the loss of weight of polyamide piston ring with
different load and quantity of Jatropha oil hence we observed that the loss
of weight is less in polyamide piston ring with Jatropha oil compared to the
polyamide piston ring with SAE 15 W40 oil.

References

1. Adhvaryu A, Liu Z, Erhan SZ (2005) Synthesis of novel alkoxylated triacylglycerols and their
lubricant base oil properties. Indus Crops Prod 21(1):113–119
2. Shahabuddin M et al (2012) Tribological characteristics of amine phosphate and acety-
lated/butylated diphenylamine additives infused bio lubricant. Energy Educ Sci Technol Part
A 30:89–102
3. Shahabuddin M et al (2013) Comparative tribological investigation of bio-lubricant formulated
from a non-edible oil source (Jatropha oil). Indus Crops Prod 47:323–330
872 M. R. Sanjay and L. B. Abhang

4. Siniawski MT, SN, Adhikari B, Doezema LA (2007) Influence of fatty acid composition on the
tribological performance of two vegetable-based lubricants. J Synthetic Lubricant 24:101–110
5. Wolff A (2014) Simulation-based study of the system piston-ring-cylinder of a marine two-stroke
engine. Tribol Trans 57:653–667
6. Arumugam S, Sriram G (2012) Effect of bio-lubricant and biodiesel-contaminated lubricant on
tribological behaviour of cylinder liner-piston ring combination. Tribol Trans 55:438–445
7. Baker CE, Theodossiades S, Rahnejat H (2012) Influence of in-plane dynamics of thin
compression rings on friction in internal combustion engines. J Eng Gas Turb Power
Garbage Monitoring and Collection
System Using RFID Technology

Amol A. Kadam, AksahyAjadhav, Dhanraj P. Narsale, Anil M. Kasture,

S. M. Karve, and Manoj A. Deshmukh

Abstract Now each day we face the matter of garbage which is scattered everywhere
in buildings also as in villages the Gathering of that garbage isn’t done on time. So,
we’ve proposed the system of garbage pickup & Monitoring. This system, it’s alleged
to be collect the bins from every registered home. The worker should collect Bin from
the house and scan the RFID tag with the RFID Reader. The RFID Reader is going to
be interfaced with the Wi-Fi module. After scanning of RFID tag the message "Bin
collected" are going to be delivered to the respective customer. If any home is going
to be missed for collection of Bin, then the actual customer can call to the customer
care with the given toll-free number. After interaction with care, the message “Bin
collection is remaining is sent to the nearby bin collector worker.

Keywords Ultrasonic sensors · RFID Reader and tag · GSM · Garbage bin

1 Introduction

As we observe the present days, the waste is spread everywhere and overflows it
from the bins kept at the important areas. These areas are often railway stations,
our homes or schools Itcreatesbadhealthconditionforthepeoplethatlivenearthedust-
binareas, to avoid such a situation, we are getting to design bins collecting system
means it contains frequency identification (RFID) CARD reader. Garbage pickup
System minimizes and improves the efficiency by including costs per km, per hour
to transfer the waste after collection from various locations of the town or villages.
Recent days, many of us live in cities just for their simple earning money all together
with the ways, and lots of people are coming from the agricultural areas for the oppor-
tunities which are largely found in urban areas. Thanks to that massive crowd of indi-
viduals, there’s a considerable amount of garbage is produced a day. By observing
all the issues were getting to propose a system which can collect the waste from

A. A. Kadam (B) · AksahyAjadhav · D. P. Narsale · A. M. Kasture · S. M. Karve ·

M. A. Deshmukh
SVERIs College of Engineering Pandharpur, Pandharpur, Maharashtra, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 873
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_87
874 A. A. Kadam et al.

every home which is registered already. The system works faithfully and doesn’t
skip collecting garbage from any home.
Advantage:
• It saves money as the overall budget of the project in a smaller amount thanks to
the database.
• RFID avoids the disadvantage of barcode scanning, which needs line-of-sight
access to every barcode, and also it scans just one item at a time.
• RFID eliminates the likelihood of errors which are created by humans.
• The RFID reader can scan the number of tags.
• RFID is often integrated with scanning, and readers are fixed.

2 Literature Survey

Samudhana Satyamanikantha makes the utilization of sensors like IR sensor, weight

sensor, photoelectric sensor and frequency identification (RFID) CARD reader.
Thereupon the load sensor below the Bin and therefore, the percentage of garbage
inside the Bin is calculated. Also, the message “Thanks for cleaning” are going to
be given to the user [1]. Dr. Sathish Kumar, B. Vijayalakshmi, R. Jenniferprarthana,
A. Shankar had used the RFID tag. That’s a very tiny device that stores and forwards
the info to an RFID reader. They’re characterized in two types’ active tag and passive
tag [4]. Nalawa di Srikantha, Khaja Moinuddin, Lokesh K S3, Aswatha Narayan had
made the utilization of models that adopts RFID and also capacity, weight sensor,
humidity and chemical sensors IC. With the assistance of that, they’re getting to clean
to wash the areas in big cities [2]. Vinayak Baradi, Devesh Rawool, Rutvji Desh-
putre, Hemanshu Ghadigaonkar ESP8266 maybe a Wi-fi module which can give this
project access to the internet. It’s cheap, but it made their project very powerful [3].
Fetulhak Abdurahman, Silesi Aweke, Chera Assifa had implemented the rubbish
monitor using sensor and Arduino Microcontroller. The small print of those dustbins
is monitored by the Municipal authorities with the assistance of GUI [5]. Norfadzlia
MohdYusof, Aiman Zakwan Jidin and Muhammad Izzat Rahim had used some ultra-
sonic sensors, the Arduino Uno Microcontroller and therefore the GSM Module. The
ultrasonic sensors are wont to give the extent of garbage in every binand it’llsendin
formation to ArduinoUnowhichis that the system controller [6]. Kesthara V., Nissar
Khan, Praveen S. P., Mahesha C., Murali N. are the author who made the utilization
of sensors which can senses the quantity of garbage also as level of garbage in the
dustbin and instant messages to the trash management system. Their main objective
is to separate the waste in several bins [7]. Sirichai Watana sophonand Sarinee Ouitra
kulma dearobo to which uses the PIC18f4550. They used the roboton beach area. The
robot cleans the beach area for that the instructions are given through the program
developed from Visual Basic application [8]. Suchit S. Purohit, Vinod M. Both-
alehave used RFID, GPS, GIS, GSM. That they had attached the tags oncontainer
and also attached readers on track. They used GPS for location tracking and GSM
Garbage Monitoring and Collection System Using RFID Technology 875

for wireless transmission [9]. P. Ramchandar Rao, S. Sanjay Kumar, Ch. Rajendra
Prasad They used Garbage monitoring system it’s an improvement of normal dustbin
by changing it to be smart using sensors. Garbage monitoring system may be an idea
of which makes a traditional dustbin smart using ultrasonic sensors for garbage
level detection. It display sends the message to the priority department person to
update the status of the Bin using GSM model [10]. The workers are carrying RFID
Tags with them. The RFID Reader is attached at every home.The RFID Reader is
interfaced with the Wi-Fi module. At the initial stage of the Gathering, the worker
will collect the waste from a particular home. He will scan the RFID Tag with the
RFID reader.The reader is going to be connected to the Wi-Fi module, which gives
knowledge about the collection of Bin. After that, the message “Bin collected at
your home” is shipped to the customer. If the Gathering of Bin isn’t wiped out the
correct way or the Gathering of bins refused, there’ll be an alternate solution for
the particular problem. The customer is provided with a toll-free number to register
their complaints. If things occurred like Bin is skipped or not collected properly,
they could register their complaints through the toll-free number. The GPS system
is employed to trace the method. If any complaint is registered by the customer, the
monitoring system will send a message to the nearby worker, and he will follow an
equivalent flow for collection and Monitoring of the waste. The method is going to
be continued until the last home’s Bin is going to be collected. The Fig. 1 show the
flow chart of the proposed system to manage collection and Monitoring of registered
customers bin from their home.

Fig. 1 Flow chart of

Garbage collection and
Monitoring System using
RFID
876 A. A. Kadam et al.

3 Methodology

In the whole process of “Garbage Collection and Monitoring System,” we are using
the pairs of RFID Reader and RFID tag. Therein we’ve homes which are registered.
The workers are carrying RFID Tags with them. The RFID Reader is attached at
every home. The RFID Reader is interfaced with a Wi-Fi module. At the initial stage
of the Gathering, the worker will collect the waste from a particular home. He will
scan the RFID Tag with the RFID Reader. There are going to be connected to the
Wi-Fi module, which gives knowledge about the collection of Bin. After that, the
message “Bin collected at your home” is shipped to the customer. If the Gathering of
Bin hasn’t wiped out a correct way or the Gathering of Bin is refused, there’ll be an
alternate solution for the particular problem. The customer is provided with a toll-
free number to register their complaints. If things occurred like Bin is skipped or not
collected properly, they could register their complaints through the toll-free number.
The GPS system is employed to trace the method. If any complaint is registered by
the customer, the monitoring system will send a message to the nearby worker and
he will follow an equivalent flow for collection and Monitoring of the waste. The
method is going to be continued until the last home’s Bin is going to be collected.
The fig no 3 show the flowchart of the proposed system to manage collection and
Monitoring of registered customers bin from their home.

4 Conclusion

This paper concludes that by using this smart garbage monitoring system using RFID
over IOT’s we can easily dispose the waste present within the garbage bins as early
as possible in comparison to the previous methods without it affecting to the people
and keep the environment clean.

References

1. Smith TF, Waterman MS (1981) Identification of common molecular subsequences. J Mol Biol
147:195–197. Samudhana Satyamanikantha (July-2017). Smart garbage monitoring system
using sensors with RFID over the internet of things. Journal of Advanced Research in Dynamical
and Control Systems (JARDCS), SaveethaUniversity,Thandalam, Chennai,TamilNadu, India.
2. Nanavati S, Moinuddin K, Lokesh KS, Narayana A (2017) Waste managementin IoT-enabled
smart cities: a survey. Int J Eng Comput Sci ISSN: 2319–7242(IJECS)
3. Baradi V, Rawool D, Deshputre R, Ghadigaonkar H (2018). Intelligent garbage collection using
IoT. Int J Innov Adv Comput Sci IJIACS ISSN2347–8616
4. Dr. Sathish kumar N, Vijayalakshmi B IOT based smart garbage alert system using
arduinoUNO.2016IEEE region10 conference (TENCON) proceedings of the international
conference
5. Abdurahman F, Aweke S, Assefa C (2016) Automated garbage monitoring system using arduino
IOSR J Comput Eng (IOSR-JCE)
Garbage Monitoring and Collection System Using RFID Technology 877

6. Yusof NM, Jidin AZ, Rahim MI (2017) Smart garbage monitoring system for waste
management. MATEC web of conferences 9701098 (2017)
7. Kesthara V, Khan N, Praveen SP, Mahesha C, Murali N (2015) Sensor based smart dustbin for
waste segregation and status alert. Int J Latest Technol Eng Manage Appl Sci (IJLTEMAS)
8. Sirichai Watanas ophonand Sarinee Ouitrakul (2014) Garbage collection roboton the beach
using wireless communications, 2014 3rd international conference on informatics, environ-
ment, energy and applications
9. SoukaynaMouatadid ZFA (2011) RFID Based Solid Waste Collection Process, International
Conference in Institute of Electrical and ElectronicsEngineers(IEEE)
10. Ramchandar Rao P, Sanjay Kumar S, Rajendra Prasad Ch (2017) Garbage monitoring system
using arduino. Int J Trend Sci Res Dev (IJTSRD)
A Survey on Mental Health Monitoring
System Via Social Media Data Using
Deep Learning Framework

Satyaki Banerjee and Nuzhat F. Shaikh

Abstract In today’s society Depression, Stress and Gloominess are some of the
most broadly perceived and increasing mental issue influencing us. The presence
of a system that is automatically capable of identifying a users mental state is of
great benefit. Due to users spending a lot of time on social media using that to
check his well-being will be helpful in many ways. There are various algorithms
such as Random Forest, SVM, ANN, CNN, RNN present using which this can be
achieved. Sentiment Analysis and deep learning techniques could provide us robust
algorithms and structure for a target also a chance for observing mental issues which
are, specifically of depression and stress. In this paper different ways of dealing
with depression shown on social media platform are studied. This will enable in
achievement of better understanding of the various mechanisms used in depression
detection.

Keywords E-health · Stress and depression · Sentiment analysis · Social media

1 Introduction

Social media is talking up a lot of time of humans so it is nowadays becoming the

richest source of human generated text inputs. Feedbacks, critiques, opinions and
views which are provided by the social media users will reflect thoughts, personal
attitudes and sentiments towards various topics. There exists a knowledge-based
system, that will include an emotional health monitoring process which is used to
detect users with a likely psychological disorder specially like depression and stress
[1, 2]. Various platforms may be studied to uncover symptoms. In many a situation,
extraction of behavior on online social platform offers a substantial opportunity to
undeviatingly identify mental illness at even a very premature stage [3]. The most
common and most disabling mental disorders are depression, stress and prolonged
sadness, which also has a relevant impact on our current society [3].

S. Banerjee (B) · N. F. Shaikh

Department of Computer Engineering, M.E.S. College of Engineering, Pune, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 879
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_88
880 S. Banerjee and N. F. Shaikh

At present, various methods for depression and stress detection do exist some of
the diagnosis rely on self-reporting linked with a informed assessment of health care
practitioners [4]. The presence of an automatic system will not only help the various
health care professionals but it will also provide a cost efficient and immediate if not
exact diagnosis. Sentiment and deep learning or machine learning technologies on
social media platforms could help to tackle such objectives. Even though the aim of
these systems is not outright replacing any professional they surely do support their
work and give immediate aid. There are various other methods of identification of
a depressed user but nowadays users spend a lot of time accessing social media so
if there is a mechanism that will utilize this data it will cause in immediate benefit
[5, 6]. The organization of the paper is as follows: Sect. 2 gives the related work
and limitations and last section concludes the paper with future work followed by
references.

2 Related Work

2.1 Convolution Neural Network, Recursive Neural Network,

Sentiment Analysis

User’s feelings about various topics may be understood using social media data.
Along these lines, systems, such as, checking and suggestion frameworks (RS) can
gather and dissect this information [1]. There are systems that can identify a user’s
stress with the help of his social media posts. Different deep learning algorithms
such as CNN, RNN, LSTM-RNN prove to be very useful in such cases [1]. The
main advantage considered here is informing some of the authorized persons about
the depression detected this will result in taking steps which might be necessary
to help the stressed and depressed user and might save them from ending up to
take any drastic steps. Also motivational messages are sent to depressed users thus
cheering them up and relaxing them the system also has a very easy to use interface
that proves to help the users in navigating and performing the needed function [1].
The recommendation system consisted of 360 different messages of relaxing, happy,
motivational etc. categories. The sentiment analysis here comprises of building a
sentiment metric the sentiment score consists of a range of −5 to +5. The users
data from their Facebook is collected and this data is used for identification of the
depressed user. Different ontological networks are considered from user data such
as their activity, sleep patterns etc. The use of CNN and BILSTM is done for the
depression identification here along with sentiment analysis inclusion thus providing
with a more accurate system [1].
A Survey on Mental Health Monitoring System … 881

2.2 Combined Feature Models, Adaptive Boosting, Support

Vector Machine

In social media networks there may be emojis too which have also been considered in
some of the works. Work has been done mainly on social media sites like Facebook
and Twitter. There is work done on other platforms like reddit too which have longer
texts and may thus give more insights about the user [7]. The dataset used by them
comprised of 1293 depressed and 548 standard tweets. The subreddits available for
depression help and other subreddits involving friends and family are included in it for
the dataset building. Total of 68 different features could be taken here. Extensive work
for obtaining the feature set was done and various methods have been considered. For
n-gram modeling unigram and bigram models have been considered, for linguistic
dimensions LIWC and for Topic modeling LDA is considered. Most related words
for different categories like job, depression, tired, friend etc. are also identified. Next,
for the purpose of classification many algorithms have been used here classifiers like
Logistic Regression, SVM, Adaboost, MLP are analyzed. The main finding here
were words related to a depressed users messages are, preoccupation with oneself,
sadness, hostile, anxiety, anger or suicidal thoughts, with a most emphasis on the
present and future [7]. For single feature case SVM using bigram model provided the
best result. It has been observed that MLP using combined features models presented
the best solution when compared to other combinations such as unigram models and
Adaboost algorithms [7]. The work provided a way to get relation between user’s
language and him facing depression. Also the combination of n-gram, LDA and
LIWC was studied and also their impact individually was considered too [7].

2.3 Sentiment Scores

With the help of social media we may identify negative words which ultimately result
in the formation of depressive words. Twitter data is considered here, the dataset built
is a large one. Data collection is over a span of 3 months and the target was kept as
Australian elections. Next the collected data cleaning and frequently occurring word
removal process is carried out. Various unigram models may be prepared to carry
out this process [2]. Here the words are given scores and using this identification
of the depressed words using the scores and hence identification of the depressed
user is done [2]. The tweets are all rated, the rating is done based on its strength.
There are two scales, the positive range consisting of +1 to +5 and the second being
the negative range varying from −1 to −5 [2]. Such methods helps to provide a
simple yet efficient mechanism for the monitoring of a users mental health. Out of
all users 13 depressed users with lowest scores of sentiment were identified. For these
users further data analysis was carried out, LDA method could separate tweets into
different types. Then the negative sentiments identification and standard deviation
representation of the same was studied. Human behavior is not constant happiness
882 S. Banerjee and N. F. Shaikh

and sadness varies always for normal users, but for a stressed user this may not be the
behavior. A stressed user’s standard deviation could indicate a continuous sadness
state. Hence words which when used in Tweets give a more probability of user being
depressed could be identified. Here, however the detailed analysis on the positive
sentiments has not been carried out [2].

2.4 Random Forest, eXtreme Gradient Boosting, Support

Vector Machine

Using Depression levels can influence the probability of taking drastic steps such as
committing suicide out of sheer stress. Such unfortunate incidents can be prevented
if early detection of such issues can be done. This can be achieved by presenting
students as well as parents with questionnaires which may include questions like:
if they are regular to school, if they have mood swings also their age groups, their
appetite, sleeping habits etc. [3]. The set of students were between age group of
15–29 and there were many factors that caused in the presence of depression in their
life. The various causes for it included failure, bullying, pressure from parents etc.
The option to fill the answers are from value of ‘0’ to ‘3’ where ‘0’ denotes not at
all and ‘3’ represents always. The data collection process is performed using API
for reddit and standard data set for Twitter. There are multiple subreddits in reddit
that have a lot of user interaction and engagement. The data responses considered
here are around 619 out of this almost 18 different features could be identified. The
feature extraction is based on Tf-idf and using the questionnaires data points as
reference. Now the data from the questionnaires can be treated to obtain features
like age, regularity etc. which is given to classifier, along with this their Twitter data
is also monitored to detect the presence of stress or depression in Tweets [3]. The
data is used for both training and testing purpose. Machine learning algorithms such
as Random Forest, XGBoost, SVM are considered for the analysis. The maximum
accuracy was obtained with the help of XGBoost here. As per the data obtained
around 48% of people faced hopelessness and around 30% students felt bad about
themselves [3].

2.5 Natural Language Processing, Support Vector Machine,

Naive Bayes

Social media websites are the place where users are nowadays expressing their
emotions the most and in the most honest manner [5] Facebook and Twitter are
the main sites where users express themselves and also connect. There has been
work done on data based on these sites, using machine learning algorithms the users
may be classified into: normal, mild, severe or extremely depressed [5]. The gap
A Survey on Mental Health Monitoring System … 883

between expression on social media site and actual thoughts may also be fulfilled by
building multiple choice questions that will help to understand and detect depressed
users [5]. Both Facebook and Twitter data are collected, then various algorithms are
used for the purpose of depression detection. The step after getting the data using tool
is using Natural Language Toolkit in order to get a clean version of it. The further
analysis deals with the presence of the formation of a datasplit in order to get the
train data. The system is considered on 100 twitter user, their data generation on one
week is considered for the analysis process. The system considered single user based
on the user name and their individual result of being depressed or not is carried out.
The system gives access to the users friends and family to know about the persons
mentals health. The model also may be implemented by the social media sites itself
to provide help to their own users. SVM, Naive Bayes and NLP techniques are used
for the identification of depression [5].

2.6 Various Machine Learning Algorithms

If the depressed user is motivated and made to feel better his complete potential
will be utilized and that person will not take any drastic life threatening steps too.
As per requirement there may be different approaches of data collection, it can be
questionnaires asked to the person, the posts they put on social media, or even text
used in verbal communication as well as expressions on face [8]. The facial detection
may be done for a user it will but depend on many factors such as the lighting or
the angle of getting image. Social Media data however depends on many lesser
parameters. Hence this will give a simpler way to collect the data and will have many
lesser dependencies to be considered also there is no requirement of any additional
resource here thus the cost is also minimized. Also data collection from Twitter may
be done using tool like Twint, set of positive and negative Tweets may be collected
by this. A sample of data collected with the help of this is presented and given as
reference. Various machine learning algorithms such as SVM, Decision Trees, Naive
Bayes Classifier, KNN Classifier are considered for the detection of a depressed user
[8]. The usage of deep learning algorithms has been suggested it will be useful for
accuracy improvement [8].

2.7 Metadata, Convolution Neural Network

The diagnosis of depression is in-fact a major issue, across the world people do
not always have access to facilities which are readily available for their access [4].
With the help of twitter data here a system has been provided, the twitter data is
used to construct graphs with relation between depression and symptoms. Statistics
and NLP is used to identify presence of depression in the tweets [4]. The social
media platform redddit is considered a total data of 135 users is used. Different
884 S. Banerjee and N. F. Shaikh

word embedding were checked such as, fastText, GloVe, word2vec also multiple
dimensions were considered. The closest neighbours for the word ‘depression’ is
identified and we can also obtain ‘insomnia’, ‘suicidal’ etc. are the terms that have the
closest association with ‘depression’. The sentence formation itself is a process that
is of a lot of depth. There are different parts of a sentence which have their associated
meanings and the way in which they are present are also having a significant impact
on the meaning. Some of the adjectives presence may be having imapact on the
system for example ‘very’ will add more weight to the sentence and will make it
more positive or more negative depending on its context of usage. A deep learning
mechanism with the use of CNN proves to be very successful in extracting deep
features and obtaining depressed users from the given set of users [4]. Different
ERDE models are considered here and CNN models provide a good result with
metadata association but the collection of metadata is a difficult task on its own [4].

2.8 Natural Language Processing and Statistics

Doctors usually work with patients via questioning to detect depression however
this data is not available for common public. So depression symptoms are not very
clearly disclosed [6]. With the help of twitter data here a system has been provided,
the twitter data is used to construct graphs with relation between depression and
symptoms. Statistics and NLP is used to identify presence of depression in the tweets.
Initially in the data collection phase a total of 120 depressed tweets were collected
in a span of around 15 days. The study of depression based network generation here
considered total of 64 thousand tweets after all cleanup activities. Work in this area
to obtain a relationship between depression and various symptoms associated with it
is done in detail [6]. The words associated with depression are thus identified, words
like anxiety, insomnia, hate have the maximum similarity in context of depression
detection [6]. The words are identified for their frequencies and also their degree
which implies to their number of occurrences. The word score may be utilized in
this process. Word2Vec mechanism is implemented and also a semantic graph with
the center as “depression” is prepared and other words weights are calculated with
reference to it [6]. There are algorithms that are there to identify the similar words with
respect to depression and they are merged together, next their statistical distribution
is found out. The merge of medical system in a social media system is the motive
here, the medical systems will be benefited if this system is utilized in them [6].
A Survey on Mental Health Monitoring System … 885

2.9 Discrete Wavelet Transform with Artificial Neural

Networks

Social media tools devices are wide spread however there may be other mechanisms
which can be used for depression detection. Humans Electroencephalography signals
can also provide a mechanism to provide an understanding of depressed vs non
depressed users [9]. It is important to monitor stress and depression as it also has
an impact on a persons general physical health. A stressed user finds it difficult to
carry on with his daily activities at ease. The study conducted comprised of people
in the age group of range 18 years to 30 years. All regular users were considered
and smokers were excluded as they consider it as a coping mechanism of stress
already. The volunteers are asked to be in different states and their EEG signals are
considered. Simple actions like read watch video are used to capture the reactions.
Also stress and depression may be a factor to cause other issues to, issues like blood
pressure increase and other lifestyle disorders may be caused due to the existence on
stress and depression in a person [9]. The EEG signals of the users may be utilized
with Wavelet transforms in order to obtain a system that is capable to identify if a
user is depressed or not [9]. The features are extracted with the help of DWT, a total
of 5 level decomposition too are used. ANNs are further trained to classify efficiently
the depressed user. The process of back propagation is carried out based on which
the values of next weight is kept on being updated every time. The main issue is the
cost and availability of EEG capture procedure in this research. Also simple human
behaviors like eye blinking or hand movements too cause in error in value of EEG
recording [9].

2.10 Convolution Neural Network

Stress and depression can also be identified with the help of analyzing a users inter-
actions on the social media sites [10]. Various user level and Tweet level attributes are
considered and they are used for analysis. The data is provided to a CNN which helps
in identification of the depressed users [10]. Here not just own tweet was considered
but also comments that a user puts on others posts are considered for analysis [10].
The features considered are based on both user level as well as tweet level aspects
here real world huge dataset is considered for analysis purpose. The use of the users
Tweet content as well as the social interaction factors are considered for the analysis.
The different features considered may be positive negative words or emoticons also
use of adjectives can be studied. The different social attention parameters which
were considered included their comments and likes. The users Social Tie analysis
is performed too and what they like or retweet are also considered along with his
own posts [10]. A factor graph model is provided along with CNN in order to obtain
various relationships amongst the user behaviours. CNN model gives a good result
when applied and provides a good accuracy. The social structure of the friends of
886 S. Banerjee and N. F. Shaikh

the stressed users appeared to be less connected when compared to the non stressed
users social structure. Also stressed users had more association to anger and sadness
where as very less association to family, friend and leisure when compared to a non
stressed user [10].

2.11 Statistical Models

An alternative way to identify stress may be by analyzing a user’s daily phone usage
activities, few more parameters may be used to enhance the process. Parameters
such as weather data, personality traits may be considered [11]. User spends his
most time with his phone so this analysis should provide us a good platform to
understand the current users state of mind [11]. Statistical Models are prepared from
the various features that are collected and the models are studied with reference to
the impact it has based on other parameters such as current weather in the given area,
the personality traits of the user. The user may have permanent personality issues
which also influences the results [11]. A multiple factor based system that is capable
for the identification of stress finding could be achieved. A total of 111 subjects were
considered and their seven months data was taken into account for the work, they
belonged to different backgrounds and had different lifestyles. The data collection is
done based on many ways including self reporting, call and sms logs and proximity
data using bluetooth data. Obtaining daily stress from a users smart phone usage
activity provides a good way for stress monitoring, it is essential to identify stress
and help the users take necessary help to overcome their issue. The non availability
of proximity data is an issue in this case, it might not be easy to obtain it and use it
[11].

3 Open Issues

Mental health is a very important aspect of well-being of humans hence its a field
of extensive research area. In the above section, few of the approaches which have
been implemented in many situations in order to achieve the same purpose have been
mentioned. Many different algorithms have been studied which provide depression
detection in social media platform. In the existing works mostly data mining and
machine learning algorithms have been used and major work is done some aspects
of user behaviour, sentiment analysis may also be included in a more complete
manner. User’s personality type may also be considered in analysis, more features
related to occupation etc. may be included too. Also more number of social media
platforms may be included for getting a better understanding of a users behavior.
A Survey on Mental Health Monitoring System … 887

4 Conclusion

In this survey, identification of online users with depression and stress that is threat-
ening to people’s mental well-being is studied. Social media account privacy is a
concern in many platforms. It has been observed that Facebook and Twitter are the
social media platforms that have most works done on. This also identifies that in
many situations deep learning’s performance is better than other algorithms. It has
also been seen that in many cases CNN, RNN have better accuracy compared to
algorithms like random forest.

References

1. Rosa RL, Schwartz GM, Ruggiero WV, Rodríguez DZ (2019) A knowledge-based recom-
mendation system that includes sentiment analysis and deep learning. IEEE Trans Indus Inf
15(4):2124–2135
2. Tao X, Dharmalingam R, Zhang J, Zhou X, Li L, Gururajan R (2019) Twitter analysis for
depression on social networks based on sentiment and stress. IEEE Int Conf Behav Econ
Socio-Cult Comput 6:1–4
3. Jain S, Narayan SP, Dewang RK, Bhartiya U, Meena N, Kumar V (2019) A machine learning
based depression analysis and suicidal ideation detection system using questionnaires and
twitter. In IEEE students conference on engineering and systems, pp 1–6
4. Trotzek M, Koitka S, Friedrich CM (2020) Utilizing neural networks and linguistic metadata
for early detection of depression indications in text sequences. IEEE Trans Knowl Data Eng
32(3):588–601
5. Al Asad N, Pranto MAM, Afreen S, Islam MM (2019) Depression detection by analyzing
social media posts of user. In IEEE international conference on signal processing, information,
communication & systems, pp 13–17
6. Long Ma, Wang Y (2019) Constructing a semantic graph with depression symptoms extraction
from twitter. IEEE Conf Comput Intell Bioinf Comput Biol, pp 1–9
7. Tadesse MM, Lin H, Bo X, Yang L (2019) Detection of depression-related posts in reddit social
media forum. IEEE Access 7:44883–44893
8. Narayanrao PV, Kumari PLS (2020) Analysis of machine learning algorithms for predicting
depression. Int Conf Comput Sci Eng Appl, pp 1–4
9. Berbano AEU, Pengson HNV, Razon CGV, Tungcul KCG, Prado Seigfred V (2017) Classi-
fication of stress into emotional, mental, physical and no stress using electroencephalogram
signal analysis. IEEE Int Conf Sig Image Proc Appl, pp 11–14
10. Huijie Lin J, Jia JQ, Zhang Y, Shen G, Xie L, Tang J, Feng L, Chua TS (2017) Detecting stress
based on social interactions in social networks. IEEE Trans Knowl Data Eng 29(9):1820–1833
11. Bogomolov A, Lepri B, Ferron M, Pianesi F, Pentland AS (2014) Daily stress recognition from
mobile phone data, weather conditions and individual traits. Proc ACM Int Conf Multimedia
22(14):477–486
An Approach to Heart Disease Prediction
with Machine Learning Using Chatbot

Chinmay Nanaware, Arnav Deshmukh, Nikhil Chougala, and Jaydeep Patil

Abstract Cardiovascular diseases are one of the leading causes of mortality and
morbidity worldwide which is a major concern to be dealt with, but it is difficult to
identify cardiovascular diseases in the earlier stages because of several contributory
factors which posses risk such as high blood pressure, high cholesterol, diabetes, and
many other factors. Prediction of cardiovascular diseases is considered as one of the
important aspects of healthcare analysis given a large amount of raw data available,
waiting to be converted into valuable information. On the other hand, the majority
of internet users are adopting one or more messenger platforms which enable us to
deploy AI-based conversational bots that can be developed to cater to individual users
to provide user satisfaction, high level of accessibility and flexibility. The aim of this
paper is to demonstrate the use of Machine Learning models using Google Cloud
Platform to assist users in the prediction of cardiovascular diseases using a chatbot
with the help of Flutter framework on Android and other messenger platforms such
as Telegram and Facebook.

Keywords BigQuery · Cardiovascular disease · Chatbot · Flutter · Google cloud

platform · Machine learning

1 Introduction

Cardiovascular Disease(CVD) is a blanket term used to describe all conditions

affecting the heart and blood vessels which include but not limited to coronary
heart disease and cerebrovascular disease. CVDs are the leading cause of the death
globally representing more than 31% of global death annually estimated to be 17.9
million people in the year 2016, out of which 85% are due to heart attack and
stroke [1]. People who are at high risk of heart diseases due to several contributing
risk factors such as diabetes, hypertension, high cholesterol levels, smoking, obesity
need early detection and management steps as appropriate. Although age is a known
risk factor for the development of heart diseases, autopsy evidence suggests that the

C. Nanaware · A. Deshmukh (B) · N. Chougala · J. Patil

IT Department, AISSMS Institute of Information Technology, Pune, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 889
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_89
890 C. Nanaware et al.

CVD development process in later years is not inescapable [2]. Detection of CVD
mostly requires a clinical diagnosis but chances of incorrect diagnosis and treatment
are probable. Additionally, not all tests contribute towards effective diagnosis, the
majority rely on doctors expertise and experience. With the availability of abundant
raw data in the healthcare sector and machine learning advancing at an astounding
speed, we can use the data to support doctors in leveraging the diagnosis accuracy
with the help of machine learning techniques. Supervised Machine Learning model
takes input data and trains the model over the data with the help of the target namely
the label we want to predict thus resulting in a model that attempts to predict the
desired label. With the progress of NLP, and the internet users adopting one or more
messenger platforms, it increased the usage of chatbots, especially the ones which are
integrated within the messenger platform. A chatbot is a computer program with AI
capabilities that interacts with the user through text or audio. Chatbots are an evolu-
tion of question answering systems which can be effectively used in any domain that
is feasible to interact for a quick response. Chatbots offer responses faster than other
means of communication and are of higher quality, it is a manifestation of customer
ease. Chatbots are trained to generate responses to users queries and this is the most
important aspect to train bot further to adapt to end-users and their preference when it
comes to designing personalized chatbot. Chatbots mainly classify user requests into
intents and entities to give an appropriate answer to specific questions asked. Intents
are used to classify exactly what the user is asking and to process the request as per
model trained on intents. Entities are used used to process the request after under-
standing the intent of user such as what or where related to which things or person
the user is asking about [3]. Multiple platforms allow us building chatbots using their
NLP base to train custom bots such as Dialogflow, a service offered by Google which
runs on Google Cloud Platform. It is a natural language processing (NLP) platform
which can be used to build conversational applications and provide experiences for a
company’s customers in various languages and on diverse platforms [4]. Dialogflow
is used to develop different digital assistants which act as a conversational agent to
identify dialog states and respond with the corresponding state which is stored in the
dialog tree as to how the conversation flows between user and agent. Additionally,
as the complexity of big data increases, smaller enterprises struggle to accommodate
the storage and computing needs required for training and deployment of machine
learning models. This is where public cloud providers come into play. Cloud makes
it easy for such companies to experiment with various machine learning capabili-
ties and scale up as projects are pushed to production and demand increases. Cloud
offers elasticity and multiple services to set an infrastructure without having to invest
heavily and create cloud-native applications. Cloud providers such as Google Cloud,
AWS, Microsoft Azure provide many options for implementing intelligent features
in enterprise applications that don’t need deep knowledge of machine learning or
AI theory. BigQuery, a service by Google Cloud Platform is a fully managed and
serverless data warehouse that enables cost-effective, scalable and fast analysis over
petabytes of data. It is a Software as a Service(SaaS) that offers to query using
ANSI SQL. BigQuery has built-in machine learning capabilities that help to train
ML models with good accuracy using ANSI SQL. BigQuery can also be used to
An Approach to Heart Disease Prediction with Machine … 891

train and test models as it is a flexible ML platform that is easy to rely on due to SQL
related queries.
This research paper dedicates section II for system design for the proposed method,
section III deeply elaborates the proposed techniques and whereas section IV eval-
uates the performance of the system and finally section V concludes the paper with
traces of future enhancement.

2 System Design

Figure 1 shows system design and explains how different software and technologies
are implemented in this project.

Fig. 1 Implemented system

892 C. Nanaware et al.

3 Proposed Methodology

The goal of the prediction methodology is to design a machine learning model that
classifies whether patients have chances of heart disease or not, and deploy the model
on the Android/iOS platform with the help of chatbot using Flutter Framework and
Telegram. The dataset that is used to train the model is the Cleveland Heart Disease
Dataset made available to download through UCI Repository which contains a total
of 76 attributes but only a subset of 14 attributes is used for the prediction. The
search was in the main direction to the causes or the main factors which have a
strong influence on the heart. Some factors are constant such as sex, age, and family
background. Among many parameters, there are some varying parameters like blood
pressure, heart rate, cholesterol, chest pain that are considered mainly in the dataset.
Figure 2 shows features which are important and contribute in prediction.
The next step is data discretization and preprocessing techniques in the form of
Data Transformation, Data cleaning, Data Reduction. After Preprocessing the data
was split into training and testing dataset and was applied to algorithms such as
Decision Tree, Logistic Regression and KNN. The model accuracy was compared to

Fig. 2 Feature importance

An Approach to Heart Disease Prediction with Machine … 893

determine the best performing model which is later used to establish communication
with Chatbot. To generate the model we used BigQuery, a Google Cloud service that
enables users to create and execute various machine learning models in BigQuery
using standard SQL queries. Dialogflow, another service offered by Google was used
to create a design of the conversational agent. Dialogflow We integrated our chatbot
using Dialogflow fulfilment with Cloud Function so that it can interact with our model
in BigQuery.Using Dialogflow we deployed the chatbot on Telegram with the help of
Telegram API Token generated by Botfather which is used to set up Telegram bots.
Further, the chatbot was integrated with our Flutter application using Dart language
and Dialogflow API with the help of flutter_dialogflow package in Flutter. Flutter
authenticates the users with the help of Firebase Authentication, enabling users to
log in using Google, thus providing ease of access.

4 Results

The proposed methodology has been implemented using Google Cloud Platform,
Dialogflow and Flutter. To deploy the most accurate classifier we compared the
accuracy of Logistic Regression, Decision Tree and KNN algorithms.It was found
that out of Logistic Regression, KNN and Decision Tree algorithms, Logistic Regres-
sion performed the best with an accuracy of 85.14% while Decision Tree and KNN
performed with an accuracy of 74.72% and 84.61% respectively. Table 1 shows the
comparison between Logistic Regression, Decision Tree and KNN algorithms.
We can describe the performance of the classifier using a confusion matrix that
allows the visualization of the performance of an algorithm. In Table 2 the number

Table 1 Predictive analysis of classifiers

Evaluation criteria Classifiers
Logistic regression Decision tree KNN
Time to build a model (in seconds) 1.932 1.154 1.762
Correctly classified instances/total instances 258/303 69/91 77/91
Predictive accuracy 85.1485 74.7252 84.6153

Table 2 Confusion matrix

Class 1 predicted Class 2 predicted
Class 1 actual TP FN
Class 2 actual FP TN
894 C. Nanaware et al.

of incorrect and correct predictions are summarized together with their respective
count values and broken down by each class. It gives us insight not only into errors
being made by a classifier but more importantly the types of errors that are being
made by the classifier.
Classification rate/Accuracy:
Classification rate or Accuracy of a classifier is given by the following relation:

TP + TN
Accuracy =
TP + TN + FP + FN

Confusion Matrix of Logistic Regression.

= = = Confusion Matrix = = =
A B
106 32 |A = TRUE
13 152 |B = FALSE
Confusion Matrix of Decision Tree.
= = = Confusion Matrix = = =
A B
33 5 |A = TRUE
17 36 |B = FALSE
Confusion Matrix of KNN.
= = = Confusion Matrix = = =
A B
41 5 |A = TRUE
9 362 |B = FALSE
The best performing classifier i.e. the Logistic Regression is integrated with the
chatbot using Cloud Functions on Telegram and Flutter Application. Figure 3 illus-
trates the interaction between the user and chatbot integrated within the Flutter Appli-
cation we developed and Fig. 4 shows telegram bot which can be accessed through
Telegram Application.

5 Conclusion

Heart Disease is one of the major concerns for society today due to its high mortality
rate. The complexity of manually detecting the heart disease can be leveraged using
Machine Learning algorithms. In this paper, we discussed the accuracy of Decision
An Approach to Heart Disease Prediction with Machine … 895

Fig. 3 Flutter interface

Tree, KNN and Logistic Regression. The accuracy of the Decision Tree is 74.72%,
KNN is 84.61% and Logistic Regression is 85.14%. The Logistic Regression clas-
sifier performed with the highest accuracy than suggested in previous papers hence
we suggest Logistic Regression be used for further prediction of diseases. Further,
this system also can be used in the form of ensembles i.e. combinations of multiple
techniques. This would increase further accuracy and high performance of the system.
896 C. Nanaware et al.

Fig. 4 Telegram interface

References

1. https://www.who.int/en/news-room/fact-sheets/detail/cardiovascular-diseases-(cvds)
2. Karvonen MJ (1988) Prevention of cardiovascular disease among the elderly. Bull World Health
Organ 66(1):7–14
3. Liu B, Xu Z, Sun C, Wang B, Wang X, Wong DF, Zhang M (2018) content-oriented User
Modeling for Personalized Response Ranking in Chatbots. IEEE/ACM Trans Audio Speech
Lang Process 26(1):3–9
4. Gupta AK, Kumar P, Joshi S (2020) Conversational agent dialog flow interface. United States
patent application publication, US2020/0004874 A1,2-6, Jan 2
Automated Early Detection of Diabetic
Retinopathy

Supriya Shegdar, Ameya Bhatlavande, Dhanashree Patil,

and Sanjivani Kadam

Abstract Diabetic Retinopathy is a common retinal complication associated with

diabetes. Early detection of Diabetic Retinopathy shields patients from losing their
vision. Thus this paper proposes an automated method for image-based classification
of diabetic retinopathy. The method is divided into three stages: image processing,
feature extraction and image classification. The objective is to naturally group the
evaluation of non-proliferative diabetic retinopathy at any retinal image. For that an
underlying image preparing stage separate blood vessels microaneurysm, and hard
exudates so as to extricate highlights that can be utilized by calculation to make sense
of retinopathy grade.

Keywords Classification · Image processing · Feature extraction · Analysis ·

Machine learning · Retinopathy

1 Introduction

Diabetic Retinopathy (DR) is a standout among the most successive reasons for
visual debilitation in created nations and is the main source of new instances of visual
deficiency in the working age populace. By and large, almost 75 individuals go dazzle
each day as an outcome of DR A viable treatment for DR require early finding and
consistent checking of diabetic patients, however this is a testing undertaking as the
malady indicates couple of manifestations until it is past the point where it is possible
to give treatment [3].
Diabetic Retinopathy is an eye issue that can cause visual deficiency. Little veins
in the back of the eye are called retinal veins. Indications of Diabetic Retinopathy are
gliding spot in vision, obscured vision and blocked vision. At the point when sugar
level in blood builds, blood vessels in the back of the eye ends up frail and in view

S. Shegdar (B) · D. Patil · S. Kadam

SVERIs College of Engineering, Pandharpur, India
e-mail: [email protected]
A. Bhatlavande
SVERIs College of Engineering (Polytechnic), Pandharpur, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 897
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_90
898 S. Shegdar et al.

of this vessel releases the blood and lipo-proteins liquid [5]. After that liquid ends
up skimming spot in vision with the goal that Diabetic patient cannot see anything
totally through the vision. In the event that we don’t do the treatment of this ailment
on the time then it might be conceivable of complete vision misfortune or visual
deficiency. On the off chance that we distinguished early the indication of Diabetic
Retinopathy, it is conceivable to keep extra loss of vision.
• Four stages of Diabetic Retinopathy are as follows:

First stage is known as Mild Non-Proliferative Diabetic Retinopathy (Mild-

NPDR). In this stage, there will be expand like swelling in the veins in the retina and
little inflatable like swelling in the veins known as Micro aneurysms.
Second stage is known as Moderate Non-Proliferative Diabetic Retinopathy
(Moderate NPDR). In this stage, a portion of the veins in the retina will end up
blocked.
Third stage is known as Severe Non-Proliferative Diabetic Retinopathy (Severe
NPDR). In this stage, more veins are hindered that is the reason the territories of the
retina won’t getting enough blood. Without appropriate stream of blood, the retina
won’t develop fresh recruits vessels and to supplant the harmed veins.
Fourth stage is known as Proliferative Diabetic Retinopathy (PDR). This is
propelled stage. Fresh recruits vessels will start to develop in the retina, however
they will be powerless veins. So frail veins can spills blood and lipo proteins liquid.
Microaneurysms are little red specks on the outside of retina. In the event that
expand like Swelling is happened in the retina’s veins and veins are blocked then we
can say that these are Microaneurysms. Exudates are yellow or white sort of structure
in the retina [9]. Both are shown in Fig. 1. There are two sorts of exudates and they
are shows up contingent upon their essence or event in the vision. Hard exudates have
limits and delicate exudates have no limits or we can say that vague limits otherwise
called cotton fleece spots. Hemorrhages happen because of bleeding and it show up
as little speck. Dot Hemorrhages are a sign of diabetic retinopathy.

Fig. 1 Eye fundus images

showing Microaneurysms
and Exudates
Automated Early Detection of Diabetic Retinopathy 899

• Why machine learning for DR?

As a cost effective way to handle the healthcare resources, systematic screening for
DR has been identified. An important screening tool for early DR detection is the
emergence of automatic retinal image analysis [13]. This can save both cost and time,
as it reduces the manual workload of grading as well as diagnostic cost and time.
Machine learning is a family of computational methods that allows an algorithm to
program itself by learning from a large set of examples that demonstrate the desired
behavior, removing the needs to specify rules explicitly. The accuracy, sensitivity
and specificity of the algorithm for detecting Diabetic Retinopathy (DR) can help
ophthalmologists and physicians raise the red flag and thus provide early treatment
to patients and bring in a more preventive care which can bring down the burden on
healthcare resources. Machine learning can thus help the old adage-prevention better
than cure, by predicting who is more liable to be at risk of DR or not.

2 Literature Survey

1. Akara Sophark, BunyaritUyyanonvara and Sarah Baraman, Automatic Exudates

Detection from Non-dilated Diabetic Retinopathy-Retinal images using Fuzzy
C-means Clustering [2].
Advantage:-The low contrast retinal image intensity increased and a number of
edge pixels were extracted.
Disadvantage:-Time consuming.
2. Walter, J. Klein, P. Massin, and A. Erginary. A contribution of image processing
to the diagnosis of diabetic retinopathy thy, detection of exudates in color fundus
images of human retina [4].
Advantage:-Time consumption is reduced as it uses mathematical morogy
techniques,
Disadvantages:- The paper ignores some types of errors on the border of
segmented exudates in their reported performances.
3. Pilar Perez Conde, Jorge de la Calleja, Antonio Benitez, Ma Auxilio
Medina,Image based classification of Diabetic Retinopathy using Machine
Learning [1]:- This method detects and classifies the diabetic retinopathy.
Preliminary results show that k-nearest neighbors obtained the best result with
68.7% for dataset with different resolutions.
Advantages:- Perform automated classification of Diabetic Retinopathy and
component analysis in less time,
Disadvantages:- The paper does not include testing the methods with larger data
sets and classifying the sub types of the retinopathy.
4. Neelam D. Panse, et al. [6], Glaucoma and Diabetic Retinopathy Diagnosis using
Image Mining. Author has mainly focus on detection of Glaucoma and Diabetic
Retinopathy Glaucoma can be detected by cup to disc ratio (CDR). Diabetic
Retinopathy can be detected by Exudates, Hemorrhages, Microneurysms and
900 S. Shegdar et al.

Cotton Wool Spots. RGB images is converted into YCbCr. Y plane is used for
detection of blood vessels, optic disc and exudates. After candy edge detection,
image will converted into binary to perform Skeletonization operation. DCT is
used for feature extraction. Author has proposed DCT (Discrete Cosine Trans-
form) for feature extraction. Extracted feature goes to SVM classifier. After that
SVM Classifier categories into Normal, DR and Glaucoma.
5. R. Radha and BijeeLakshman [7], Retinal image analysis using morpholog-
ical process and clustering technique. This paper proposes a method for the
Retinal image analysis through efficient detection of exudates and recognizes
the retina to be normal or abnormal. Morphology operators are applied to the
enhanced image in order to find the retinal image ridges. A simple thresholding
method along with opening and closing operation indicates the remained ridges
belonging to vessels. The clustering method is used for effective detection of
exudates of eye.
6. Yun et al. (Neural Network): Automatic classification of different stages
of diabetic retinopathy-mild nonproliferative retinopathy, moderate non-
proliferative retinopathy, severe non-proliferative retinopathy and proliferative
retinopathy using neural network from six features extracted from the retinal
images and accuracy obtained was 72.00 [11, 12].

3 Proposed System

The above Fig. 2. Shows the overall structure of the proposed system. It includes all
the modules which we are going to implement in this project. The architecture gives
idea of input to the system, processing on that input and what will be the output of
the project.

Feature Extraction
Image 1.Removal of optic Classification
Processing disc and blood vessels KNN
Techniques 2.Exudates detection Classifier
3. Micro-aneurysms
detection

Normal NPDR PDR

Fig. 2 Proposed systems for detection and classification of different stages of diabetic retinopathy
Automated Early Detection of Diabetic Retinopathy 901

• Image Database:

The Messidor database [10] includes 1200 eye fundus shading numerical photos
of the back shaft obtained by 3 ophthalmologic workplaces using a shading video
30CD camera on a Topcon TRC NW6 non-mydriatic retinograph with a 45-degree
field of view. The pictures were caught implementing 8 bits for each shading plane
at 1440 X 960, 2240 X 1488 or 2304 X1536 pixels. 800 images were acquired with
pupil dilation (one drop of Tropic amide at 0.5%) and 400 without dilation. The
1200 pictures are bundled in 3 sets, one for every ophthalmologic division, utilizing
the TIFF position. What’s more, an Exceed expectations document with therapeutic
judgments for each picture is given.In this work, we utilize the pictures of only one
ophthalmologie division 1,69 with moderate NPDR (grade 2), and 149 with serious
NPDR (grade 3).

4 Implementation

4.1 Module 1

Image Processing:
The aim of this stage is to enhance those features that will allow, in the next stage,
classify the data set of fundus images. We have investigated the performance of two
image processing techniques: the canny edge detector and the histogram equalization
method.
(1) The Canny edge detector:

Edge detectors are local image processing techniques to detect edge pixels (pixels at
which the intensity of an image function changes abruptly). Particularly we have used
Canny algorithm because its performance is superior in general to the traditional edge
detectors [8]. This method consists of the following basic steps: Smooth the input
in age with a Gaussian filter, compute the gradient magnitude and angle images:
apply non-maxima suppression to the gradient magnitude image and use double
thresholding and connectivity analysis to detect and link edges.
Canny edge detection is used to find out the areas where intensity changes abruptly.
In the above Fig. 3., first image shows the original image and in second one it shows
the areas where intensity change abruptly i.e. the areas where defects are present.
(2) Histogram Equalization:

The histogram equalization is a widely used nonlinear method designed for the
enhancement of images. It is a non-parametric method to match the cumulative
distribution function of some given image to a reference distribution. This method
902 S. Shegdar et al.

Original Image Edge Image

Fig. 3 Output of canny edge detection method

employs a monotonic nonlinear mapping which reassigns the intensity values of

pixels in input image, in order to control the shape of output image intensity histogram
to achieve a uniform distribution of intensities or to highlight certain intensity levels
[8].
Extracted features
In order to automatically detect NPDR we have implemented three main processes
to extract some important features, highlights. Furthermore, the retina edge was
recently fragmented from whatever remains of the image utilizing the red segment
of each retinal image.
(1) Blood Vessels

This process determines the density of blood vessel in a retinal image. For that, the
RGB picture is changed to its CMY portrayal and the magenta segment is isolated.
On the magenta component, morphological operations (i.e. erosion, opening, and
dilation) hide blood vessels. The difference between the magenta component and
the resulting image of the morphological processing is binarized after a histogram
matching that increases its contrast, as shown in Fig. 4. The noise existing in the
binarized image is reduced through dilation and erosion operations. At last, the
density of white pixels (i.e. blood vessels) is computed.

Fig. 4 Extraction of blood vessels

Automated Early Detection of Diabetic Retinopathy 903

Fig. 5 Extraction of microaneurysms

(2) Microaneurysms

Microaneurysms are small lumps in the blood vessels, looking as little and round
shape spots near to tiny blood vessels. In order to determine the number of microa-
neurysms, the green component is extracted and the blood vessels are concealed
utilizing the commotion diminished picture of the previous process as shown in
Fig. 5. Basically, the pixels corresponding to blood vessels are painted with the
average retina color. And then, a disc-based dilation operation is applied to detect
the microaneurysms. It is important to find the difference between the resulting image
and the image of edges to finally remove those edges. Toward the end, the conceivable
microa-neurysms are shifted by utilizing morphological tasks, a number of pixels to
get the actual micro-aneurysms.

5 Result Analysis

The evaluation of our proposal is implemented in python and it has two sections DR
detection and DRNP grade classification. The main idea of this is to detect DRNP
grades. We trained the KNN classifier with all features of the images and then tasted
it. According to the features the system will detect the type as DR or NPDR and
mild, moderate and severe. With reference to Table 1.
The metrics used to evaluate the performance of the machine learning methods
were accuracy, precision, recall, and f-measure, defined as follows:

Table 1 Confusion matrix

Predicted: No Predicted: Yes
for the accuracy-optimized
DRNP detector Actual: No Tn = 150 FP = 20 170
Actual: Yes Fn = 18 TP = 212 230
168 232
904 S. Shegdar et al.

Accuracy = (TPTN)/(TP + TN + FP + FN)

precision = TP/(TP + FP)
recall = TP/(FN + TP)
f measure = 2(recall ∗ precision)/(recall + precision)

where TP (True Positive) is the number of correct predictions of a positive example,

FP (False Positive) is the number of incorrect predictions of a positive example, TN
(True Negative) is the number of correct predictions of a negative instance and FN
(False Negative) is the number of correct predictions of a positive instance.
Performance of the KNN Classifier.

Accuracy − (TP + TN)/(TP + TN + FP + FN)

(150 + 212)/(150 + 20 + 18 + 212) = 0.88

References

1. Conde PP, de la Calleja J, Benitez A, Ma Auxilio Medina Depar-tamento de Posgrado 2010,

Image-Based Classification of Diabetic Retinopathy using Machine Learning, In en Sistemas
y ComputoInteligente Universidad Politecnica de Puebla Puebla, Mexicogperezjdelacalleja,
abenitez, [email protected]
2. Sophark A, Uyyanonvara B, Baraman S (2010) Automatic exudate detection from non-dilated
diabetic retinopathy-retinal images using fuzzy c-means clustering
3. Gurudath N, Celenk M, Riley HB (2014) Machine learning identification of diabetic retinopathy
from fundus images. In: School of Electrical Engineering and Computer Science Stocker
Center, Ohio University Athens, OH 45701 USA
4. Walter TI, Klein JC, Erginay MP (2002) A contribution of image processing to the diagnosis
of diabetic retinopathy-detection of exudates in color fundus images of the human retina, In
IEEE Trans Med Imaging. Oct: 21(10):1236–43
5. https://algoanalytics.com/diabetic retinopathy-machine learning
6. Panse ND, Ghorpade T, Jethani V (May 2015) Glaucoma and diabetic retinopathy diagnosis
using image mining. Int J Comput Appl 5
7. Radha R, Lakshman B (Dec 2013), Retinal image analysis using morphological process and
clustering technique, signal and image processing. Int J (SIPIJ) 4(6)
8. Gonzalez R, Woods R (2007) In Digital image processing Prentice Hall
9. Thomas N, Mahesh T (2014) Detecting clinical features of diabetic retinopathy using image
processing. Int J Eng Res Technol (IJERT) 3(8)
10. Ong G, Ripley L, Newsom R, Cooper M, Casswell A (2004) Screening for sight-threatening
diabetic retinopathy: comparison of fundus photography with automated color contrast
threshold test. Am J Ophthalmol 137(3):445–452
11. Gandhi M, Dhanasekaran D (2013) Diagnosis of diabetic retinopathy using morphological
process and svm classifier. Int Conf Commun Sig Process
12. Yun LW, Acharya UR, Venkatesh YV, Chee C, Min LC, Ng EYK (2008) Identification of
different stages of diabetic retinopathy using retinal optical images. Inf Sci 178:106–121
13. Jiang X, Mojon D (2003) Adaptive local thresholding by verification based multi threshold
probing with application to vessel detection in retinal images. IEEE Trans Patt Anal Mach
Intell 25(1):131–137
Product Lifecycle of Automobiles

V. K. Bupesh Raja, Ajay Shivsharan Reddy, Suraj Ramesh Dhavanapalli,

D. R. Sai Krishna Sanjay, BH. Jashwanth Varma,
and Puskaraj D Sonawwanay

Abstract The product lifecycle concept is acknowledged in both the general

economy and management studies. In line with the concept, every product features
an estimative lifecycle. The car’s lifecycle is divided into four stages; they are, intro-
duction, growth, maturity, and decline. Moreover, after a decline, it is sent to the
scrap yard. Electric vehicles, when linked with low-carbon electricity sources, offer
the perspective for diminishing greenhouse emissions. The sales of the vehicles also
are different in indifferent stages. Consistent with the concept, every product features
an estimative lifecycle. The present study addresses these concepts of the product
lifecycle for automobiles.

Keywords Product lifecycle · PLC · Automobiles · India · Economic analysis ·

Automobile industries

1 Introduction

A study was made on the prevailing ELV from different countries for comparing it
with the Indian scenario. From the findings, it had been realized that the ELV solution
could not be the sole solution towards reducing, recycling waste, and eco-friendly
process. However, this can be one of the various steps that must be taken and are also
crucial. The advantages are often reaped only in concurrence with the development
of the infrastructure of our country. While marketers plan to capture high revenues
at all the stages, they alter their sales techniques noticeably throughout the product
life cycle. Some economists believe that somehow sales of a brand-new product can
be predicted—right from the instant of its entry to its withdrawal from the market.
We will assess if the PLC (Product Lifecycle Concept) is sensible within the Indian

V. K. Bupesh Raja · A. S. Reddy · S. R. Dhavanapalli · D. R. S. K. Sanjay · BH. J. Varma

School of Mechanical Engineering, Sathyabama Institute of Science and Technology, Chennai
600119, India
P. D. Sonawwanay (B)
School of Mechanical Engineering, Dr. Vishwanath Karad MIT World Peace University, Pune
411038, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 905
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_91
906 V. K. Bupesh Raja et al.

automobile industry and can showcase the Lifecycle Analysis of the models sold by
various OEMs in India [1].

2 Economic Analysis

2.1 Material Recycling Policy

For motor recycling, the car is sent to scrap. The vehicle is demolished, and its parts
are sent for sale in the market. In the process, the scrapyard owner earns more profit
than the vehicle owner [2].

2.2 Component Remanufacturing Policy

The cost of a remanufacturing engine is less than half the price of the new engine.
Remanufacturers tend to earn an outstanding amount of money. By recycling the
engine and resale parts to the scrapyard, owners can earn a fair amount of money.
Also, consumers can save money on engine replacement. Apart from this, setting up
this new industry can open doors for new job opportunities [2].

2.3 The Captivating Reality About the Life Cycle of a Car

Many individuals keen on a new car. Although the feeling of driving a novel car is
exciting and exhilarating—there is nothing like getting behind the wheel of a car that
you love and cherish and have driven for years [2, 3].

2.4 What is the Life Cycle of Your Automobile Look like?

Like a human being, cars have a lifecycle, which starts well before the primary time
you drove it and ends well after the last. Here is the fascinating truth about cars: The
lifecycle of a car begins and ends in a very factory. During the manufacturing phase,
countless materials are accustomed to making many different parts. Metal is required
for the vehicle’s body, plastic is employed within the interior and parts of the body,
glass is required for the windows, and cloth and leather are used to upholster the
interior [2, 3].
Product Lifecycle of Automobiles 907

2.5 What Happens During Your Automobile’s Retirement?

Once retired, a car is shipped to a recycling plant to be diminished and repurposed.

Although some parts, like the seats, are unusable—nearly 75% of a car is often
recycled. First, the tires are removed, and also the oil, gas, and other fluids are
drained and sent to a separate processing plant to be repurposed. The metal, plastic,
and glass components are melted down and turned back to usable materials. These
materials are used in making new cars. They drive until the end of their life and are
sent to recycling plants. So, the time that you own your car encompasses just one
small fraction of its lifecycle [2, 3].

3 The Usefulness of Product Life Cycle

In the different phases of the product life cycle, firms will focus on different aspects
of promoting and sales.

3.1 Introduction Phase

We are raising product awareness through advertising/word of mouth.Offering the

product at a discount—penetration pricing to tempt customers to undertake the
product.Target early adopters and influential market leaders. For instance, firms may
offer free product reviews to influential bloggers within the market [3, 4].

3.2 Growth

Firms need to maximize growth to increase product sales from small retailers to big
supermarkets.Firms can change marketing from niche areas to a more mass-market
[3, 4].

3.3 Maturity

With peak penetration, the firm may seek to extend prices to increase profitability.
However, if the market is extremely competitive, the firm may feel the requirement
to keep prices low to defend market share.The firm may seek to enhance the product
to achieve market differentiation and extend the period of maturity[3, 4].
908 V. K. Bupesh Raja et al.

3.4 Decline

In the decline phase, the firm may feel it is best to let the product go—e.g. diesel cars
cannot solve problems with pollution and damage to its brand reputation. However,
with an iPhone, Apple lets old models go to get replaced by the successive model.
Decline and discontinuing the product is often the simplest way to force customers
to shop for an upgrade—next time their contract expires [3, 4].

4 The Lifecycle Impact on the Environment by Material

This can be set apart into different sections like pre-assembly, assembly, use, post-use.

4.1 Pre-Assembly

1. Extraction of minerals for raw materials and their transportation. (example:—

iron ore bauxite, oil, etc.)
2. Production of secondary materials and their transportation to assemblers and
suppliers. (example:—Plastic, steel, aluminium).
3. Production of components and making subassemblies for further transportation
[5].

4.2 Assembly

1. Energy is used in an assembly plant.

2. Pollution caused during assembly and paint shop.
3. Disposal of waste materials in landfills, water and into the recycling system.
4. Transport of finished vehicles to outlets [5].

4.3 Use

1. Emissions caused by driving.

2. Land used for creating new roads, parking, and other facilities.
3. Damages caused by accidents to environment and people
4. Pollution is caused by waste materials at the end of lifecycle. (example:—tyres,
batteries, etc.) [5, 6].
Product Lifecycle of Automobiles 909

4.4 Post-Use

1. Transportation to scrapyards.
2. Toxic wastes from scrapping.
3. Energy used for dismantling/scrapping.
4. Transportation of recycled parts to the place of reuse [5, 6].

5 Conclusion

There has been an unexpected gush in Indian Automobile Industry in the past two
decades. It has attracted global attention to investing in Indian Market. The product
life cycle in an automobile is significant because, on this cycle, many ancillary
industries depend. It’s expected that the automotive industry will play a vital role in
helping the economy to continue India’s growth.

References

1. Comparative analysis of scrap car recycling management policies: The 7th International confer-
ence on waste management and technology conference (2012) Kun Yue, Procedia Environmental
Sciences 16:44–50
2. Broch F, Warsen J, Krinke S (2015) Implementing life cycle engineering in automotive devel-
opment as a helpful management tool to support design for environment. In Sonnemann G,
Margni M (eds) Life cycle management. LCA compendium—the complete world of life cycle
assessment. Springer, Dordrecht, pp 319–329
3. Ribeiro C, Ferreira JV, Partidário P (2007) Life cycle assessment of a multi-material car
component. Int J LCA 12(5):336–345
4. https://www.ausmed.com/cpd/articles/product-life-cycle-relates-healthcare
5. Sabadka D, Molnár V, Fedorko G (2019) Shortening of Life Cycle and Complexity Impact on
the Automotive Industry. TEM J 8(4):1295–1301
6. Agnihotri D, Chaturvedi P (2013) Indian automobile industry: a life-cycle. VSRD Int J Bus
Manage Res 3(8):323–329
Analysis of Heavy Metal Pollutants
in the Sediments from Coastal Sites
of Al-Hodiedah Governorates, Yemen

Majeed Hazzaa Nomaan, Dipak B. Panaskar, and Ranjitsinh S. Pawar

Abstract Contamination of heavy metals in sediments and soils are of increasing

concern. Heavy metals sources in sediments and soils mainly include presence of
sediments as well as anthropogenic sources. Heavy metals are not biodegradable
they tend to persist over long-term periods in the sediments and soil template. They
can be mixed or elated to underground and surface water as well as uptake by agri-
cultural crops or food crop which leads to concerns over animal and human health.
Heavy metal contaminated sediments have direct adverse effects on aquatic life and
ecosystems. Poisoning of food chain and loss of recreational enjoyment are the most
potential problems caused by contaminated sediments.

Keywords Distribution · Heavy metal · Sediment · Coastal · Pollutant · Etc

1 Introduction

Heavy metals term is used to describe more than metal that are metals or metalloids
(elements that have both metal and nonmetal characteristics). The heavy metals
include Chromium, Arsenic, Cadmium, lead, Mercury, and Manganese. Generally
the heavy metals have densities above 5 g/cm3 . They cannot be degraded or destroyed
and are persistent in all parts of the Environment. The Human activities affect the
natural geological and biological redistribution of heavy metals through pollution of
the air, water, and soil. All the primary anthropogenic sources of heavy metals are
point sources such as mines, foundries, smelters, and coal-burning power stations
as well as diffuse sources such as combustion by products and car exhaust. Humans
also affect the natural geological and biological redistribution of heavy metals by

M. H. Nomaan
Amran University, Amran, Yemen
D. B. Panaskar
SRTM University, Nanded, Maharashtra, India
R. S. Pawar (B)
SVERIs College of Engineering, Pandharpur, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 911
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_92
912 M. H. Nomaan et al.

altering the chemical form of heavy metals released into the environment. Such
change often affect a heavy metal’s toxicity by allowing it to bio-accumulate in
plants and animals, bio-concentrate in the food chain, or attack specific organs of
bodies. They are named after those metals between atomic number 21 (scandium)
and atomic number 84 (polonium), except for aluminum, which has atomic number
13, but it is also considered a heavy metal [1]. Heavy metals can also bio-accumulate
in the aquatic flora and fauna, such as plankton, fish, etc. Therefore, an early detection
of high heavy metal concentrations in water bodies such as rivers or wetlands are
vital for nature conservation [2]. The main transport of heavy metals in water bodies
is mainly through sediment [3], in the form of suspended solids.

1.1 Environmental Threats to the Yemen Seas

Recently, several studies have been carried out describing the concentration and
distribution of heavy metals along the Yemeni Red Sea region. They have focused
on the existence of heavy metals in fish, mussels and shoreline sediments. In general,
data revealed that the levels of cadmium (Cd), cobalt (Co), crome (Cr), cupper (Cu),
ferrous (Fe), manganese (Mn), nickel (Ni), lead (Pb), and zinc (Zn) in the region were
generally within the range of levels reported for other regions in the world [4, 5].
However, the concentrations of Cd, Cr, Mn, Ni, Pb and Zn in some fish collected from
As Salif and Al Hodiedah sites were higher than those in other sites in the region. It is
largely believed that human activities attribute to such high concentrations, which was
mainly due to human activities in both sites. Concentrations of some metals, including
Zn, Cu, Mn and Fe in both shrimps and sea snails collected were higher more than in
other fish collected from the same sites. Another study, conducted by DouAbul and
Hebba [6], revealed that the concentration of heavy elements in fish, molluscs and
shoreline sediments in the region were rather low concentrations. Additionally, the
shrimps (P. semiculatus) collected showed a lower concentration (5.93 ppm) of Cu,
compared with that (29–171 ppm) showed by Karbe [7]. However, in most recent
studies [8], the mean concentrations of some heavy elements in dissolved particulates
and sediments, collected from several location in the region were slightly higher than
those found in other parts of the world. Sources of these metals were most likely
natural origin.

2 Materials and Methods

Fifteen representative sediment samples have been collected from Al-Hodiedah

coastal areas during the years 2011 by adopting scientific methods used to collect
sediment samples at five locations was taking the amount of 10 g per sample. The
sediment were immediately stored in sealed plastic containers and kept refrigerated.
Analysis of Heavy Metal Pollutants in the Sediments … 913

Salinity was also measured at each sample location. For each location, surface sedi-
ment samples were collected at 5 cm depths after removing any plant debris and
large stones. The sampling locations include areas of different environmental back-
grounds like fish aquaculture, recreational areas, ports, industries and urban areas.
All samples preserved in the Central Laboratory of the University of Aden at the time
of processing for digestion and analysis. A total of 15 samples from sediments were
collected from the sites during low tide in December 2011. The scientific method
used to collect sediment samples, where the study samples were collected from the
province of Al-Hodeidah five regions are as follows: Hodiedah port, Al-kateef shore,
Cornish location, Almehwate site and Al-Manjer location (Fig. 1) [9].

YEMEN
SANAA

Al Hudaydah

Scale
0 100 200 km
Aden
100 200 mi

0 0
42 44
0
16

1 Sanaa
2
3
4
Red Sea
5
Scale
0 50 100 km
1
Sampling Station
0
14

Fig. 1 Map of costal Al-Hodiedah Governorate

914 M. H. Nomaan et al.

2.1 Digestion of Samples and Heavy Metals Analysis

The trace elements from sediments were determined according to Oregioni and
Aston, [10] procedure. The concentration of heavy metals like Cu, Pb, Zn, Cd, Co,
Ni, Cr, Fe and As for all samples of Al-Hodiedah were determined. Heavy metal
analysis was performed on the < 63 µm fraction of the sediment which had been
separated by sieving after drying and grinding. The determination of Heavy metals
in particulate and sediment samples was done according to the procedure described
by Sturgeon [11]. The Perkin-Elmer 2380 atomic absorption spectrophotometer was
used for heavy metal analysis. The general method involves atomization of samples
by thermal sources and the absorption of a specific wavelength by the atomic source
as it is excited. The radiation used is a hollow cathode lamp containing, as its cathode,
the same element under analysis.

3 Result and Discussion

Heavy metals that will be referred to in this study as contaminants of concern that
will form the focus of this study are As, Cd, Cr, Cu, Co, Fe, Ni, Pb, and Zn. These
metals are known to be present in landfills and their presence poses serious problems
to human health and environmental concern. Such focus is because of their known
toxicity with regard to human health and ecology or their ability to bio accumulates
and moves through the food chain. Bioaccumulation processes may adversely affect
food resources by making them unfit for human consumption [12]. The Heavy Metal
analysis of sediments from Al-Hodiedah Governorate during year 2011 has been
presented in Table 1.

3.1 Arsenic (as)

Arsenic of sediment samples varies from 0.27 to 1.22 µg/g in the year 2011 Table
1. The minimum As has been recorded from Cornish location, while maximum As
has been recorded from Hodiedah Harbour. The average As and Standard Deviation
of the sediment is 0.6957 µg/g and 0.3490 respectively Fig. 2.

3.2 Cadmium (Cd)

Cadmium of sediment samples varies from 7.8 to 9.5 µg/g in the year 2011 Table 1.
The minimum Cd has been recorded from Al-Manjer location, while maximum Cd
Table 1 Heavy metal concentration (µg/g dry wt.) of sediment samples from Al-hodiedah governorate
Zones Heavy metal concentration during 2011
As Cd Co Cu Cr Fe Ni Pb Zn
µg/g dry wt
Hodiedah harbour 1.22 9.5 31.25 23.45 117.25 2591.75 15.83 62 52
Al-Kathib shore 0.63 9.25 26.7 18 46.2 2526 13.33 62 39.54
Cornish location 0.27 7.82 25.25 14.87 58.65 2541.18 17.15 63 36.33
Almehwate Site 0.53 7.94 25.55 21.45 87.14 2379.25 18.33 65 48.74
Analysis of Heavy Metal Pollutants in the Sediments …

Al-Manjer location 0.73 7.8 23.1 18.33 56.27 2677.71 19 62 37.33

Min 0.27 7.8 23.1 14.87 46.2 2379.25 13.33 62 36.33
Max 1.22 9.5 31.25 23.45 117.25 2677.71 19 65 52
Avg 0.6957 8.462 26.37 19.22 73.102 2543.18 16.728 62.8 42.788
StDev 0.3490 0.8398 3.0225 3.3192 28.9914 109.12 2.2505 1.3039 7.1122
915
916 M. H. Nomaan et al.

Concentration of As during 2011

1.4

1.2

1
As (μg/g)

0.8

0.6

0.4

0.2
Hodiedah Al-Kathib Shore Cornish Location Almehwate Site Al-Manjer
Harbour Location
Al Hodiedah Governorate Locations

Fig. 2 Concentration of Arsenic in Sediments from Al Hodiedah Governorate during 2011

Concentration of Cd during 2011

10

9
Cd (μg/g)

7
Hodiedah Al-Kathib Shore Cornish Location Almehwate Site Al-Manjer
Harbour Location
Al Hodiedah Governorate Locations

Fig. 3 Concentration of cadmium in sediments from Al Hodiedah Governorate during 2011

has been recorded from Hodiedah Harbour. The average Cd and Standard Deviation
of the sediment is 8.462 µg/g and 0.8398 respectively Fig. 3.

3.3 Cobalt (Co)

Cobalt of sediment samples varies from 23.1 to 31.25 µg/g in the year 2011 Table 1.
The minimum Co has been recorded from Al-Manjer location, while maximum Co
has been recorded from Hodiedah Harbour. The average Co and Standard Deviation
of the sediment is 26.37 µg/g and 3.0225 respectively Fig. 4.
Analysis of Heavy Metal Pollutants in the Sediments … 917

Concentration of Co during 2011

33

29
Co (μg/g)

25

21
Hodiedah Al-Kathib Shore Cornish Location Almehwate Site Al-Manjer
Harbour Location
Al Hodiedah Governorate Locations

Fig. 4 Concentration of cobalt in sediments from Al Hodiedah governorate during 2011

Concentration of Cu during 2011

25

21
Cu (μg/g)

17

13
Hodiedah Al-Kathib Shore Cornish Location Almehwate Site Al-Manjer
Harbour Location
Al Hodiedah Governorate Locations

Fig. 5 Concentrations of copper in sediments from Al Hodiedah governorate during 2011

3.4 Copper (Cu)

Copper of sediment samples varies from 14.87 to 23.45 µg/g in the year 2011 Table
1. The minimum Cu has been recorded from Cornish Location, while maximum Cu
has been recorded from Hodiedah Harbour. The average Cu and Standard Deviation
of the sediment is 19.22 µg/g and 3.3192 respectively Fig. 5.

3.5 Chromium (Cr)

Chromium of sediment samples varies from 46.2 to 117.25 µg/g in the year 2011
Table 1. The minimum Cr has been recorded from Al-Kathib Shore location, while
918 M. H. Nomaan et al.

Concentration of Cr during 2011

120

100

80
Cr (μg/g)

60

40

20
Hodiedah Al-Kathib Shore Cornish Location Almehwate Site Al-Manjer
Harbour Location
Al Hodiedah Governorate Locations

Fig. 6 Concentration of Chromium in Sediments from Al Hodiedah Governorate during 2011

maximum Cr has been recorded from Hodiedah Harbour. The average Cr and
Standard Deviation of the sediment is 73.102 µg/g and 28.9914 respectively Fig. 6.

3.6 Iron (Fe)

Iron of sediment samples varies from 2379.25 to 2677.71 µg/g in the year 2011 Table
1. The minimum Fe has been recorded from Almehwate location, while maximum Fe
has been recorded from Al-Manjer location. The average Fe and Standard Deviation
of the sediment is 2543.18 µg/g and 109.12 respectively Fig. 7.

Concentration of Fe during 2011

2700

2600
Fe (μg/g)

2500

2400

2300
Hodiedah Al-Kathib Shore Cornish Almehwate Site Al-Manjer
Harbour Location Location
Al Hodiedah Governorate Locations

Fig. 7 Concentration of iron in sediments from Al Hodiedah governorate during 2011

Analysis of Heavy Metal Pollutants in the Sediments … 919

Concentration of Ni during 2011

20

17
Ni (μg/g)

14

11
Hodiedah Al-Kathib Shore Cornish Location Almehwate Site Al-Manjer
Harbour Location
Al Hodiedah Governorate Locations

Fig. 8 Concentration of nickel in sediments from Al Hodiedah governorate during 2011

3.7 Nickel (Ni)

Nickel of sediment samples varies from 13.33 to 19 µg/g in the year 2011 Table 1.
The minimum Cd has been recorded from Al Kathib Shore, while maximum Cd has
been recorded from Al-Manjer Location. The average Cd and Standard Deviation of
the sediment is 16.728 µg/g and 2.2505 respectively Fig. 8.

3.8 Lead (Pb)

Lead of sediment samples varies from 62 to 65 µg/g in the year 2011 Table 1.
The minimum Pb has been recorded from Al-Manjer location, Al Kathib Shore and
Hodiedah Harbour, while maximum Pb has been recorded from Almehwate Site.
The average Pb and Standard Deviation of the sediment is 62.8 µg/g and 1.3039
respectively Fig. 9.

3.9 Zink (Zn)

Zink of sediment samples varies from 36.33 to 52 µg/g in the year 2011 Table 1.
The minimum Zn has been recorded from Cornish location, while maximum Zn has
been recorded from Hodiedah Harbour. The average Zn and Standard Deviation of
the sediment is 42.788 µg/g and 7.1122 respectively Fig. 10.
920 M. H. Nomaan et al.

Concentration of Pb during 2011

66

64
Pb (μg/g)

62

60
Hodiedah Al-Kathib Shore Cornish Location Almehwate Site Al-Manjer
Harbour Location
Al Hodiedah Governorate Locations

Fig. 9 Concentration of lead in sediments from Al Hodiedah governorate during 2011

Concentration of Zn during 2011

60

50
Zn (μg/g)

40

30
Hodiedah Al-Kathib Shore Cornish Location Almehwate Site Al-Manjer
Harbour Location
Al Hodiedah Governorate Locations

Fig. 10 Concentration of zink in sediments from Al Hodiedah governorate during 2011

4 Conclusion

As can be seen from the tables from previous chapter mean Heavy metals level in all
samples from Al-Hodiedah Governorate in the following order: Fe > Cr > Pb > Zn
> Co > Cu > Ni > Cd > AS. Arsenic, copper, Nickel and Zinc has been found with
lowest concentration in marine sediments of Costal Al-Hodiedah Governorate. All
concentrations are low and can be attributed to natural background levels. Therefore,
it can be deduced that they poses little or no treat to this most of this Environ-
ment. Cadmium, cobalt, Chromium, Iron and lead has been found with the highest
concentration in the marine sediments of Costal Al- Hodiedah Governorate. All the
concentrations can be attributed to natural background levels. Therefore, they pose
threat to the surrounding environment.
Analysis of Heavy Metal Pollutants in the Sediments … 921

References

1. Schnoor JL (1996) Environmental modelling. Willey, New York

2. Ulbrich K, Marsula R, Jeltsch F, Hofmann H, Wissel C (1997) Modelling the ecological impact
of contaminated river sediments on wetlands. Ecol Model 94:221–230
3. Lumborg U, Windelin A (2003) Hydrography and cohesive sediment modelling: application
to the Romo Dyd tidal area. J Marine Syst 38:287–303
4. Al Mudafar N, Hebba H, Fawzi I (1994) Trace metals in fish from Red Sea coast of Yemen
(fish death phenmenon)
5. Hebba HMA, Al Mudafar N (2000) Trace metal in fish, mussels, shrimp and sediment from
Red Sea coast of Yemen. Bull Nat Inst Oceanogr Fish A.R.E 26:151–165
6. DouAbul, A.A-Z. and Haddad, A. (1999). The Red Sea and Yemen’s Red Sea Environments.
In DouAbul, A., Rouphael, T., and Sylvia & Marchant, R. Protection of Marine Ecosystem of
the Red Sea Coast of Yemen. UNOPS Press.
7. Karbe L, Schnier C, Siewers HO (1977) Trace elements in mussels (Mytilus edulis) from
coastal areas of the North Sea and the Baltic. Multielement analysis using instrumental neutron
activation analysis
8. Hebba HA, Majed AM, Hamed T, AL-Saad MA, Al Mouniam (2002) Trace elements in
dissolved, particulate water and sediment from the Red Sea coast of Yemen, Faculty of Marine
Science and Environment, University of Hudaydah, Republic of Yemen
9. Nomaan MH, Pawar RS, Panaskar DB (2012) Assessment of heavy metals in sediments from
coastal Al-Hodiedah Governorate, Yemen. Univ J Environ Res Technol 2(3):168–173
10. Oregioni B, Aston SR (1984) Determination of selected trace metals in marine rediments by
flame/flameless atomic absorption spectrophotometer. IAEA Monaco lab., Internal re-port.
Now cited in reference method in pollution studies No. 38, UNEP
11. Sturgeon RE, Desaulniers JA, Berman SS, Russell DS (1982) Determination of trace metals
in estuarine sediments by graphite furnace atomic absorption spectrophotome- try. Anal Chim
Acta 134:283–291
12. Denton GRW, Wood HR, Concepcion LP, Siegrist HG, Eflin VS, Narcis DK, Pangelinan GT
(1997) Analysis of in-place contaminants in marine sediments from four harbor locations on
guam: a pilot study, water and environmental research institute of the western pacific, Technical
Report No. 87, University of Guam, Mangilao, Guam
Development of Low Cost PCR Product
Detection System for Screening
and Diagnosis of Infectious Diseases

Patil Yogesh Navalsing, Suri Vinod Kumar, Suri Aseem Vinod,

Kar Harapriya, and Thakur Mansee Kapil

Abstract Rapid and accurate detection of Infectious disease at rural setup is one of
the most challenging tasks. Advancement in molecular biology-based techniques had
revolutionized the field of diagnosis. Among several advantages, these techniques
are restricted to tertiary healthcare centers due to high cost and specific operating
procedures. We had developed a fast, battery-operated, portable device which can be
clubbed with conventional PCR for screening of Infectious diseases at rural setup.
The proof of concept was tested on standard reference samples of Mycobacterium
tuberculosis (H37RV strain) DNA as a positive control and ATCC Strain of E.coli
as a negative control in duplicates by clubbing the system with Conventional PCR
machine (For Amplification) and compared the result with Real-Time PCR and Fluo-
rescence microscopy techniques. The results of the detection system were found
consistent with conventional diagnostic techniques and can be used for screening of
infectious diseases at rural setup.

Keywords PCR · DNA detection · Fluorescence · Real-Time PCR ·

Tuberculosis · E.coli · Mycobacterium tuberculosis

1 Introduction

Infectious diseases are one of the oldest health problems associated with mankind and
consuming millions of people every year. Pathogenic microorganisms are responsible
for infectious diseases and can be transmitted between individuals and population,
thereby threatening the general public health and, potentially, the economy [1]. In

P. Y. Navalsing (B) · T. M. Kapil

Department of Med. Biotechnology, MGM School of Biomedical Sciences, MGMCRL,
MGMIHS, Kamothe, Navi Mumbai, Maharashtra, India
S. V. Kumar · S. A. Vinod
Sinergy Nano Systems, Navi Mumbai, Maharashtra, India
K. Harapriya
Department of Medical Microbiology, MGM Medical College, MGMIHS, Kamothe, Navi
Mumbai, Maharashtra, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 923
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_93
924 P. Y. Navalsing et al.

order to solve the problem of over increasing rates of these infectious diseases, rapid
and specific identification of the causative agent plays a crucial role in successful
treatment and curing of the infected individuals. Conventional methods for iden-
tification of microorganisms associated with infectious diseases involves culturing
and sub-culturing of organisms, which is time-consuming and labor-intensive [2]
[3]. After discovery of the Nucleic Acid amplification and detection technologies
(PCR assays), various methods have been developed for the detection and character-
ization of microorganisms. These assays are highly sensitive & specific and can be
utilized for screening of all classes of microorganisms which causes various infec-
tious diseases [4, 5]. Nucleic acid based assays are continuously expanding in various
fields and are being adopted by clinical laboratories for identification of the disease
specific pathogens accurately in a time bound manner [6–9]. These methods had
identified numerous microbial pathogens responsible for acute and chronic diseases
of the mankind [10–17]. Nucleic acid based testing (NAT) relies on simple princi-
ples of direct amplification and detection of pathogen specific nucleic acid sequences,
and have better sensitivity than immunological and cell culture based detection tech-
niques [18]. These techniques offer very simple and effective way for screening of
infectious diseases such as TB and HIV, which are very difficult to diagnose by
conventional microbiological methods. Nucleic acid based testing’s are being devel-
oped for screening of drug resistance associated with particular pathogens and had
great advantages as compared to conventional antibiotic susceptibility testing. In spite
of the various advantages of Nucleic acid testing (NAT) based assays, these tests are
restricted in centralized laboratories due to use of high-end technologies and complex
operating procedures [19]. A Currently available instrument for nucleic acid detection
are bulky and expensive, which are available only at sophisticated tertiary healthcare
centers, and are not adopted currently by primary healthcare centers [20]. Due to
lack of skilled manpower and limited financial resources in rural areas of developing
countries well established commercially available NAT systems, cannot be deployed
at rural setup through the network of centralized clinical laboratories for serving
the need of poor strata of the society [21, 22]. The major challenge faced by devel-
oping world in diagnostic field is the development of robust, portable and low-cost
assays which will assist the detection of disease specific genetic biomarkers reliably
in places which are far away from the tertiary healthcare centers [23]. ASSURED
(affordable, sensitive, specific, user-friendly, fast, equipment free, and delivered) has
been invented for more than 10 years to define the ideal test to meet the needs of the
developing world [24].
In order to address these problems, various efforts are being made to develop
simple and low cost NAT-based assays and have focused mainly on the amplifica-
tion and detection steps for the detection of nucleic acids which can suite resource
limited settings [25]. Lot of simpler and cost effective nucleic acid amplification
technologies are emerging in the field of medical diagnostics, but the problem of
detection is still. Here we introduce a very simple and robust, but sensitive fluores-
cence method for end point detection of amplified nucleic acid using a miniaturized
low cost optical detection system which can be clubbed with low cost conventional
DNA amplification systems and can be operated at remote areas as well, for detection
Development of Low Cost PCR Product Detection System … 925

of Nucleic acids. Real time detection involves real-time monitoring of the amplifi-
cation reactions which need highly sensitive and accurate optical instrumentation
which leads to increase overall cost of the equipment on the other hand endpoint
detection technique needs simple instrumentation and provide less complex outputs
processing for interpretation which reduces the overall cost of detection. Hence we
had worked out on the development of simple and robust endpoint amplified nucleic
acid detection system for rural parts of the developing nations. We had conducted the
pilot study on standard reference positive and negative samples and found that the
detection system results matches with existing real-time PCR assay and fluorescence
detection systems.

2 Materials and Methods

2.1 Development of PCR Product Detection System

Development of the end point amplified nucleic acid detection system is based on
the principle of fluorescence signal excitation and emission from the targeted nucleic
acids and fabricated by using off the shelf components in order to reduce develop-
ment cost of nucleic acid detection system [26]. We had utilized 6—fluorescein
amidite (FAM) dye specific filters and light source for excitation and emission of
these molecules as these molecules are attached to the probes in most of the commer-
cially available Nucleic acid based amplification kits. The basic design of the DNA
detection system was developed as per the basic principle of excitation and emission
of light by fluorescence molecules attached to Target DNA molecules which is as
shown in following Fig. 1.
Here we are amplifying the targeted Nucleic acid with the help of conventional
PCR protocol by using sets of primers and FAM Labelled probes and then visualizing
the samples post amplification on PCR Product Detection system, if tube shaped
fluorescence is observed inside the detector then the sample is positive and if tube
shaped fluorescence is absent then sample is considered as negative for the targeted
nucleic acid.

2.2 Fabrication of Detection System

The excitation assembly was fabricated for excitation of fluorescence molecules

tagged with targeted DNA molecules. This assembly mainly includes the Light
Source as a monochromatic Laser of Wavelength 450–480 nm plus minus 10 nm the
power source for this laser was designed and integrated with the laser for emission
of the desired wavelength of light. After the fabrication of the excitation assembly
the emission assembly was done to filter the light coming from the source as well as
926 P. Y. Navalsing et al.

Fig. 1 The basic design of DNA detection system

targeted molecules, which includes a fluorescence filter with a peak wavelength of

520 nm ± 10 nm.

2.3 Study of Response for Developed PCR Product Detection

System and Testing of Proof of Concept

2.3.1 Sample Selection

Mycobacterium tuberculosis standard reference strain of H37RV has been selected

as a positive control and an ATCC Standard reference strain of E. coli has been
selected as a negative control for the experiments, which are obtained from MGM
Microbiology, Laboratory, Kamothe, Navi Mumbai.

2.3.2 DNA Extraction

DNA was extracted from the standard reference samples by using commercially
available HiPurA™ Mycobacterium tuberculosis DNA Purification Kit (Cat No.
MB545) from HiMedia.
Development of Low Cost PCR Product Detection System … 927

2.3.3 Amplification Protocol

The protocol has been standardized to amplify the control samples of Mycobac-
terium tuberculosis (Positive control) and E.coli (Negative Control) by using Hime-
dias Mycobacterium Tuberculosis real time Probe based PCR kit (Cat No. MBPCR
108) on Conventional low cost PCR machine for detection of the Tuberculosis. The
amplification protocol was standardized as per the kit insert. The mastermix was
prepared by mixing 10 µL of 2 × PCR Taq Mixture (MBT061), 2 µL of Primer Mix
for M. tuberculosis (DS0132A), 0.5 µL MTB Probe (FAM Labelled) (DS0253) and
5 µL of Extracted Template DNA, the volume of reaction mixture is then balanced
by adding 2.5 µL of Molecular Biology Grade Water for PCR (ML065), and the
prepared reaction mixture was amplified on Himedias Wee-16™ Thermal Cycler by
subjecting the samples to PCR cycles as Initial denaturation for 95 °C for 15 min
and 40 cycles of Denaturation at 95 °C for 15 s and annealing at 60 °C for 1 min.

2.3.4 Detection of Amplified PCR Product on the Developed Detection

System

After amplification of samples they are quickly transferred to detection system for
end point analysis, After placing the samples in the sample resting slot the samples
were excited by using monochromatic source of light and then there after the signal is
visualized by naked eyes in order to conclude the results as if tube shaped fluorescence
is observed inside the filter of detection system then sample is considered as positive
and if tube shaped fluoresce is absent inside the filter then samples is labelled as
negative for Tuberculosis. The detection system is useful for screening purpose.

2.4 Real Time PCR Analysis

The amplification of Mycobacterium tuberculosis (H37RV) Strain specific sequences

mainly mpt64 gene [27], was targeted for amplification by using Mycobacterium
Tuberculosis real time Probe based PCR kit (Cat No. MBPCR 108) on HimediaInsta
Q48™ Real-time PCR Detection System by the standard PCR cycles mentioned in the
kit manual as Initial denaturation for 95 °C for 15 min and 40 cycles of Denaturation
at 95 °C for 15 s and annealing at 60 °C for 1 min. The reaction mixture was prepared
by mixing 10 µL of 2X PCR Taq Mixture (MBT061), 2 µL of Primer Mix for M.
tuberculosis (DS0132A), 0.5 µL MTB Probe (FAM Labelled) (DS0253) and 5 µL of
Extracted Template DNA, the volume of reaction mixture is then balanced by adding
2.5 µL of Molecular Biology Grade Water for PCR (ML065), the main objective
behind real-time PCR analysis is to compare the result of detection system with this
gold standard technique in order to conclude about the sensitivity and specificity of
the developed detection system.
928 P. Y. Navalsing et al.

2.5 Fluorescence Microscopic Analysis of Amplified Nucleic

Acids

Post analysis of samples on the detection system the samples were analyzed on
FL Auto Fluorescence Microscope (EVOS® FL Auto Imaging System; Cat No:
AMAFD1000) under 2X magnification for detection of visible signal generated from
amplified samples. The significance behind performing Fluorescence microscope
analysis is to test proof of concept as well as to compare the results of the detection
system. If tube shaped fluorescence signal is observed under the microscope, then
samples are considered positive for targeted nucleic acids and if fluorescence signal
is absent, then they are labeled as negative for the targeted nucleic acid.

3 Results and Discussion

3.1 Fabrication of the PCR Product Detection System

The Detection system was fabricated by suing Laser as an Excitation source of

monochromatic light to excite the targeted molecules and then subsequent filtration
of the excited light from the light from source (Laser) by using the fluorescence filter
specific for the excited light as mentioned in the methodology section. The integrated
assembly of the detection system is as shown in the following Fig. 2. In this system
the samples can directly visualized inside the system without need of any expensive
optical devices such as lenses and ultra-sensitive cameras, the samples can directly
analyzed by naked eye for detecting the target Nucleic acid molecules which are
specific to particular disease.

Fig. 2 Developed PCR

Product detection system
Development of Low Cost PCR Product Detection System … 929

a b c d

Fig. 3 Results of samples on developed PCR Product detection system; Image A & B denotes-
PCR Tube shaped fluorescence signal for positive control sample 1 & 2 Indicated by Arrow; Image
C & D denotes- No fluorescence Signal from Negative control sample 3 & 4

3.2 Study of Response of Developed PCR Product Detection

System and Testing of Proof of Concept

After completion of the conventional PCR assay the samples are quickly transferred
inside the detection system and the results are captured for both positive and negative
samples as shown in the Fig. 3.
Clear visible tube shaped fluorescence Signal is observed in both the positive
control 1 and 2 in triplicate indicated by arrow in image A and B and no fluores-
cence was seen in negative control 3 and 4 indicated by image C and D of Fig. 3,
which indicated that the results are consistence with the Real Time PCR assay and
fluorescence microscopy results. Thus the developed detection system in the present
study can be directly utilized for detection of FAM Labelled amplified nucleic acid
for specific screening of infectious diseases at rural setup. The developed detection
system in present study is mainly consist of low cost laser as excitation source and
fluorescence filter for filtration of excited light emitted from sample by eliminating
light directly coming from laser (Excitation source), the whole assembly is created by
using of the shelf components in order to minimize the cost of the detection system.
The basic operating procedure to use this system will be, amplification of Pathogen
specific sequences by using FAM labelled probes by conventional PCR machines
and then subsequent detection of the amplified samples by using our PCR product
detection system. This system can be used for identification of both bacterial and
viral pathogens by using commercially available FAM probes based real time PCR
kits.

3.3 Real Time PCR Analysis

The samples were successfully amplified by real time PCR (HiMediaInsta Q48TM
Real time PCR machine) by targeting mpt64 gene specific region of Mycobacterium
tuberculosis bacteria by using Hi-media Mycobacterium tuberculosis real time FAM
Probe based PCR kit, the amplification curve for the samples is as shown in the
Fig. 4. The curve A and B of Fig. 4 represents the successful amplification of mpt46
gene targeted by the FAM labelled probes for detection of Tuberculosis. The results
930 P. Y. Navalsing et al.

Fig. 4 Real Time PCR Amplification plot for the Positive and Negative control samples on HiMedi-
aInsta Q48 System; Curve A& B denotes- Amplification plot of H37RV strain as a Positive control
samples; Line C & D denotes- Amplification plot of E.coliATCC Strain as a Negative control
samples

shows that, threshold cycles for the positive control (H37RV Strain) was found to be
19.3 and 19.5 respectively and for negative control samples flat line was observed in
comparison to positive samples which indicates that no amplification was observed
in the negative samples as shown by C and D overlapped lines of Fig. 4, which
indicates the samples are negative for targeted mpt64 sequences as it is specific for
Mycobacterium Tuberculosis, hence no amplification is observed inside the E.coli
samples (Negative control).

3.4 Fluorescence Microscope Analysis

Amplified PCR products were screened by using Fluorescence microscopy technique

by directly placing PCR tube on slides under 2X magnification, the positive control
samples shown clear visible fluorescence glow and no fluorescence was observed in
negative control samples. Figure 5 describes the representative images for Positive
and negative control samples Image A& B of Fig. 5, denotes positive control sample
1 & 2 Fluorescence signal under 2X magnification and Image C & D of Fig. 5
denotes absence of fluorescence signal from Negative control samples 3 & 4. This
observation indicated that the proof of concept has been tested and samples can be
analyzed on the developed detection system.
Sensitivity, specificity, positive and negative predictive values (with 95% confi-
dence intervals) of DNA detector were compared to with RTPCR and fluorescent
Development of Low Cost PCR Product Detection System … 931

a b c d
Fig. 5 Images for PCR Products analyzed on EVOS® FL Auto Fluorescence microscope, under
2x magnification; Image a & b denotes - PCR Tube shaped fluorescence signal for positive control
sample 1 & 2; Image C & D denotes- No fluorescence Signal from Negative control sample 3 & 4

microscopy, which were taken as a reference standard for Mycobacterium tubercu-

losis as a positive control and E.coli as a negative control. Inter-test agreement for
both results of positive and negative readings was expressed overall agreement.

4 Conclusion

In-order to screen the infectious diseases in rural setup in developing nations, the
systems should be developed which will be easy to operate, robust, low cost, light
weight (portable) and particularly battery operated, so that paramedical persons can
operate it. Several researchers had made an effort to develop such systems for the
endpoint PCR product detection but they are complex. Other researchers are trying to
develop low cost lab on chip detection systems but currently they are under develop-
ment phases our system offers a great reliability over other methods as the results can
be directly visualized by naked eyes and there is no need of photo detectors as well as
complex optoelectronic components which leads to increase in the overall cost of the
detection system. The only limitation associated with our study, we neither performed
culture and PCR to screen statistically sufficient number of samples patients to be
considered as gold standard for comparison with our detector which in the near future
we will be tested. The inclusion of either of it would have enabled us to calculate
sensitivity and specificity of the DNA end point detector for assessing the routine
applicability of and the system will be optimized for screening of various infectious
diseases other than Tuberculosis.

References

1. Chen H, Liu K, Li Z, Wang P (2019) Point of care testing for infectious diseases. Clinica Chim
Acta 493:138–147
2. Fantahun M, Kebede A, Yenew B, Gemechu T, Mamuye Y, Tadesse M, Yaregal Z (2019)
Diagnostic accuracy of Xpert MTB/RIF assay and non-molecular methods for the diagnosis
of tuberculosis lymphadenitis. PLoS ONE 14(9):
932 P. Y. Navalsing et al.

3. Kaltsas P, Want S, Cohen J (2005) Development of a time-to-positivity assay as a tool in the

antibiotic management of septic patients. Clin Micro Inf 11(2):109–114
4. Christensen DR, Hartman LJ, Loveless BM, Frye MS, Shipley MA, Bridge DL, Richards MJ,
Kaplan RS, Garrison J, Baldwin CD, Kulesh DA (2006) Detection of biological threat agents
by real-time PCR: comparison of assay performance on the RAPID, the lightcycler, and the
smart cycler platforms. Clin Chem 52(1):141–145
5. Emanuel PA, Bell R, Dang JL, McClanahan R, David JC, Burgess RJ, Thompson J, Collins L,
Hadfield T (2003) Detection of Francisellatularensis within infected mouse tissues by using a
hand-held PCR thermocycler. J Clin Micro 41(2):689–693
6. Timbrell JA (1998) Biomarkers in toxicology. Toxicology 129(1):1–12
7. Tanner M, Goebel B, Dojka M, Pace N (1998) Specific ribosomal DNA sequences from diverse
environmental settings correlate with experimental contaminants. Appl Environ Microbiol
64:3110–3113
8. Brinkmann B (1998) Overview of PCR-based systems in identity testing. Forensic DNA
profiling protocols. Humana Press, pp 105–119
9. Mothershed EA, Whitney AM (2006) Nucleic acid-based methods for the detection of bacterial
pathogens: present and future considerations for the clinical laboratory. Clinica Chim Acta
363(1–2):206–220
10. Challoner PB, Smith KT, Parker JD, MacLeod DL, Coulter SN, Rose TM, Schultz ER, Bennett
JL, Garber RL, Chang M (1995) Plaque-associated expression of human herpes virus 6 in
multiple sclerosis. Proc Natl Acad Sci 92(16):7440–7444
11. Chang Y, Cesarman E, Pessin MS, Lee F, Culpepper J, Knowles DM, Moore PS (1994) Iden-
tification of herpesvirus-like DNA sequences in AIDS-associated Kaposi’s sarcoma. Science
266(5192):1865–1869
12. Choo QL, Kuo G, Weiner AJ, Overby LR, Bradley DW, Houghton M (1989) Isolation of
a cDNA clone derived from a blood-borne non-A, non-B viral hepatitis genome. Science
244(4902):359–362
13. Lipkin WI, Travis GH, Carbone KM, Wilson MC (1990) Isolation and characterization of
Borna disease agent cDNA clones. Proc Natl Acad Sci 87(11):4184–4188
14. Nichol S, Spiropoulou C, Morzunov S, Rollin P, Ksiazek T (1990) Genetic identification of a
hantavirus associated with an outbreak of acute respiratory illness. Science 262:914–917
15. Relman DA, Loutit JS, Schmidt TM, Falkow S, Tompkins LS (1990) The agent of bacillary
angiomatosis: an approach to the identification of uncultured pathogens. New Eng J Med
323(23):1573–1580
16. Relman DA, Schmidt TM, MacDermott RP, Falkow S (1992) Identification of the uncultured
bacillus of Whipple’s disease. New Eng J Med 327(5):293–301
17. VandeWoude S, Richt JA, Zink MC, Rott R, Narayan O, Clements JE (1990) A borna virus
cDNA encoding a protein recognized by antibodies in humans with behavioral diseases. Science
250(4985):1278–1281
18. Soldan K, Barbara JAJ, Ramsay ME, Hall AJ (2003) Estimation of the risk of hepatitis B virus,
hepatitis C virus and human immunodeficiency virus infectious donations entering the blood
supply in England, 1993–2001. VoxSanguinis 84(4):274–286
19. Weigl BH, Boyle DS, Santos TDL, Peck RB, Steele MS (2009) Simplicity of use: a critical
feature for widespread adoption of diagnostic technologies in low-resource settings. Exp Rev
Med Dev 6(5):461–464
20. Singh RR, Ho D, Nilchi A, Gulak G, Yau P, Genov R (2010) A CMOS/thin-film fluores-
cence contact imaging microsystem for DNA analysis. IEEE Trans Circ Sys I: regular papers
57(5):1029–1038
21. Puren A, Gerlach JL, Weigl BH, Kelso DM, Domingo GJ (2010) Laboratory opera-
tions, specimen processing, and handling for viral load testing and surveillance. J Inf Dis
201(Supplement_1):S27–S36
22. Stevens WS, Marshall TM (2010) Challengea in implenting HIV load testing in South Africa.
J Inf Dis 201(Supplement_1):S78–S84
Development of Low Cost PCR Product Detection System … 933

23. Giljohann DA, Mirkin CA (2009) Drivers of biodiagnostic development. Nature

462(7272):461–464
24. Land KJ, Boeras DI, Chen XS, Ramsay AR, Peeling RW (2019) REASSURED diagnostics
to inform disease control strategies, strengthen health systems and improve patient outcomes.
Nat Micro 4(1):46–54
25. Dineva MA, Mahilum-Tapay L, Lee H (2007) Sample preparation: a challenge in the devel-
opment of point-of-care nucleic acid-based assays for resource-limited settings. Analyst
132(12):1193–1199
26. Patil Y, Thakur M (2020) Development and validation of indigenously developed DNA
detection system for M. tuberculosis. I.J.E.T 11(3)
27. Arora J, Kumar G, Verma AK, Bhalla M, Sarin R, Myneedu VP (2015) Utility of MPT64
antigen detection for rapid confirmation of Mycobacterium tuberculosis complex. J gloinf Dis
7(2):66
Determination of Viscous, Coulomb
and Particle Damping Response in SDOF
by Forced Oscillation

S. T. Bodare, S. D. Katekar, and Chetan Chaudhari

Abstract When a vibrating system is damped with more than one type of damping,
it is necessary to determine which of these types of damping are more effective to
control the resonant response. In such a case, identification of damping parameters
from the responses of a vibrating system becomes an important factor. Therefore
when the system is damped due to coulomb friction, viscous friction, it is necessary
to develop theoretical and experimental methods for identification of these damping
parameters from the responses of the vibrating system. As such, for identification
of coulomb and viscous friction (and also with particle damping) parameters, it is
proposed to develop methods irresponsible for the control of resonant response of
vibrating systems. The paper contains experimental setup and results about these
different types of mechanical vibrations.

Keywords Vibration · Types of vibration · Coulombs friction · Viscous friction ·

Particle vibration

1 Introduction

When oscillations occur about an equilibrium point, an observed phenomenon is

known as mechanical vibrations. These oscillations may be periodic, such as the
motion of a pendulum or random, such as the movement of a tire on a gravel road.
When a mechanical system is set in motion with an initial input and allowed to
vibrate freely then free vibration occurs. Examples of this type of vibration are
pulling a buried metal bar in cement and letting it vibrate freely, or hitting a drum
and letting it sound. In this type, the mechanical system under observation vibrates at
one or more of its natural frequencies, and damps down to motionlessness. When a
time-varying disturbance (load, displacements, or velocity) is applied to a mechanical
system, the phenomenon obtained is Forced Vibrations. The disturbances can occur

S. T. Bodare (B)
SKN Sinhgad College of Engineering, Korti, Pandharpur, Maharashtra, India
S. D. Katekar · C. Chaudhari
Mech Engineering Department, Sandip University, Nashik, Maharashtra, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 935
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_94
936 S. T. Bodare et al.

after particular time intervals and steady-state input, a transient input, or a random
input. The periodic input can be varied as a harmonic or a non-harmonic disturbance.
Example is loading machines shaking due to an imbalance. In case of transportation,
vibration caused by an engine or uneven road is an example too. The vibration of a
building due to an earthquake also causes vibrations. In the case of linear systems,
the frequency of the steady-state vibration response resulting from the application of
a periodic, harmonic input is equal to the frequency of the applied force or motion.
The vibrations are said to be damped, when the energy of a vibrating system is
gradually dissipated by friction, and other resistances. in such case of vibration the
amplitude of vibrations gradually decrease or change in frequency or intensity or
cease and the system rests in its equilibrium imposition. One of the examples of this
type of vibration is the vehicular suspension dampened by the shock absorbers. Den
Hartog first solved the forced response of a SDOF with both viscous and dry friction
damping. Viscous and coulomb friction is more important in different applications of
mechanical vibrations. These applications are like dynamics, and control problems.
Most friction dampers are used to reduce the resonant forces by providing sliding
contact between points experiencing relative motion due to vibration, thereby dissi-
pating resonant vibration energy. The above study is conducted to determine various
types of vibrations and effect of forced vibrations over mechanical elements such as
spring. Also study of viscous, coulomb and particle damping, and has been done.
This study will be useful in investigation of damping performance over a wide range
of excitation frequency and amplitude. Also, this study contributes in investigation
of frequency response analysis for various types of damping combinations.

2 Literature Review

Liang and Feeny [1] this gives a method for determination of Coulomb and Viscous
friction coefficients from responses of a harmonically excited dual-damped oscillator
with linear stiffness. The method of identification has been depended on existing
analytical solutions of non-sticking responses excited near resonance.
M.R. Duncan [2] have investigated the performance of a single particle vertical
impact damper over a range of forcing oscillation amplitudes and frequencies, mass
ratios, structural damping ratios, impact damper lid heights, and damper/structure
coefficients of restitution. Previous studies have not examined such a large parameter
space and instead have focused primarily on conditions at the structure’s natural
frequency and periodic particle trajectories.
Levitan [3] analyzed the motion of a system (refer Fig. 1) with harmonic displace-
ment of the base as shown in figure. The friction forces in the model proposed by
him are in between the base and the mass. Also in this paper, analytical solution
for the response of the support-excited system has been presented. The solution
to the equation of motion has been developed through the application of a Fourier
series to represent the frictional force opposing the relative motion between mass
and supporting structure.
Determination of Viscous, Coulomb and Particle Damping … 937

Fig. 1 System analyzed by

Levitan

Den Hartog [4] has presented an exact solution for the steady-state vibration of a
harmonically excited oscillator damped by combined dry and viscous friction (refer
Fig. 2). The system, as shown in figure, it consists of a forced excited mass m with
friction forces acting in between ground and them. The quite a few experimental
tests to verify solutions have been performed to find out the forced response of a
single-degree-of-freedom system with both viscous and dry friction damping.
Hundal [5] has done experimental study a base-excitation frictional oscillator
in which close form analytical solutions of the equation of motion were obtained.
The consequences of these experiments have been unfilled in non-dimensional form
as magnification factors in opposition to frequency ratios as functions of Viscous
and Coulomb friction parameters. It has been shown that the mass motion may
be continuous or it has a single stop during each cycle, depending upon system
parameters. The response of a single degree of freedom spring-mass system with
Viscous and Coulomb friction, with harmonic base excitation, has been determined.
Mao et al. [6] have examined the characteristics of Particle damping with respect
to a simple single-mass impact damper and a dry-friction damper. The analysis of

Fig. 2 System analyzed by

Den Hartog
938 S. T. Bodare et al.

the damping uniqueness of particle vibration dampers based on the 3D discrete

element method has been carried out. An incompetent procedure for the dynamics
of a large number of particles without undue computing complexity has been devel-
oped. Particle damping is a derivative of single-mass impact damper that has been
thoroughly studied on the influence of mass ratio, particle size, particle/slot clear-
ance, excitation levels, and direction of excitation. The distribution and arrangement
of the multiple particle dampers on the structure usually have a significant impact on
the damping effect of the dampers.
Stanway et al. [7] have anticipated a nonlinear least squares estimator scheme
which involves the online solutions to determine the parameters of viscous damping
and Coulomb from the harmonic response of a vibrant system.

3 Methodology

3.1 Identification of Coulomb and Viscous Friction Damping

in a Harmonic Base Excited SDOF System

When the vibrating system has both viscous and coulomb friction damping, it is
observed that the coulomb friction has noticeable effect on the resonant amplitude.
However, it is not clear from the steady state response analysis how much is the
contribution of the viscous friction in the system towards the reduction of resonant
amplitude [8, 9]. In more than few circumstances, such as in control systems, it is
necessary to know the content of the viscous and coulomb friction damping. in such
case, the identification of these two types of friction damping in a vibrating system
will play a significant role when the method used to control the resonant amplitude
of a system is the use of damping.

3.2 Forced Response and Friction Identification

A mass-spring-damper system on a rigid surface with the coulomb friction law, a

mass-spring-damper system on a rigid surface can have either stable pure-sliding or
stick-slip motion when subjected to harmonic base excitations. A schematic diagram
depicting the linear SDOF vibrating system with viscous and coulomb friction and
subjected to base excitation is presented below [10].
Here both the viscous and Coulomb damper elements have been put in between
the mass and ground as shown in figure. The equation of motion can be written as
by using Newton’s law of motion as,
←− −−→ − →
m ẍ = − F(ẋ) − c ẋ − k[x − y(t)] (1)
Determination of Viscous, Coulomb and Particle Damping … 939

or

m ẍ + c ẋ + kx + F(x) = ky(t) (2)

where,
m, c, f (x) and k are the mass, viscous damping coefficient, Coulomb friction
force, and spring constant respectively, y(t) = Y0 cos(ωt + ∅) is the base excitation
motion.
One can write the Eq. (2) as,
c  
ẍ + ẋ + ωn2 x − x f = ωn2 Y0 cos(ωt + ∅) (3)
m
where,
x f = Fk0 denotes the equivalent “friction displacement” and mk = ωn2 denotes
un-damped natural frequency. The peak values of non-sticking response occur at the
turning points and the peak amplitude x0 of the response is given as,

    2
x0 xf 1 xf
= −G + −H 2 (4)
Y0 Y0 q2 Y0

The values of G, H, and q can be represented as, [1].


sinh(ξβπ ) − ξ/ 1 − ξ 2 sin(ωd π/ω)
G= (5)
cosh(ξβπ ) + cos(ωd π/ω)
 
β/ 1 − ξ 2 sin(ωd π/ω)
H= (6)
cosh(ξβπ ) + cos(ωd π/ω)

and,

 2   
ω ω 2
q= 1− + 2ξ (7)
ωn ωn

above expressions, ξ is the non-dimensional damping ratio, β = ω/ωn, and

In the 
ωd = ωn 1 − ξ 2 represents the frequency of damped oscillation. The G and H are
functions of ξ, ωn, and ω only.
The analytical input-output relation as given by Eq. 4 forms the basis of the
scheme of identification of viscous and Coulomb damping of a SDOF system. With
the assumption that the ξ is very small and ξ= 0 and = ωd , the value of function
H approaches to zero (H = 0). When the system is excited near the resonance such
that ω ≈ ωn ≈ ωd , it is reasonable to assume that H ≈ 0. (When ξ= 0) Therefore,
neglecting H at resonance, the input-output amplitude relationship in Eq. (4) reduces
940 S. T. Bodare et al.

approximately to, [1]

Y0
x0 = −Gx f + (8)
q

The above equation is nothing but equation of line of Y0 vs. X0 in which the slope
and intercept are q and Gx f respectively.
If the system is excited at two excitation levels with the same frequency (close
to the damped natural frequency) with input output pairs denoted as, (Y1 , X1 ), and
(Y2 , X2 ) one can write using Eq. (8) as,

Y1
x1 = −Gx f + (9)
q
Y2
x2 = −Gx f + (10)
q

Subtracting Eq. (9) from Eq. (10) and after rearranging yields,

Y2 − Y1
q= (11)
x 2 − x1

Also from Eqs. (9) and (10) one can write,

x1 Y2 − x2 Y1
xf = (12)
G(Y1 − Y2 )

An expression for ξ can be obtained by putting Eq. (7) into Eq. (11) with the
condition ω ≈ ωn ≈ ωd

Y2 − Y1
ξ= (13)
2(x2 − x1 )

And, G reduces to,

sinh(ξ π)
G≈ (Approximately) (14)
cosh(ξ π) − 1

When more than two input-output amplitudes pairs are measured, the intercept
land slope of Eq. (8) is estimated by a least-squares fit of the (Yi, xi) data. It should
be noted that implementing the identification process, the system is required to be
excited lat the resonance.
Determination of Viscous, Coulomb and Particle Damping … 941

3.3 Effect of Coulomb Friction Damping Parameter

on the Steady State Response of a SDOF System

To examine the dynamic characteristics of a servo mechanism and a machine tool

slide way considering behave more sensibly of the friction is important in practice.
The coulomb friction model takes for granted that the friction force is constant in
magnitude and is directed so as to resist relative movement of two surfaces in contact.
The amplitude of friction force is proportional to the normal connecting force with
a factor l μ l defined as the coefficient of friction. Mechanical vibration systems
with viscous damping land coulomb friction are of importance in the applications of
dynamics and control problems.

3.4 Particle Damping

Particle damping is a derivative of impact damping technology with additional advan-

tages. Particle dampers appreciably lessen the noise and impact forces produced by
an impact damper and are less sensitive to alters in the cavity dimensions or excita-
tion amplitude. The benefits of impact dampers are that these dampers are econom-
ical, simple designs that provide operative damping performance over a range of
accelerations and frequencies. In addition to this, impact dampers are vigorous and
can operate in environments that are too severe for other conventional damping
methods. Impact vibration dampers have been used in a spacious variety of applica-
tions including vibration lessening of cutting tools, turbine blades, television aerials,
structures, plates, tubing, and shafts.
Particle Vibration Damping is a mixture of impact damping and friction damping.
In a PVD, metal or ceramic particles or powders of tiny size (0.05 to l5 mm in
diameter) are placed inside cavities within or attached to the vibrating structure. Metal
particles of great density such as lead or tungsten provide great damping performance
due to indulgence of kinetic energy [11]. Particle Vibration Damping includes the
potential energy absorptions and dissipation through momentum exchange between
moving particles and vibrating walls, friction impact restitutions.

4 Experimental Set up

i. An experimental set up has been designed and developed to obtain input output
amplitude relationship for SDOF Harmonically excited system with Viscous,
Viscous and Coulomb friction, and Viscous, Coulomb friction and Particle
damping
ii. For this purpose, design and development of viscous fluid damper, dry friction
damper and particle damping system has been carried out.
942 S. T. Bodare et al.

Fig. 3 Experimental setup

iii. A schematic of the experimental test set-up is shown in the Fig. 3 of the
“C” frame. The frequency response curve of the system is obtained using
the accelerometer pick up with its necessary attendant equipment and a FFT
analyzer, over a small range of excitation frequency.

In this case, the SDOF spring mass system with different damping types, system
has been excited from 5 Hz to 12 Hz using electro dynamic exciter and the steady
state response X has been measured using an accelerometer and FFT analyzer. Only
in case of PVD the readings are taken with increase in the number of balls from no
ball to full of balls with regular intervals.

5 Theoretical Analysis

(a) The non-sticking response of a Single Degree Of Freedom (SDOF for the short)
system subjected to harmonic excitation with both Viscous and Dry friction
has been determined in the approach of Cheng and Zu.
(b) Also the necessary expressions for viscous friction and dry friction parameters
have been obtained by establishing the input output amplitude relationship
near resonance of the SDOF system with viscous and Coulomb damping and
subjected to harmonic mass excitation and using two input output amplitude
pairs the values of dry and viscous friction has been determined in the approach
of Liang and Feeny.
(c) Using more than two input output response pairs, the values of Viscous and
Coulomb friction damping parameters have been estimated using a least square
fit of input output response parameters.
Determination of Viscous, Coulomb and Particle Damping … 943

Table 1 Spring details

Parameters Spring stiffness 3776 N/m
Number of turns (N) 28
Wire diameter (d) 4 mm
Mean diameter (D) 28 mm
Outer diameter (D0 ) 30 mm
Inner diameter (Di ) 26 mm
Spring index (C) 7

5.1 Determination of System Damping (ξ S)

In this case, the SDOF spring mass system has been excited from 5 Hz to 12 Hz using
electro dynamic exciter and the steady state response X has been measured using an
accelerometer and FFT analyzer.
The value of system damping ratio ξs has been determined as,

f2 − f1
ξs =
2 × fn

5.2 Spring

The springs selected are helical compression with both ends squared and ground.
The dimensions of springs are obtained using standard formulae of design of helical
compression spring (refer Table 1). The material of springs is selected as high grade
spring steel.

6 Results

Using the experimental set up, the frequency response analysis has been carried out
for a harmonically based excited SDOF system. Table 2 shows Damping Rations of
Various Combinations
1. System with mass m and stiffness k and system damping (S).
2. System with m. k and system damping plus viscous damping (S + V).
3. System with m, k and system damping with coulomb damping (S + C).
4. System with m, k, and system damping S with Viscous Damping and Coulomb
damping (S + V+C).
5. System with m, k, coulombs damping, viscous damping, and particle damping
(S + V+C + P).
944 S. T. Bodare et al.

Table 2 Damping rations of various combinations

Type of damping fr (Hz) F1 f2 ξ
S 11.86 11.50 12.01 0.0215
S+C 11.86 11.72 12.10 0.0435
S + P(T:-No balls) 10.75 10.252 11.20 0.044
S + P(T:-25% Balls) 10.75 10.350 11.52 0.0544
S + P(T:-50% balls) 10.75 10.121 11.672 0.0721
S + P(T:-75% balls) 10.75 10.425 11.80 0.0641
S + P(T:-fully filled with balls) 10.75 9.95 11.92 0.0961
S+V 12 11.92 13.50 0.0658
S + V+C 12 11.78 12.92 0.0475
S + V+C + P (25%full balls) 12 11.99 13.557 0.0694
S + V+C + P (50%full balls) 12 11.45 13.952 0.1042
S + V+C + P (full balls) 12 11.20 14.02 0.1175

7 Conclusion

In viscous, coulomb friction & particle damping, the coulomb friction damping is
non-linear damping so it should be minimized, for that purpose we found coulomb
friction parameters & the result obtained are nearly the same in theoretical and exper-
imental method. The unique aspect of PID is that high damping has been achieved
by absorbing the kinetic energy of the structure as opposed to the more traditional
methods of damping that uses elastic, where strain energy stored in the structure
is converted to heat. From the above literature study, it is necessary to understand
contents of viscous damping, coulomb damping, and particle damping in a system.
The dominant damping identification in a system is useful to understand the behavior
of system.

References

1. Liang JW, Feeny BF (2004) Identifying coulomb and viscous friction in forced dual-damped
oscillators. ASME J Vib Acoustics 126:118–125
2. Steven E (2003) Olson: an analytical particle damping model. J Sound Vib 264:1155–1166
3. Duncan MR, Wassgren CR, Krousgrill CM (2005) The damping performance of a single
particle impact damper. J Sound Vib 286:123–144
4. Mao (2004) Simulation and characterization of particle damping in transient vibrations. ASME
J Vib Acoustics 126:202–211
5. Den Hartog JP (1931) Forced vibrations with combined coulomb and viscous friction. Trans
ASME 53:107–115
6. Hundal MS (1979) Response of a base excited system with coulomb and viscous friction. J
Sound Vib 64:371–378
Determination of Viscous, Coulomb and Particle Damping … 945

7. Levitan ES (1960) Forced oscillation of a spring-mass system having combined coulomb and
viscous damping. J Acoust Soc Am 32:1265–1269
8. Perls Thomas A, Sherrard Emile S (1956) Frequency response of second order systems with
combined coulomb and viscous damping. J Res Natl Bureau Standard 57:45–64
9. Liang JW, Feeny F (1999) The estimation of viscous and coulomb damping in harmonically
excited oscillators. Proceedings of DETC 99, Las Vegas
10. Elliott SJ, Tehrani MG, Langley RS (2005) Nonlinear damping and quasi-linear modelling
Phil. Trans R Soc A.37320140402
11. Duifhuis H (2012) 2012 Cochlear mechanics: introduction to a time domain analysis of the
nonlinear cochlea. Springer, New York
Assessment of Godavari River Water
Quality of Nanded City, Maharashtra,
India

P. R. Shaikh, Girish Deore, A. D. Pathare, D. V. Pathare, and R. S. Pawar

Abstract Water resources are sources of water that useful or potentially useful to
humans. Uses of water are including domestic, agricultural, industrial, recreational as
well as environmental activities. Virtually, all these human uses require fresh water.
The water samples from the Godavari River of the five sampling site are taken and
analyzed for the physicochemical parameters such as Colour, Odour, Temperature,
pH, Electrical Conductivity (EC), Total Solids (TS), Total Dissolved Solids (TDS),
Total Suspended Solids (TSS), Total Hardness (TH), Calcium (Ca), Magnesium
(Mg), Chloride (Cl), Alkalinity (TA), Dissolved Oxygen (DO), Chemical Oxygen
Demand (COD), Biological Oxygen Demand (BOD) and Turbidity, etc. parameters
in seven weeks. Godavari River water is one of the important sources of water in
Nanded city. As the Nanded is drought-prone area alteration within major or minor
in the characteristics of River water results in great attention of day to day life of
citizens of Nanded. There are five samples taken weeks of March and April months
of Godavari River to access the quality of water and the results are compared with
WHO standards.

Keywords River water · Quality · Nanded · Characteristics · Etc

P. R. Shaikh (B)
School of Earth Sciences, S.R.T.M. University, Nanded, Maharashtra, India
G. Deore
Matoshri College of Engineering, Nanded, Maharashtra, India
A. D. Pathare
Pravara Rural Engineering College, Loni, Maharashtra, India
D. V. Pathare
Padmashri DVVP Institute of Technology and Engineering (Polytechnic) College, Pravaranagar,
Loni, Maharashtra, India
R. S. Pawar
SVERIs College of Engineering Pandharpur, Solapur, Maharashtra, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 947
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_95
948 P. R. Shaikh et al.

1 Introduction

Water is well known as a natural solvent. Before, it reaches the consumer’s tap,
it comes into contact with different pollutants, including organic and inorganic
matter, chemicals, and other contaminants which pollute them. Many public water
systems, treated water with chlorine to kill disease-producing contaminants that may
be present in the water. Although disinfection is an important step in the treatment of
potable water, the taste and odor of chlorine is objectionable. The disinfectants that
are used to prevent disease can create byproducts which may pose significant health
risks. The water cycle explains interactions between the atmosphere, hydrosphere,
and lithosphere. The water or hydrologic cycle is a major driving force on our planet.
A river is a natural watercourse, usually freshwater, flowing towards an ocean, a lake,
a sea, or another river. In a few cases, a river simply flows into the ground or dries
up completely at the end of its course, and does not reach another body of water [1].

1.1 Background of the Study

In recent years, because of continuous growth of population, rapid industrialization

and the technologies involving waste disposals, the rate of discharge of the pollutants
into the environment is far higher than the rate of their purification. The implications
of deteriorating quality of the receiving waters are considerable both in the imme-
diate situation and over the longer term. In this context, water quality assessment is
critical for pollution control and the protection of surface and ground waters. In India,
disposal of untreated domestic sewage from cities, towns and villages is the major
source of pollution of surface water bodies leading to the outbreak of water borne
diseases. Biodegradable organic matter is the contaminant of concern for dissolved
oxygen concentration which is the principal indicator of pollution of surface water.
According to world health organization (WHO), about 80% of water pollution in
developing countries like India is caused by domestic wastes [2].

2 Study Area

The district is situated on Maharashtra-Karnataka-Andhra Pradesh boundary. Nanded

is the second largest city in the Marathwada region of Maharashtra state. It is an
important holy place for the Sikh faith and is famous for the Hazur Sahib Gurudwara.
It is the district headquarters once very famous as district of Sanskrit poets. Nanded
is a town of great antiquity and famous for Muslim Sufi shrines. This district of great
antiquity is important for the Hindu faith and is known for the Renukadevi temple at
Mahur. The official languages are Marathi, Hindi, Urdu and Panjabi. The sampling
locations plotted in the map of Godavari River in Nanded City (Fig. 1) [3, 4].
Assessment of Godavari River Water Quality of Nanded City … 949
Nanded City Map
0
700 80 900 City Boundary 1:10000

0
40

Nanded Taluka
India Nanded District
19011'0"

300

Maharashtra 19009'0"

200

0 0 200 Km
0 1000 Km 10 190 07'0"

Scale
Nanded City
Scale
0 2 Km

77017'0" 0
77 19'0" 77021'0"

77o18' 20" 77 o19' 18"

19 o09' 17"

Shree Nagar

Vazirabad

Old Mondha
4

3
1

Index
5 Sample Number
19 o08' 23"

Fig. 1 Location map of the study area

3 Materials and Methods

For the study of River water quality we have select the Godavari River at Nanded city
because the growth and development of the city affect the water quality of this River.
As Nanded city is situated in Marathwada region of Maharashtra (draught zone).
Population of this city is mainly dependent on the River water to sustain. Therefore
any major alteration in the physico-chemical characteristics of River water of this
region affects the day to day activities of peoples in this region adversely.
950 P. R. Shaikh et al.

3.1 Sample Collection

There are 5 locations selected as per the pollution sources and total 7 times (Weeks of
March and April Months) water samples were collected from the study of River water
quality. Samples are collected in the polythene bottles. The collected samples are
transfer to the laboratory for further analysis. The physico-chemical parameters such
as Colour, Odour, Temperature, pH, Electrical Conductivity (EC), Total Solids (TS),
Total Dissolved Solids (TDS), Total Suspended Solids (TSS), Total Hardness (TH),
Calcium (Ca), Magnesium (Mg), Chloride (Cl), Alkalinity (TA), Dissolved Oxygen
(DO), Chemical Oxygen Demand (COD), Biological Oxygen Demand (BOD) and
Turbidity, etc. are analysed. The physico-chemical parameters are checked by the
titrometric method [3, 5].

4 Result and Discussion

The water samples from the Godavari River of the five sampling site are taken. The
samples are filtered with the help of filter paper and analyzed for the physico-chemical
parameters such as Colour, Odour, Temperature, pH, Electrical Conductivity (EC),
Total Solids (TS), Total Dissolved Solids (TDS), Total Suspended Solids (TSS), Total
Hardness (TH), Calcium (Ca), Magnesium (Mg), Chloride (Cl), Alkalinity (TA),
Dissolved Oxygen (DO), Chemical Oxygen Demand (COD), Biological Oxygen
Demand (BOD) and Turbidity, etc. parameters in seven weeks as shown in the Table 1.
The all samples show the greenish colour, which is the indicator of algal blooms.
The odour is objectionable, because of the mixing of sewage in the River water. The
temperature is varies from 35 to 38 °C. It is increases because of the summer season.

4.1 pH

The pH of samples in seven weeks of March and April months is ranges from 7.4
to 7.8. The pH value of all the weeks of two months falls in Neutral category. The
average and standard deviation of pH is 7.52 and 0.1 respectively. The pH of sample
no. 1 is higher as compare to the other samples of first three weeks ref Fig. 2.

4.2 Electrical Conductivity

The EC of samples in seven weeks of March and April months is ranges from 1952
to 2178 µS/cm. The average and standard deviation of EC is 2080.9 and 77.12
respectively. The EC of sample no. 1 is higher as compare to the other samples and
Assessment of Godavari River Water Quality of Nanded City … 951

Table 1 Physico-chemical parameters of Godavari river water samples in the weeks of March and
April months
Parameters Unit Min Max Avg StDev
pH – 7.4 7.8 7.52 0.1
EC µS/cm 1952 2178 2080.9 77.12
TS mg/l 2253 3394 2743.2 520.6
TSS 1000 2000 1411.4 489.76
TDS 1249 1394 1331.4 49.35
TH 154 302 191.2 33.8
Ca 36.07 84.17 47.18 11.31
Mg 16.52 25.33 20.23 2.4
Cl 99.4 205.9 148.97 21.89
TA 130 150 137.57 5.86
DO 3.2 5 3.81 0.43
COD 78 128 104.71 12.25
BOD 31 51 41.66 4.84
Turbidity NTU 220 663 370.09 81

7.9 Graphical Representation of pH

7.8

7.7
7.6
pH

7.5
7.4

7.3

7.2
12345123451234512345123451234512345

01-Mar 08-Mar 15-Mar 22-Mar 29-Mar 05-Apr 12-Apr

Sampling Location with Date

Fig. 2 Graphical representation of pH

952 P. R. Shaikh et al.

2200 Graphical Representation of EC

2100
µS/cm

2000

1900
12345123451234512345123451234512345

01-Mar 08-Mar 15-Mar 22-Mar 29-Mar 05-Apr 12-Apr

Sampling Location with Date

Fig. 3 Electrical conductivity

lower in sample no. 5 of all the sampling weeks. From the Fig. 3 it is seen that the
EC is gradually increased during the sampling periods.

4.3 Solids (TS, TSS and TDS)

The TS of samples in seven weeks of March and April months is ranges from 2253 to
3394 mg/l. The average and standard deviation of TS is 2743.2 and 520.6 respectively.
The TSS of samples in seven weeks of March and April months is ranges from 1000
to 2000 mg/l. The average and standard deviation of TSS is 1411.4 and 489.76
respectively. The TDS of samples in seven weeks of March and April months is
ranges from 1249 to 1394 mg/l. The average and standard deviation of TDS is 1331.8
and 49.35 respectively. All the samples are falls in above the highest desirable limit,
but below the maximum permissible limit given by WHO [6, 7]. From the Fig. 4, 5
and 6 it is seen that the TDS is gradually increased during the sampling periods.

4.4 Total Hardness (Th)

The TH of samples in seven weeks of March and April months is ranges from 154 to
302 mg/l. The average and standard deviation of TH is 191.2 and 33.8, respectively.
The all samples falls hard water category. According to WHO [6, 7], all samples falls
Assessment of Godavari River Water Quality of Nanded City … 953

3600 Graphical Representation of TS

3100
mg/l

2600

2100

1600
12345123451234512345123451234512345

01-Mar 08-Mar 15-Mar 22-Mar 29-Mar 05-Apr 12-Apr

Sampling Location with Date

Fig. 4 Graphical representation of TS

2500 Graphical Representation of TSS

2000

1500
mg/l

1000

500

0
12345123451234512345123451234512345

01-Mar 08-Mar 15-Mar 22-Mar 29-Mar 05-Apr 12-Apr

Sampling Location with Date

Fig. 5 Graphical representation of TSS

above the highest desirable limit, but below the maximum permissible limit except
one sample. It can be seen that the river water samples are more total hardness
content because of mixing of sewage. From the Fig. 7 it is seen that the TH is
954 P. R. Shaikh et al.

1500
Graphical Representation of TDS

1400
mg/l

1300

1200
12345123451234512345123451234512345

01-Mar 08-Mar 15-Mar 22-Mar 29-Mar 05-Apr 12-Apr

Sampling Location with Date

Fig. 6 Graphical representation of TDS

320
Graphical Representation of TH
mg/l

220

120
12345123451234512345123451234512345

01-Mar 08-Mar 15-Mar 22-Mar 29-Mar 05-Apr 12-Apr

Sampling Location with Date

Fig. 7 Graphical representation of TH

gradually increased during the sampling periods. The sample no. 4 TH content is
higher in the 8 March, 15 March and 22 March sampling period and in the sample
no 2, TH content is higher in next all sampling periods.
Assessment of Godavari River Water Quality of Nanded City … 955

100
Graphical Representation of Ca
mg/l

50

0
12345123451234512 345123451234512345

01-Mar 08-Mar 15-Mar 22-Mar 29-Mar 05-Apr 12-Apr

Sampling Location with Date

Fig. 8 Graphical representation of Ca

4.5 Calcium (Ca)

The Ca of samples in seven weeks of March and April months is ranges from 36.07 to
84.17 mg/l. The average and standard deviation of Ca is 47.18 and 11.31 respectively.
All samples falls below the highest desirable limit, only two samples falls above the
highest desirable limit but below the maximum permissible limit given by WHO [6,
7]. From the Fig. 8 it is seen that the Ca is gradually increased during the sampling
periods. The sample no. 4 Ca content is higher in the 8 March, 15 March and 22
March sampling period and in the sample no 2, Ca content is higher in next all
sampling periods.

4.6 Magnesium (Mg)

The Mg of samples in seven weeks of March and April months is ranges from 16.52 to
25.33 mg/l. The average and standard deviation of Mg is 20.23 and 2.4 respectively.
All samples falls below the highest desirable limit given by WHO [6, 7]. From the
Fig. 9 it is seen that the Mg is gradually increased during the sampling periods. The
sample no. 4 Mg content is higher in the 8 March, 15 March and 22 March sampling
period and in the sample no 2, Mg content is higher in next all sampling periods.
956 P. R. Shaikh et al.

30
Graphical Representation of Mg
mg/l

20

10
12 3451 2345 1234 512 3451 2345 1234 5123 45

01-Mar 08-Mar 15-Mar 22-Mar 29-Mar 05-Apr 12-Apr

Sampling Location with Date

Fig. 9 Graphical representation of Mg

4.7 Chlorides

The Cl of samples in seven weeks of March and April months is ranges from 99.4 to
205.9 mg/l. The average and standard deviation of Cl is 148.97 and 21.89 respectively.
According to WHO, all samples falls below the highest desirable limit except one
sample. It can be seen that the river water samples are more chloride content because
of mixing of sewage. From the Fig. 10 it is seen that the Cl is gradually increased
during the sampling periods. The sample no. 5 Cl content is higher except first two
sampling periods.

4.8 Total Alkalinity

The TA of samples in seven weeks of March and April months is ranges from 130 to
150 mg/l. The average and standard deviation of TA is 137.57 and 5.86 respectively.
From the Fig. 11 it is seen that the TA is gradually decreased during the sampling
periods.
Assessment of Godavari River Water Quality of Nanded City … 957

230 Graphical Representation of Cl

180
mg/l

130

80
123451234 51234512345123451 234512345

01-Mar 08-Mar 15-Mar 22-Mar 29-Mar 05-Apr 12-Apr

Sampling Location with Date

Fig. 10 Graphical representation of Cl

160 Graphical Representation of TA

mg/l

140

120
123451 23451234512 345123451234 512345

01-Mar 08-Mar 15-Mar 22-Mar 29-Mar 05-Apr 12-Apr

Sampling Location with Date

Fig. 11 Graphical representation of TA

4.9 Dissolved Oxygen (Do)

The DO of samples in seven weeks of March and April months is ranges from 3.2
to 5 mg/l. The average and standard deviation of DO is 3.81 and 0.43 respectively.
958 P. R. Shaikh et al.

6 Graphical Representation of DO
mg/l

2
1 23 45 12 34 51 23 45 12 3 45 12 34 51 23 45 12 34 5

01-Mar 08-Mar 15-Mar 22-Mar 29-Mar 05-Apr 12-Apr

Sampling Location with Date

Fig. 12 Graphical representation of DO

From the Fig. 12 it is seen that the DO is gradually decreased during the sampling
periods. The sample no. 5 DO content is higher in the first 4 sampling period and in
sample no 3, DO content is higher in next all sampling periods.

4.10 Chemical Oxygen Demand (Cod)

The COD of samples in seven weeks of March and April months is ranges from 78
to 128 mg/l. The average and standard deviation of COD is 104.71 and 12.25 respec-
tively. According to WHO [6, 7], all samples falls below the maximum permissible
limit. From the Fig. 13 it is seen that the COD is gradually increased during the
sampling periods.

4.11 Biological Oxygen Demand (Bod)

The BOD of samples in seven weeks of March and April months is ranges from 31 to
51 mg/l. The average and standard deviation of BOD is 41.66 and 4.84 respectively.
According to WHO [6, 7], all samples falls below the maximum permissible limit.
From the Fig. 14 it is seen that the BOD is gradually increased during the sampling
periods.
Assessment of Godavari River Water Quality of Nanded City … 959

150
Graphical Representation of COD
mg/l

100

50
123 4512345 1234512 3451234 5123451 2345

01-Mar 08-Mar 15-Mar 22-Mar 29-Mar 05-Apr 12-Apr

Sampling Location with Date

Fig. 13 Graphical representation of COD

55 Graphical Representation of BOD

45
mg/l

35

25
1 234 512 34 512 345 12 345 123 451 23 451 234 5

01-Mar 08-Mar 15-Mar 22-Mar 29-Mar 05-Apr 12-Apr

Sampling Location with Date

Fig. 14 Graphical representation of BOD

960 P. R. Shaikh et al.

700 Graphical Representation of Turbidity

600

500
NTU

400

300

200
12 34512 34512 34512 34512 34512 34512 345

01-Mar 08-Mar 15-Mar 22-Mar 29-Mar 05-Apr 12-Apr

Sampling Location with Date

Fig. 15 Graphical representation of Turbidity

4.12 Turbidity

The Turbidity of samples in seven weeks of March and April months is ranges from
220 to 663 mg/l. The average and standard deviation of Turbidity is 370.09 and 81
respectively ref Fig. 15.

5 Conclusion

Godavari River water is one of the important sources of water in Nanded city. As the
Nanded is drought prone area alteration within major or minor in the characteristics
of River water results in great attention of day to day life of citizens of Nanded. There
are five samples taken weeks of March and April months of Godavari River to access
the quality of water and the results are compared with WHO standards. All the values
are found within limit (except EC, TDS, and TH). The Chloride value is found little
higher. This indicates that River water is contaminated with sewage pollution, fecal
matter. Probable solutions are suggested to avoid the further contamination.
1. Analysis of physico- chemical parameters of water shows that all the parameters
are within (except EC, TDS, and TH) Standard limit provided by WHO.
2. River water of this region is contaminated with sewage waste
3. Disposal site for sewage waste should be out of the premises of the city to avoid
further contamination of River water of the area.
Assessment of Godavari River Water Quality of Nanded City … 961

References

1. Verma S (2009) Seasonal variation of water quality in Betwa river at Bundelkhand region, India.
Glob J Environ Res 3(3):164–168
2. Srivastava A, Kr R, Gupta V, Agarwal G, Srivastava S, Singh I (2011) Water quality assess-
ment of Ramganga river at Moradabad by physico-chemical parameters analysis. VSRD-TNTJ
2(3):119–127
3. Panaskar DB, Wagh VM, Pawar RS (2014) Assessment of groundwater quality for suitability
of domestic and irrigation from Nanded Tehsil, Maharashtra, India. SRTMUs J Sci 3(2):71–83
4. Pawar RS, Panaskar DB (2014) Characterisation of groundwater in relation to domestic and
agricultural purposs, Solapur Industrial Belt, Maharashtra, India. J Environ Res Dev (JERAD)
9(01):102–112
5. APHA, AWWA, WEF (1985) Standard methods for the examination of water and wastewater,
16th ed. American Public Health Association. Washington DC, pp 6–187
6. World Health Organization (WHO) (1998) International standards for drinking water quality-
geneva, WHO
7. World Health Organization (WHO) (2002) Guideline for Drinking Water Quality, 2nd ed. Health
criteria and other supporting information. World Health Organization, Geneva, pp 940–949
Groundwater Quality Assessment
in an Around Thermal Power Plants
in Central India

V. U. Deshmukh, D. B. Panaskar, P. R. Pujari, and R. S. Pawar

Abstract More than hundred million tons of coal fly ash is produced annually in
India from combustion of coal in power plants. It is expected that about 150 million
tons of coal ash will be produced due to burning of coal in power plants by the year
of 2015. This will require about 30,000 hectare of land for the disposal of ash. One
of the biggest problems due to disposal of large quantities of coal ash is the possible
leaching of different hazardous pollutants, including TDS, fluoride and sulphate. The
present study investigated the leaching of soluble compounds in ash ponds and its
impact on the groundwater quality in the sub-watershed around the ash ponds in the
vicinity of Koradi and Khaperkheda near Nagpur in Maharashtra, India. A network of
twenty three observation wells set up for monitoring of water level and groundwater
quality for major cations and anions during pre-monsoon season, 2010. The results
indicate that the SO4 concentration is very high (>1000 mg/L) in the samples which
is much closer to the ash ponds. The TDS and Fluoride concentration is also elevated
with respect to BIS standards in few samples.

Keywords Coal fly ash · Power plants · Ash ponds · Groundwater quality · Koradi

1 Introduction

In India, groundwater is the principal source for drinking, irrigation and domestic as
well as industrial requirements especially in the rural areas, even in peri urban areas,
where centralized water supply is not available; groundwater is used for drinking and
domestic purposes. Hence, Protection of the groundwater resource assumes impor-
tance as it constitutes 96% of the available freshwater globally [1]. The huge amount

V. U. Deshmukh (B) · D. B. Panaskar

School of Earth Science, SRTM University, Nanded, Maharashtra, India
P. R. Pujari
Water Technology and Management Division, CSIR-NEERI, Nagpur, India
R. S. Pawar
SVERIs College of Engineering, Pandharpur, Maharashtra, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 963
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_96
964 V. U. Deshmukh et al.

of coal consumption for electricity generation used in thermal power plants in most
part of India. Due to this, lot of waste in the form of ash is generated and is often
mixed with water and disposed in ponds in the form of slurry. The ash is rich in the
major constituents namely, CaO, MgO, Na2 O, K2 O, SiO2 , Fe2 O3 , MnO, TiO2 and
P2 O5 . Besides, the ash is also rich in trace elements like Cu, Pb, Zn, Ni, Co etc.
[2]. The ponds do not have a secured liner to prevent the seepage of contaminants
to the aquifer. Hence, it is essential to study the likely threat to groundwater regime
from such sources of groundwater pollution. Among the various sources of ground-
water pollution, the ash ponds of the Thermal power plants have been one of the
major concerns recently both on a global as well as national level. The soil and water
contamination from ash disposal ponds has been a subject of much research the world
over [3–5]. The contamination is a matter of much concern as huge amounts of ash
are generated on a daily basis and they are continuously disposed on the ground. The
leachate can ultimately get into the aquifer and contaminate it. The present work is an
attempt to study the impact of the ash ponds on the groundwater resources. It seeks to
adopt a holistic approach in examining the groundwater contamination vis-à-vis the
ash ponds. The study has been undertaken in a sub-watershed surrounding the ash
ponds of Koradi and Khaparkheda power plants near Nagpur in Maharashtra. The
groundwater quality is assessed vis-à-vis the BIS [6] (Bureau of Indian Standards)
guidelines for standards of drinking water. The focus is on the major cations /anions
and fluoride.

2 Study Area

The study area (57.8 km2 ), covers the ash ponds of the Koradi and Khaparkheda
Power plants ref Fig. 1. It is located approximately 15 km from the Nagpur city
and it is bounded by longitudes 79 ° 05´ E to 79 ° 15´ E and latitudes 21 ° 10´
N to 21 ° 17´ N (SoI Toposheet No. 55 O/3 and 55 O/4 on 1:50,000 scale). The
region experiences an arid climate with temperature varying from 10º to 46 °C. The
average annual rainfall is approximately 1,100 mm [7]. The area receives rainfall
from the southwest monsoon during the period June to September. The ash from
Koradi Thermal Power Plant is disposed in Pond 1 and Pond 2, whereas Pond 3
receives the ash from Khaparkheda power plant. The ash from the power plant is
generated in two different modes. It is disposed as fly ash in the atmosphere and
as ash slurry in ash ponds. The ash slurry is made by mixing the bottom ash with
water and transported to the pond through pipelines. Pond 1 and Pond 2 are spread
over 254 ha whereas Pond 3 is spread over 282 ha. The Koradi Power Plant and the
Khaparkheda Power Plant are in operation since 1974 and 1989 respectively. The
ponds do not have liners at the bottom as well as the flanks.
Geologically, the area comprises formations of Sausar group (quartzite’s, marbles,
schist’s and gneisses), sedimentary formations (sandstone, shale and clay) equivalent
of Kamthi stage of Gondwana super group [8, 9]. The southern part of the study area
is covered by the Archeans whereas the northern part is covered by the Gondwanas.
Groundwater Quality Assessment in an Around Thermal Power … 965

Fig. 1 Location of the study area

3 Materials and Methods

A network of twenty three observation wells both dug wells as well as hand pumps
were set up in the study area during pre monsoon 2010 season. The pre-monsoon
sampling was done during May months and analyzed for major cations (Na, K, Ca,
Mg), anions (CO3, HCO3, Cl, SO4), fluoride following standard protocol [10] and pH
and temperature were measured in the field. The samples were filtered by Whatman
filter paper (no. 42) prior to the analysis in the laboratory. Water level was measured
with the help of an electric contact-gauze (KL 010 A) with a buzzer. The reduced
level of the observation wells has been measured by surveyors. The reduced level
of the observation wells has been used to obtain the groundwater level (above mean
sea level). Subsequently, the groundwater level (amsl) contours were prepared for
identifying the groundwater flow direction in the study area. The QA/QC was ensured
by analyzing the samples in duplicate manner for effective accuracy of results and
maintaining various quality criteria like analysis of externally supplied standards,
analysis of reagent blanks and calibration with standards during the laboratory. As
a minimum requirement of the accuracy in the analytical instruments used in the
analysis, three different dilutions of the standards were measured when an analysis
is initiated. Subsequently, the calibration curve was plotted by analyzing one or
more standards within the linear range. The concentration of the specific parameters
was obtained from the calibration curve against the corresponding wavelength. The
majority of analyzed groundwater samples had ionic charge balance <2–3%.
966 V. U. Deshmukh et al.

4 Result and Discussion

The groundwater table varied between 3.4 m to 15.7 m (bgl) in pre-monsoon season,
2010. A groundwater level contour (amsl) indicates ref Fig. 2 that the groundwater
flows towards the south-east direction.
A close look towards groundwater quality ref Table 1 showed that the sulphate is
one of the problematic parameter contaminating the groundwater in the study area.
The sulphate concentration varied in between 6 – 943 mg/L (pre 2010). It is indicated
that the higher sulphate concentration (>400 mg/L) continuously observed in KD-8,
KD-9 and KD-10 which immediate downstream of the ash ponds ref Fig. 3. In 2010,
TDS concentration varied in range of 127 – 2322 mg/L for pre-monsoon season. The
fluoride concentration continuously increasing in KD-23 which is much closer to
the fly ash carrying stream. The literature said that, due to the ash pond effect there
is chance of fluoride contamination in the groundwater [11]. The piper diagram ref
Fig. 4, reveal that the distribution pattern of anions and cations for pre-monsoon and
post monsoon seasons of the study period. The chloride and sulphate anions play
major contribution followed by sodium and potassium in pre monsoon season 2010.
Chemical characteristics ref Table 2 and trace elements ref Table 3. of pond
ash analyzed by XRF, the results showed that the major constituent in the pond
ash samples is Si, Ca, Al, K, Na, Mg, S, Ti and P as prominent elements in the

KD-1
KD-2 KD-3
KD-4
KD-5

21.24 KD-21

KD-11
KD-10
Latitude

KD-9
KD-23 KD-20
KD-19
21.22 KD-7 KD-22 KD-15 KD-18
KD-8 KD-6 KD-17

KD-12
21.2 KD-13

Pre-10

79.1 79.12 79.14 79.16 79.18 79.2 79.22

Longitude

Fig. 2 Water table (amsl-m) contour plot of pre monsoon season, 2010
Groundwater Quality Assessment in an Around Thermal Power … 967

Table 1 Physico-chemical parameters of groundwater quality of study area during pre-monsoon

season, 2010
Sr. No. Parameters Range (mg/L) Average BIS Limit For No. of samples
with sample (mg/L) drinking water exceeding bis
number (mg/L) permissible limit
1 Total Dissolve 127–2322 1016 2000 3
Solid (TDS) (KD3)–(KD10)
2 Calcium (Ca++ ) 48–285 124 200 2
(KD23)–(KD10)
3 Magnesium 29–140 78 100 6
(Mg++ ) (KD19)–(KD18)
4 Chloride (Cl− ) 40–760 261 1000 0
(KD20)–(KD18)
5 Sulphate 6–943 243 400 3
(SO4 − ) (KD20)–(KD9)
6 Nitrate (NO3 − ) 0.3–90 38 45 6
(KD20)–(KD8)
7 Fluoride (F− ) 0.001–2.3 1 1.5 1
(KD15)–(KD23)

Fig. 3 Contour plot of sulphate (mg/l) of pre-monsoon season, 2010

form of oxides, silicates and alumino-silicates. A result indicates that the pond ash
comes into the Class F category depending on the percentage of SiO2 (53.38%)
followed by Al2 O3 (25.70%) and Fe2 O3 (5.32%) respectively. The concentration of
Mg (2800 mg/kg) is increased in the respective pond ash followed by P (850 mg/kg),
Mn (310 mg/kg) and Cr (220 mg/kg).
968 V. U. Deshmukh et al.

Fig. 4 Piper diagram of pre-monsoon season, 2010. (Green and blue colour represent dug wells
and hand pumps respectively)

Table 2 Chemical characteristics of pond ash

Samples ID SiO2 Al2 O3 TiO2 Na2 O SO3 Fe2 O3
%
Koradi Ash-Pond 53.38 25.7 1.82 0.55 ND 5.32

Table 3 Heavy metals concentration of pond ash

Samples ID Co Cr Cu Mg Mn Ni P V Zn
mg/kg
Koradi ash-pond 47 220 155 2800 310 55 850 ND 105
Groundwater Quality Assessment in an Around Thermal Power … 969

5 Conclusion

Combustion of coals in thermal power plants is one of the major sources of envi-
ronmental pollution due to generation of huge amount of ashes which are disposed
of in large ponds in the vicinity of thermal power plants. During the study period
it was observed that parameters namely TDS, Sulfate and fluoride concentration
were significant in few samples. The major concern is about SO4 concentration
which increased in the few samples in the downstream of ash ponds. The severity of
fluoride concentration with respect to groundwater quality is also alarming around
the ash ponds. It is desired that regular monitoring in the critical season i.e. pre-
monsoon season need to be carried out to generate long term data for trend analysis
and compliance measures.

References

1. Shiklomanov IA, Rodda JC (2003) World water resources at the beginning of the twenty-first
century, UNESCO/IHP int. Cambridge University Press, Hydrology Series, Cambridge, UK,
p 2003
2. Deshmukh AN, Shah KC, Shrivastava BN (1994) Impact of rainy season (monsoon) on fly
ash dispersal—a case study of Koradi thermal power plant. Maharashtra Gondwana Geol Mag
8:1–17
3. Theis TL, Westrick JD, Hsu CL, Marley JJ (1978) Field investigations of trace metals in
groundwater from fly ash disposal. J Water Pollut Control Fed 50:2457–2469
4. Theis TL, Richter RO (1979) Chemical speciation of heavy metals in power plant ash pond
leachate. Environ Sci Technol 13:219–224
5. Theis TL, Gardner KH (1990) Environmental assessment of ash disposal. Crit Rev Environ
Control 20:21–42
6. BIS (1991) Drinking water specifications IS: 10500. Bureau of Indian standards, New Delhi
7. Central Ground Water Board (CGWB) (2009) Report on district groundwater management
studies in parts of Nagpur district, Maharashtra. Central Region, Nagpur
8. Suryanarayana K (1965) A new lower gondwana basin between Nagpur and Kampee. Curr Sci
34:610–621
9. Suryanarayana K (1968) The nara-pawangaon extension of the bhokara basin between Nagpur
and Kamptee. Ind Geo Sci Asso, Hyderabad 8:67–78
10. APHA (2005) Standard methods for analysis of water and waste water, 21st (edn) Washington,
DC, 2005
11. Deshmukh AN, Shah KC, Appulingam S (1995) Coal Ash: a source of fluoride contamination,
a case study of koradi thermal power station, district Nagpur, Maharashtra. Gondwana Geol
Mag. 9:21–29
Micro, Nano Manufacturing, Fabrication
and Related Applications
XML Based Feature Extraction
and Process Sequence Selection
in a CAPP System for Micro Parts

G. Gogulraj and S. P. Leokumar

Abstract In this work, a process sequence module in a CAPP system using feature-
based modeling (FBM) for prismatic and axis-symmetric parts is developed. It
extracts feature information in Extensible Markup Language (XML) format and
thereon generates the process sequence for efficient manufacturing of the micro
parts. The system consists of two components: (1) development of FBM through
design by feature (DBF) approach and automatically extract the feature information
and verify micro machine tool capability feature details and (2) Process sequence
determination to produce the micro features in prismatic and axis-symmetric micro
parts using a knowledge-based system (KBS) approach. A CAPP system for pris-
matic and axis-symmetric parts can be realized only with the incorporation of many
micro features and the development of other activities in process planning.

Keywords Feature extraction · Knowledge-based system · Micro parts · Process

planning

1 Introduction

To manufacture the micro parts, a set of instructions that involve design specification
are given and this task is termed process planning. CAPP is the tie between CAD
(Computer-aided design) and CAM (computer-aided manufacturing) as it furnishes
the plan of the process to be used in producing a designed part. Normally, CAPP
incorporates various activities depending on the sophistication of the tool, such as
feature extraction from the part model, tool, fixtures and machine selection, process
and sequence determination, set up planning, and machining parameters. This work
aims to employ a feature-based modeling approach (FBM) to extract features in
XML format and sequence the selected process parameters using object-oriented
programming and a knowledge-based system (KBS) approach for prismatic and
axis-symmetry parts [1, 2].

G. Gogulraj (B) · S. P. Leokumar

Department of Production Engineering, PSG College of Technology, Coimbatore, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 973
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_97
974 G. Gogulraj and S. P. Leokumar

2 Review of Literature

Process planning translates the design information of a product into process instruc-
tions for effective manufacturing. With the development of CAD and CAM systems,
process planning evolved into CAPP systems that can act as a bridge. While the latter
creates a process plan of a new part based on the analysis of the geometry, material,
and other miscellaneous factors that affect manufacturing [3, 4]. Among the several
activities of CAPP, feature recognition is considered to be the dreariest job as it fails to
comprehend the geometrical specification of a part drawing given by the CAD system
in terms of its engineering meaning. It started in the late 1990s, an object-oriented
approach to interface CAD and CAPP systems developed. Further on, FBM and
based models have been developed and much research has been contributed towards
their development. FBM is a model that allows the designer to specify features in
engineering terms rather than geometric terms and besides can store non-graphic
information [5]. Sivankar et al. [6] developed an intelligent CAPP system based on
FBM and neural networks (NN) which recommends appropriate manufacturing oper-
ations and their sequence. The most commonly used FBM approaches are designed
by feature (DBF) and standard for the exchange of product model data (STEP). STEP
based feature modeling has been presented for prismatic parts [7]. A high-level 3D
solid model was used as a base for the part of the design and the designed part was
extracted in XML format. It is regarded that XML-based feature extraction ensures
trouble-free integration with the decision-making system [8–10].

3 Methodology

It comprises two components involving DBF-XML based feature extraction with

interactive feature recognition and KBS based process sequence generation for the
operations based on the feature information. Subsequently, the details generated in
XML format by the preprocessor using the part features as the base. This automati-
cally extracts manufacturing features and model process parameters from the XML
file in a Visual Studio development environment (decision making) via a user inter-
face. Further on, the module executes the process sequencing activity in the Visual
Studio environment based on the manufacturing features and database developed for
various operations performed on prismatic and axis-symmetric parts.
XML Based Feature Extraction and Process Sequence Selection … 975

4 Feature-Based Modeling and Automatic Feature

Extraction and Verification of Micro Machine Tool
Capability

To successfully integrate CAD with CAPP, a component model that contains infor-
mation more than just topology and geometry is essential. High-level information
such as a selection of materials, tolerance, surface finish, etc. is limited in the
existing CAD system. Both high and low-level process parameters are extracted
by the pre-processor inbuilt in modeling software and are converted to XML data.
By converting process parameters into XML complex feature recognition processes
can be neglected, which is one of the advantages of this work. The feature extraction
is automated by the object-oriented approach developed in our previous work [11].
XML files are converted to an appropriate input format to be fed to the process
planning task by the system-designed post-processor. Feature and model information
is automatically extracted from the XML file using an algorithm developed with an
object-oriented approach and the extracted information is sent to the post-processor.
The function of the post-processor is to convert the extracted features and model
parameters to a format suitable for verification. This is done by randomly searching
for the features and model parameters in the XML file using a developed algorithm.
The signal to the post-processor to extract parameters from the XML file is given
by selecting the extract option shown in the user interface. Complex features are
eliminated by an intelligent and interactive system developed as shown in Fig. 4. A
part CAD model is an elementary input for a CAM system. For the layout, details
have been introduced, the definition of the XML schema and an algorithm has been
developed to extract function and model parameters as shown in Figs. 1, 2 and 3.
In addition to extracting features, the proposed system also performs verification
of micro machine tool capability feature detail in an extract XML file in Fig. 2.
It offers expert advice when feedback is beyond the capabilities of the machine
tool. It contains the proposed methodology and the pseudo-code for the extraction
and verification of the part feature. When confirmed, Process plan generation and
process parameter generation feature and model information further used for decision
making.

5 Development of Process Sequence Module

The primary task in process planning is to select the operations to be performed

depending on the information received from the feature extraction process. For pris-
matic parts, the operations are 1. Micro-End milling, 2. Micro-End milling, and
Micro-drilling, and 3. Micro-Ball End milling, End milling, and Micro-drilling.
In the case of axis-symmetric parts, there are seven operations namely, 1. Step
turning, 2. Step turning and Micro-drilling and 3. Step turning, Taper turning and
Micro-drilling, 4. Step turning, Grooving, and Micro-drilling, 5. Step turning, Taper
976 G. Gogulraj and S. P. Leokumar

Fig. 1 Methodology for feature extraction and verification for micro machine tool capacity from
XML files

turning, grooving and Micro-drilling, 6. Step turning, Grooving, Micro-End milling,

and Micro-drilling, and finally 7. Step turning, Taper Turning, Grooving, Micro-
End milling, and Micro-drilling. Knowledge derived from experiments was used to
generate the possible sequence of micro-machining operations to be performed in
prismatic and axis-symmetric parts. Tables 1 and 2 shows a possible combination of
micro-machining processes and their sequence.

6 Implementation

The Integrated system development consists of (1) modeling environment—

Autodesk Inventor (2) XML data file (3) Decision-making module—Visual studio
XML Based Feature Extraction and Process Sequence Selection … 977

Fig. 2 Feature Extraction in XML file

Fig. 3 Process sequence generation

978 G. Gogulraj and S. P. Leokumar

Table 1 Different micro-machining operations and their process sequence for prismatic parts
1. Prismatic parts
Sl.no Sequence Sequence Justification
1 1. End milling-(Roughing) The open-source operation of the end
2. End milling-(Finishing) milling operation is initially performed
2 1. End milling-(Roughing) Open source end milling is initially
2. End milling-(Finishing) performed, followed by closed source
3. Micro-drilling-(Roughing) drilling. Open source slot milling is done
4. Micro-drilling-(Finishing) first to avoid burr deposition in the drilled
hole
3 1.Ball End milling-(Roughing) Open source milling is initially done to
2.Ball End milling-(Finishing) form a chamfer at the edge. To avoid burr
3.End milling-(Roughing) deposition in the machined part, slot
4.End milling-(Finishing) milling is performed first, followed by
5.Micro-drilling-(Roughing) sequential drilling by hole diameter (from
(For different hole sizes in Cad model larger to smaller diameter)
operation sequence = D1 > D2 > D3 > D4)
6.Micro-drilling-(Finishing)

and (4) user interface [Figs. 1 and 2]. Based on the decision made by the decision-
making environment, the process sequences are generated step by step manner. The
decision-making environment receives the input from the user about the micro-
machining process through the user interface. Here, the visual studio environment
plays a vital role in providing the decision-making environment to select the process
sequencing required for the part.
The developed system has a menu bar where feature extraction and plan gener-
ation are available. Once selected, a pop-up window will generate a red color label
and feature name and add-on dimension limits exceed for feature extraction and
on selecting the extraction option the feature extraction takes place automatically.
The design of the system will not allow the user to input values that are beyond
the specification of the micro-machine tool and if the value is beyond the specifi-
cation the system will give expert advice for alternate options as shown in Fig. 3.
The features that are highlighted in red color and feature name to add on dimension
limits exceeded are beyond the micro-machine tool specification, so from this, the
user can identify that the values have to be modified accordingly.
In the process sequence, the user will be allowed to select the order in which
the sequences are to be executed. The process sequence module will be carried out
following the material of the part and surface finish. After the sequence is set, the tool
and fixture module will select the respective tools and fixtures required concerning
the process sequence module and process parameter module by the material and
geometry of the part. After all the parameters are fixed the developed system will
generate the process plan output in a suitable format for low-level manufacturing as
shown in Fig. 4.
XML Based Feature Extraction and Process Sequence Selection … 979

7 Conclusion

A knowledge-based process sequencing module in a CAPP system was developed

with the extraction of features from the models of prismatic and axis-symmetric parts
using the FBM approach. The interactive feature recognition system helps the user
to easily extract the features from the part model which is usually complex in earlier

Table 2 Different micro-machining operations and their process sequence for Axis-Symmetry
parts
2. Axis-Symmetry parts
Sl.no Sequence Sequence Justification
1 1. Facing A facing and plain turning
2. Plain turning-(Roughing) operation are initially performed
3. Plain turning-(Finishing) to reduce the extra material size,
4. Step turning-(Roughing) followed by a gradual step-by-step
5. Step turning-(Finishing) turning operation with different
cutting depths
2 1. Facing Facing and plain turning
2. Plain turning-(Roughing) operations are carried out
3. Plain turning-(Finishing) sequentially followed by step
4. Step turning-(Roughing) turning with varying depth of
5. Step turning-(Finishing) cutting. Micro-drilling is
6. Micro-drilling-(Roughing) performed as a final operation to
7. Micro-drilling-(Finishing) facilitate a change of plane (as the
workpiece must be removed from
the spindle and placed on the vice)
3 1. Facing Facing and plain turning
2. Plain turning-(Roughing) operations are carried out
3. Plain turning-(Finishing) sequentially followed by step
4. Step turning-(Roughing) turning with varying depth of
5. Step turning-(Finishing) cutting and taper turning with the
6. Taper turning-(Roughing) required dimension. Micro-is
7. Taper turning-(Finishing) performed as a final operation to
8. Micro-drilling-(Roughing) facilitate a change of plane (as the
9. Micro-drilling-(Finishing) workpiece must be removed from
the spindle and placed on the vice)
4 1.Facing Facing and plain turning
2.Plain turning-(Roughing) operations are carried out
3.Plain turning-(finishing) sequentially followed by step
4.Step turning-(Roughing) turning with varying depth of
5.Step turning-(Finishing) cutting. Grooving is performed to
6.Grooving-(Multi-pass) reduce the diameter of a particular
7.Micro-drilling-(Roughing) region. Micro-is performed as a
8.Micro-drilling-(Finishing) final operation to facilitate a
change of plane (as the workpiece
must be removed from the spindle
and placed on the vice)
(continued)
980 G. Gogulraj and S. P. Leokumar

Table 2 (continued)
2. Axis-Symmetry parts
Sl.no Sequence Sequence Justification
5 1. Facing Facing and plain turning
2. Plain turning-(Roughing) operations are carried out
3. Plain turning-(Finishing) sequentially followed by step
4. Step turning-(Roughing) turning with varying depth of
5. Step turning-(Finishing) cutting and taper turning with the
6. Taper turning-(Roughing) required dimension. Grooving is
7. Taper turning-(Finishing) performed to reduce the diameter
8. Grooving-(Multi-pass) of a particular region. Micro-is
9. Micro-drilling-(Roughing) performed as a final operation to
10. Micro-drilling-(Finishing) facilitate a change of plane (as the
workpiece must be removed from
the spindle and placed on the vice)
6 1. Facing Facing and plain turning
2. Plain turning-(Roughing) operations are carried out
3. Plain turning-(Finishing) sequentially followed by step
4. Step turning-(Roughing) turning with varying depth of
5. Step turning-(Finishing) cutting. Grooving is done next to
6. Grooving-(Multi-pass) reduce the diameter of a particular
7. Micro-slot-(Roughing) region, followed by a
8. Micro-slot-(Finishing) micro-milling operation to avoid
9. Micro-drilling-(Roughing) burr deposition in micro holes.
10. Micro-drilling-(Finishing) Micro-drilling is performed as a
final operation to facilitate a
change of plane (as the workpiece
is to be removed from the spindle
and placed on the vice)
7 1. Facing Facing and plain turning
2. Plain turning-(Roughing) operations are carried out
3. Plain turning-(Finishing) sequentially followed by step
4. Step turning-(Roughing) turning with varying depth of
5. Step turning-(Finishing) cutting and taper turning with the
6. Taper turning-(Roughing) required dimension. Grooving is
7. Taper turning-(Finishing) done next to reduce the diameter
8. Grooving-(Multi-pass) of a particular region, followed by
9. Micro-slot-(Roughing) a micro-milling operation to avoid
10. Micro-slot-(Finishing) burr deposition in micro holes.
11. Micro-drilling-(Roughing) Micro-drilling is performed as a
12. Micro-drilling-(Finishing) final operation to facilitate a
change of plane (as the workpiece
is to be removed from the spindle
and placed on the vice)
XML Based Feature Extraction and Process Sequence Selection … 981

Fig. 4 XML File Extraction Exceeding the Dimension Limits

systems. Through heuristic methods, the database for the sequence of the process to
be followed for each operation was developed for both the part models under study.
The visual Studio environment was used to integrate the two components namely
feature extraction and process sequencing through the user interface.
The system developed is interactive, intelligent, user friendly, and can be extended.
Nevertheless, the present system is proposed for prismatic parts and axis-symmetric
parts that possess features like micro mills, turns, and grooves. With the addition
of various micro features and incorporation of other CAPP activities like parameter
selection, tool and fixture selection, setup selection, and automatic generation of NC
part program for the available model information, a complete CAPP system can be
realized in the future for prismatic and axis-symmetric parts.

References

1. Rahman M, Senthil Kumar A, Prakash JRS (2001) Micro milling of pure copper. J Mater
Process Technol 116:39–43
2. Shunmugam MS, Mahesh P, Bhaskara Reddy SV (2002) A method of preliminary planning for
rotational components with c-axis features using genetic algorithm. Comput Indus 48(3):199–
217
3. Amaitik SM, Kiliç SE (2005) STEP-based feature modeller for computer-aided process
planning. Int J Prod Res 43(15):3087–3101
982 G. Gogulraj and S. P. Leokumar

4. Asad ABMA, Masaki T, Rahman M, Lim HS, Wong YS (2007) Tool-based micro-machining.
J Mater Process Technol 192–193:204–211
5. Lai X, Li H, Li C, Lin Z, Ni J (2008) Modelling and analysis of micro-scale milling considering
size effect, micro cutter edge radius, and minimum chip thickness. Int J Mach Tools Manuf
48:1–14
6. Sivasankar R, Asokan P, Prabhakaran G, Phani AV (2008) A CAPP framework with optimized
process parameters for rotational components. Int J Prod Res 46(20):5561–5587
7. Amaitik SM (2013) Fuzzy logic models for selection of machining parameters in capp systems.
Int J Comput Inf Technol 2(2):279–285
8. Leo Kumar SP, Jerald J, Kumanan S (2013) “An expert system for cutting tool selection
for micromachining processes”. International conference on precision, meso, micro and nano
engineering, Calicut, pp 133–139
9. Leo Kumar SP, Jerald J, Kumanan S (2014) An intelligent process planning system for micro
turn-mill parts. Int J Prod Res 52:6052–6075
10. Leo Kumar SP, Jerald J, Kumanan S (2015) Feature-based modelling and process parameters
selection in a capp system for prismatic micro parts. Int J Comput Integr Manuf 28:1046–1062
11. Leo Kumar SP, Jerald J, Kumanan S, Prabakaran R (2017) A review on current research aspects
in tool-based micromachining processes. Mater Manuf Process 29:1291–1337
To Study and Optimize the Effects
of Process Parameters on Taper Angle
of Stainless Steel by Using Abrasive
Water Jet Machining

Meghna K. Gawade and Vijaykumar S. Jatti

Abstract Abrasive water jet machining (AWJM) is a non-conventional manufac-

turing process, has a potential to cut wide range of materials. For processing various
engineering materials abrasive water jet cutting has been proven to be an effective
technology. The motive of the paper is to analyze the process parameters on taper
angle in abrasive water jet machining having grade type of 304 stainless steel material.
Design of experiment were conducted according to response surface methodology
(RSM), based on Box-Behnken design. Influence of process parameters on taper
angle is shown by main effect plots and 3D surface plots. Evaluation of process
parameters were done by ANOVA technique. For optimization of process parame-
ters so as to achieve minimum taper angle, multi-objective response methodology
is used which resulted desirability 0.9195 of the developed model. The optimal
process parameters obtained were traverse rate 80 mm/min, abrasive flow rate 300
gm/min, and stand-off distance 1 mm. For validation of results, confirmation analysis
is performed and resulted percentage error showed is less than 6% for taper angle.

Keywords Abrasive water jet machining · Response surface methodology · Taper

angle

Nomenclature

AWJM Abrasive water jet machining

RSM Response surface methodology
TS Traverse speed
ANOVA Analysis of variance
AFR Abrasive flow rate
DOE Design of experiment
SOD Stand-off distance

M. K. Gawade (B) · V. S. Jatti

Department of Mechanical Engineering, D.Y. P. COE, Pune, Maharashtra, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 983
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_98
984 M. K. Gawade and V. S. Jatti

1 Introduction

Abrasive water jet machining (AWJM) is a non conventional machining process

and is proven effective technology to cut various materials. Stainless Steel is one
of the popular materials in automobile sector due to their heat resistance, corro-
sion, strength, durability, high hardness, fabrication flexibility, and low maintenance
characteristics. The purpose of stainless steels used in the areas of automobile and
various manufacturing field have been demonstrated many studies. During machining
of stainless steels with conventional method face many problems such as poor
chip breaking, work hardening, huge amount of coolant supply and these results
in increased in production cost and time. They mainly get prone to edge chipping
due to vibrations occurred by machines, chattering of tool, and o traditional tools.
To overcome this problem, operations carried out by abrasive water jet machining
(AWJM) is an effectively proven method and have several advantages such as any
manufacturing of intricate shapes with precision, also fine finishing surface of the
material can be obtained [1]. Hardness of abrasives plays important in cutting of the
material. Garnet abrasives produces small width of cut compared with other abrasives
material such as silicon carbide, aluminium oxide [2]. A open tapered slot was inves-
tigated, it was observed that width of top is wider as compared to width of bottom
and normally known as kerf taper angle and is represented as ‘θ ‘as a characteristic
[3].

2 Materials and Methods

The material used is iron based alloy, which is manufactured commercially having
grade type 304 stainless steel sheet with dimension 330 mm × 100 mm × 6 mm. Its
yield strength is 205 MPa, tensile strength is 515 MPa, hardness 92 HRB. Chemical
composition of type 304 stainless steel specification is shown in Table 1.
Experiment was performed on Water Jet German—3015 Machining facilitated
with CNC equipment. Table 2 shows process machining parameters. The experiments
were carried under water pressure 3100 bar and the abrasive material was garnet 80
mesh size. Figure 1 shows the abrasive water jet machine.
The DOE is a systematized technique to establish the correlation between control
factors and final response. Table 3 shows the 3 control factors levels.
RSM (response surface methodology) is a statistical method used to find the
relationship of control factors on response process. RSM is used to design with

Table 1 Chemical composition of 304 steel

C% S% P% Ni% N% M% Si% C%
0.08 0.03 0.05 8 0.1 2 0.75 18
To Study and Optimize the Effects of Process Parameters … 985

Table 2 Parameters of
Sr No. Parameters Values
AWJM machine
1 Diameter of nozzle 1.1 mm
2 Orifice diameter 0.35 mm
3 Water pressure 2500–3100 bar
4 Abrasive flow rate 100–300 g/min
5 Traverse speed 50–200 mm/min

Fig. 1 Abrasive water jet machine

Table 3 Control factors and their levels

Symbol Parameters Level 1 Level 2 Level 3
A Traverse speed (mm/min) 80 120 160
B Abrasive flow rate (g/min) 100 200 300
C Stand-off distance (mm) 1 2 3

minimum number of experiments performed and also used to obtain optimum condi-
tions to produce desirable responses. Box-Behnken design (BBD) model were used
for 3 level and 3 factors based on the experiments were performed. The relationship
between the independent variables and response variables [4] is given in Eq. 1.


k 
k 
k
Y = β0 + βi X i + βii X ii2 + βi j X i X j (1)
i=1 i=1 i, j=1,i= j

where Y is represented as variable response; Xi, X 2 , and Xi Xj are input variables;

β0 are coefficient of intercept model; β0 , βi, βij and βii are regression coefficients
of linear, quadratic, and second-order terms, respectively. Design of experiment is
986 M. K. Gawade and V. S. Jatti

generated by using RSM approach in MINITAB 19 software. The response obtained

during the experimental trials is listed in Table 4.
Figure 2 shows square slots was cut having dimension as 20 × 20 mm using AWJM
After the experiment the width at top of material and width at the bottom of each
slot is calculated by digital vernier caliper. Taper angle is calculated by formula [5]

Wt − Wb
Taper angle(θ ) = tan−1 ( )degree (2)
2∗t

where, Wt is width cut at top in mm, Wb is width cut at bottom in mm, t- thickness
in mm.

Table 4 Responses obtained for Trials

Trials Parameters Taper angle
A B C
1 80 200 1 1.59
2 80 300 2 1.49
3 80 200 3 1.6
4 80 100 2 1.55
5 120 200 2 1.63
6 120 300 1 1.41
7 120 200 2 1.62
8 120 100 1 1.66
9 120 100 3 1.7
10 120 200 2 1.62
11 120 300 3 1.51
12 160 200 3 1.6
13 160 300 2 1.5
14 160 100 2 1.69
15 160 200 1 1.66

Fig. 2 Stainless steel

machined by AWJM
To Study and Optimize the Effects of Process Parameters … 987

3 Results and Discussion

3.1 Statistical Analysis of Stainless Steel for Taper Angle

The mathematical relationship between independent parameters and taper angle

regression equation is obtained and is presented in Eq. (3).

Taper angle(θ) = 1.6562 + 0.000687 A − 0.000863 B + 0.0113 C (3)

From Eq. (3), it has been observed that parameter SOD and TS have positive
effects and AFR has the negative effect on taper angle.
Further, to check the most significant parameter affecting on taper angle is deter-
mined by ANOVA (Analysis of Variance) about 95% confidence interval (CI). It is
a computational method helps us to evaluate the significance of every control factor
on response factor.
Table 5 shows the ANOVA results. The F value is the fisher’s statistical test and
P value is the probability of significance acts in accordance with coefficient of R-sq
and R-sq (adj) is determined. This coefficient implies the acceptability and suitability
of the model. It is observed that, process variables having lager F value and P value
less than 0.05 implies that the variable is significant. For the value of AFR, P value
less than 0.05 implies that the variable is significant statistically. The value of R-sq
is 89.47% and R-sq (adj) is 70.53% for taper angle implies accuracy and fitness of
model. Higher the value of R-sq satisfies the accuracy and fitness of model.
For better analysis, we have plotted the graphs of probability and residuals for
taper angle of stainless steel material. The graph of normal probability shows that
residual are distributed normally have a close fit to line. Versus fit graph implies
there is randomly distribution of residuals. Versus order graph implies observation
a constant variance. This graphs shows clarity that the observation are reliable and
has 95% confidence interval (CI). Figure 3 shows the residual graph.
Figure 4 shows the effects of various process parameters on taper angle. It is
observed that there is rise in taper angle as traverse rate increases this is because as
traverse speed increases cutting action gets reduced at bottom width of cut this leads

Table 5 ANOVA for taper angle

Source DF Seq SS Adj SS Adj MS F-Value P-Value
A 2 0.00628 0.006553 0.003276 1.61 0.258
B 2 0.070436 0.070346 0.035173 17.32 0.001
C 2 0.001015 0.001015 0.000508 0.25 0.785
Error 8 0.016242 0.016242 0.00203
Lack-of-fit 6 0.016175 0.016175 0.002696 80.88 0.012
Pure error 2 0.000067 0.000067 0.000033
Total 14 0.093973
988 M. K. Gawade and V. S. Jatti

Fig. 3 Residual graph plots

(a)

(b)

(c)

(d)
To Study and Optimize the Effects of Process Parameters … 989

Fig. 4 Main effects plot for taper angle

to increases in the value of taper angle. It is observed that as flow rate of abrasives
increases there is decrement in taper angle. Also, as stand-off distance increases there
is increment in taper angle is due to increase in focus area of jet results increase in
width of cut.
Figure 5 shows the surface plots, to find the effect of different control factors,
surface plots are also generated. Figure 5a shows interaction between AFR and SOD
keeping value constant of TS at 120 mm/min. It is noticed that for lower value of SOD
and higher value of AFR there is decrement in angle of taper and similarly for higher
SOD and lower AFR there is rise in angle of taper, this is due to abrasive particles
which cannot pierce sufficiently inside the specimen material [6]. Figure 5b shows
the interaction between SOD and TS keeping value constant AFR at 200 g/min. As
there is increase in TS and SOD as there is rise in angle of taper. Similarly, at lower
SOD and TS there is fall of angle of taper. Figure 5c shows that at higher TS and
lower of AFR there is rise in the angle of taper keeping the value constant of SOD
at 2 mm. Similarly, at the higher value of AFR and lower TS there is decrement in
angle of angle of taper.

4 RSM Optimization

Optimization of control parameters, multi-objective optimization is executed by the

response surface optimization (RSO) technique with respect to response parameters.
Table 6 shows constraints of parameters for optimization. A desirability (D) function
is used in response surface optimization, to optimize the response factors. Desirability
value changes in range of 0 to 1. If the value of d = 0, response value are unacceptable
where as d = 1 shows desirability of the model. RSO technique is based on principle
990 M. K. Gawade and V. S. Jatti

(a) (b)

(c)

Fig. 5 Surface Plot for taper angle

Table 6 Constraints of
Condition Goal Lower limit Upper limit
parameters for optimization
Traverse speed, A In range 80 160
Abrasive flow rate, B In range 100 300
Stand-off distance, C In range 1 3
Taper angle (degree) Minimum 1.41 1.70

of calculating weights and individual desirability for every response. Higher the value
of‘d’ implies higher desirables optimum values obtained for response parameter [6].
RSM optimization was performed for input parameters, multi-objective optimiza-
tion were used for response parameter. Figure 6 shows plot for response optimization.
It implies the composite desirability of 0.9195 for optimizing the taper angle. The
desirability value 0.9195 indicates acceptance of the response parameter. Optimal
values obtained TS 80 mm/min, AFR 300 gm/min, and SOD of 1 mm, so as to
achieve minimum value of angle of taper.
To Study and Optimize the Effects of Process Parameters … 991

Fig. 6 Plot of optimization of process parameters

5 Confirmation Analysis

Confirmation analysis is the analysis is used to analyze the accuracy of the model
performed by RSM. Table 7 is referred to plot the value of experiment and predicted
results. Percentage error is calculated between experimental and predicted results. It
is observed that percentage error is less than 6% within the limit of permissible. The
predicted values of RSM model show a high level of accuracy.
Figure 7 shows the difference between experimental results and modeling results.
It can be concluded to get low value of the angle of taper for the input parameters
such as TR, AFR and SOD the obtained RSM model is fulfilled.

6 Conclusions

The analysis reveals the impact of input parameters on angle of taper of stainless
steel material and thus we conclude as follows—
(i) The mathematical model is formulated for the taper angle showed the corre-
lation between the flow rate of abrasive, traverse speed and SOD. The most
significant parameter is AFR influencing on angle of taper significant statis-
tically followed by TS and SOD. The value of R-sq is 89.47% and R-sq (adj)
992 M. K. Gawade and V. S. Jatti

Table 7 Confirmation analysis

Exp No. A B C Experimental result of taper Modelling result of taper % Error
angle angle
1 80 300 2 1.47 1.47 0
2 160 200 3 1.6 1.62 −1.25
3 80 200 3 1.6 1.57 1.87
4 120 200 2 1.63 1.59 2.45
5 120 300 1 1.41 1.49 −5.67
6 160 300 2 1.5 1.52 −1.33
7 80 200 1 1.59 1.55 2.51
8 120 200 2 1.62 1.59 1.85
9 120 100 1 1.66 1.66 0
10 160 100 2 1.69 1.70 −0.59
11 120 100 3 1.7 1.68 1.17
12 120 200 2 1.62 1.59 1.85
13 120 300 3 1.51 1.51 0
14 80 100 2 1.55 1.64 −5.80
15 160 200 1 1.66 1.60 3.61

Fig. 7 Comparison experimental results and modeling results

is 70.53% for taper angle implies accuracy and fitness of model by ANOVA
technique
(ii) Main plots and 3D surface plots were studied for the influence of control
parameters on angle of taper. AFR is comparably slighter influential on angle
of taper.
(iii) Optimization of process parameter using multi-objective RSM for minimum
taper angle which results the desirability of 0.9195. Optimal values obtained
To Study and Optimize the Effects of Process Parameters … 993

for TS 80 mm/min, AFR 300 gm/min, and SOD of 1 mm, so as to achieve

minimum value of angle of taper.
(iv) Confirmation analysis of the model was performed showed percentage error
less than 6% for taper angle is in limit of permissible. Comparison of the
experimental and predicted values is shown by graph.
Thus it is concluded that mathematical model is developed and optimization for
prediction of response parameter for stainless steel is almost significant.

References

1. Srinivas S (2018) Machinability Studies on Stainless steel by abrasive water jet—Review. Mater
Today: Proc 57:2871–2876
2. Khan AA, Hague MM (2007) Performance of different abrasive material during abrasive water
jet machining of glass. J Mater Process Technol 41:404–407
3. Gupta V, Pandey PM, Garg M, Khannaand R, Batra NK (2014) Minimization of kerf taper angle
and kerf width using Taguchi’s method in abrasive water jet machining of marble. Proc Mater
Sci 6:140–149
4. Jagadish Bhowmik S, Ray A (2016) Prediction and optimization of process parameters of green
composites in AWJM process using response surface methodology. In: Int J Adv Manuf Technol,
Springer, London
5. Hascalik A, Aydas Guru UC (2007) Effect of traverse speed on abrasive water jet machining of
Ti–6Al–4 V alloy. Mater Des 18:1953–1957
6. Dumbhare P, Dubey S, Deshpande Y, Andhare A, Barve P (2018) Modelling and multi-objective
optimization of surface roughness and kerf taper angle in abrasive water jet machining of steel.
J Brazilian Soc Mech Sci Eng 40:48–55
Effect of Process Parameters on Response
Measures of Cartridge Brass Material
in Photo Chemical Machining

Bandu. A. Kamble, Abhay Utpat, Nitin Misal, and B. P. Ronge

Abstract Now a days to produce stress free and burr free micro components, photo-
chemical machining is one of the emerging technologies. The micro level etching of
components of different materials is carried out by using chemical etching process.
The very precision parts such as microchannels, heat sinks, Fine mesh or screens
and printed circuit boards are developed by this process. In this work, the cartridge
brass was selected as a base metal which has good electrical and thermal conduc-
tivity. The process parameters are etching temperature and time and responses are
etching depth and surface roughness. The objective of this study is to achieve etching
depth on the Cartridge brass plate having thickness 0.5 mm. The etching depth and
surface roughness of material increases with increase in time and temperature. The
study of copper and brass were reported by many researchers. Very few literatures
are available on an alloy of copper and brass. The effect of process parameters is
changing by changing composition of material. This article is focused on the study
of etching depth and surface roughness of cartridge brass material at different time
and temperature. The cartridge brass has been widely used in industrial applications.

Keywords Photo chemical machining · Photo tool · Etch depth · Photo resist

1 Introduction

The method also named as chemical etching, acid etching, Photochemical milling.
The photoresist used PCM has two types given below.
• Positive photoresist—The part of UV exposure becomes soluble in developing
solution.
• Negative photoresist—The part of UV exposure becomes insoluble in developing
solution.

Bandu. A. Kamble (B) · A. Utpat · B. P. Ronge

SVERI’s College of Engineering, Pandharpur, India
N. Misal
SVERI’s College of Polytechnic, Pandharpur, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 995
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_99
996 Bandu. A. Kamble et al.

The method starts with generating a photo tool i.e. the design created by using
drawing softwares like AutoCAD, Catia etc. and printing it on trace paper. The nega-
tive photoresist solution is applied on a base metal at uniform thickness of coating.
The photo tool is kept on base metal and the surface is exposed by using ultraviolet
light. The transferred pattern of design will be visible after the developing process. In
the final step of method, the base metal is dipped into the etchant solution for etching
or machining. Davis et al. [1] developed a process to etch micro-fluidic channels
and other micro components. Wangikar et al. [2] optimized process parameters of
brass and German silver and concluded with brass has a better etching performance.
Agrawal et al. [3] creates microfluidic channels, micro holes by using the PCM
process. The manufacturing of such parts by using cutting tools are very challenging
work and many defects will be occurring. Jang et al. [4] and Sadaih et al. [5] devel-
oped a 3D system in the package (SIP) process. It has been considered as a superb
microelectronics packaging system. The selection of etchant for machining of parts
is depending on the surface topology and hardness of material. The surface finishing
of the etched part is varying with etchant to etchant.
The major steps involved on 3D PCM is given below
• Creation of photo tool
• Selection of material for etching
• Modelling of workpiece
• Film coating using photoresist
• Developing
• Etching
• Stripping and inspection.
Misal et al. [6, 7] shows the surface topography of Inconel 718 and responses
are measured at different conditions and checked effect of process parameters on
response measure for hard materials like Monel and Inconel 718. Saraf et al. [8]
presented a mathematical model of PCM and optimized conditions of PCM process.
Cakir et al. [9] etched Cartridge brass by cupric chloride and suggested a suitable
process for regenerated waste treatment. Allen D.M et al. [10] demonstrate the three-
dimensional etching process and gave characterisation of ferric chloride etchant.
the development of three-dimension components is possible by using lamination
techniques. Bruzzone et al. [11] generated simulation model of 2D photochemical
machining process. The simulation was carried out in six steps. An analysis of the
PCM process was carried out at differentconditions. Kamble et al. [12] shows variable
etching process for different colours. The depth of etching is changing by changing
UV light exposing at different intensity of light. The light passing intensity is different
for different colours due to this hardening of photoresist is different for different inten-
sity of light or colour. The more and less cross linking of photoresist results in uneven
etching of base metal. Jadhav et al. [13] have focused the etching depth variation
of brass material for different operating conditions. The parametric study of process
was reported. Mazarbhuiya et al. [14] studied PCM process for Aluminium material.
The effect of different parameters such as etching time, temperature and concentra-
tion were taken into consideration to measure responses like depth, roughness and
Effect of Process Parameters on Response Measures of Cartridge … 997

edge deviation. Manoharan et al. [15] compared the PCM process to binder jetting
to develop micro post arrays. PCM found the cheapest process for developing micro
arrays. Further, the study related to photochemical machining on copper, brass, etc.
for producing microchannels, micro textures in bearing, heat sinks, etc. were reported
by many researchers [16–25]. Very few literatures are available on an alloy of copper
and brass. The study of PCM process for copper and Brass alloys has wide scope. In
this work, the etching behaviour of Cartridge Brass was analysed.

2 Material and Method

The flowchart for PCM is presented in Fig. 1, which explains the different steps
in PCM. The cartridge brass was selected for study which has good electrical and
thermal conductivity and it is widely used in industrial applications. The selected
material for etching has been cut into a size of 40 mm × 40 mm × 0.6 mm (L × B×
T). The design was created by using drawing softwares like AutoCAD, Catia etc.
and printed it on trace paper. The 2400 × 2400 dpi printer was used to print photo
tool as shown in Fig. 2a. The accuracy of PCM is mainly depending on the quality
of printing images. The thickness of the photo tool is 80 microns.
The selected metal is cleaned by using paper or cleaning solution i.e. thinner
to remove all dust particles on it. After cleaning of base metal, in the next step
negative type photoresist solution applied on metal by using spin coating method.
The uniform coating thickness was maintained throughout the coating process. After

Fig. 1 PCM flowchart [2] Selection of material

Photo tool design

Cleaning

Photoresist coating

UV light exposing

Developing

Etching

Stripping
998 Bandu. A. Kamble et al.

a) b) c)

d)

Fig. 2 a Photo tool b Exposure unit c Etching unit d Finished component

that, the coating is baked for 2 min for better adhesion between photoresist and brass
material. The high intensity UV light is used to expose coating on material at 90 s. In
the exposing process, the UV light is passed through white color and restricts through
black color, the exposed area is hardened (white) and unexposed area becomes soft
(Black). After exposing, the sample is developed in a developing solution (Mixture of
sodium carbonate and water) at time 90 s. The hardened area of photoresist remains
on base metal and unexposed area is washed out during the development process.
The development process required to transfer required design onto a coating surface
which is open to the atmosphere. The parts open to the atmosphere are etched in
a chemical etching solution (mixture of ferric chloride and water. After the etching
process, the remaining photoresist is washed out by using a striping solution (sodium
hydroxide).
The etching process setup is shown in Fig. 2c.

3 Design of Experiments

The numbers of preliminary experiments were carried out to study the etching mech-
anism of Cartridge brass material. The process parameters are etching time, temper-
ature etc. and response parameters are etching depth and surface roughness. The
etching depth was measured by using 0.001 L.C Digital Micrometre. The concen-
tration of etchant is constant throughout the experiments which is 40 Baume scale
measured by hydrometer. The level of experiments is selected by studying the etching
Effect of Process Parameters on Response Measures of Cartridge … 999

Table 1 Control parameters with their levels

Control parameters Level 1 Level 2 Level 3
Temperature (°C) 43 46 49
Time (Min) 3, 6, 9 3, 6, 9 3, 6, 9

Table 2 Etching parameters and etching depth relation

Etching temperature(°C) Time (Min) Etching depth (µm) Surface roughness (Ra)
43 °C 3 20 0.312
6 45 0.395
9 92 0.480
46 °C 3 35 0.320
6 79 0.412
9 135 0.519
49 °C 3 123 0.325
6 256 0.435
9 371 0.555

behaviour of Cartridge brass material at different temperatures. The etching depth

and surface roughness increases by increasing temperature and time.
The levels of experiments are shown in Table 1.
The coating thickness of photoresist is 4–5 µm, exposure time of UV machine is
85 s and developing time is 90 s. These parameters were taken constant throughout
the experimentation.
The experimental values are shown in Table 2.

4 Results and Discussion

The performance of photo chemical machining of Cartridge brass has been analyzed
at different time and temperature. The etching depth of each specimen was recorded
at three locations, and mean value is recorded. The etching depth of etching material
increases by changing process parameters such as etching temperature and etching
time. The depth of etching increases by increase in etching time and temperature. At
the low temperature, the depth of etch is low due to i.e. less molecules are available
to etching reaction i.e. less collision of molecules on the Cartridge brass surface. As
temperature increases the collisions in etchant molecules is increasing, it results in
higher depth of etch.
The depth of etching material is changed by changing process parameters such
as etching temperature and etching time. Figures 3 and 4 show the relation between
etching depth and process parameters. The depth of etching increases by increase in
1000 Bandu. A. Kamble et al.

Etching depth vs Temp vs Time

400
350
Etching depth (μm)

300
250
200
150
100
50
0
3 6 9
Etching time (Min)

43 46 49

Fig. 3 Etching depth versus temperature versus time

Surface Roughness vs Temp vs Time

0.6
Surface Roughness (μm)

0.5

0.4

0.3

0.2

0.1

0
3 6 9
Time (Min)
43 46 49

Fig. 4 Surface roughness versus temperature versus time

etching time and temperature. The depth of etching and roughness at low temperature
i.e. for 43 °C is low at time 2 min and it increases with time its highest at time 6 min
which is 80 microns. Similarly, for other temperatures, the depth of etching is highest
at higher temperature and time.
Effect of Process Parameters on Response Measures of Cartridge … 1001

4.1 Comparison Between Etching Temperatures and Etching

Depth

The depth of etching is maximum at 49 °C i.e. 371 µm due to more collision of

etchant molecules on the surface of Cartridge brass. The etchant species develop
more reactive molecules at this temperature compared to others. Low temperature
means less reactive species collide on the Cartridge brass surface which results in
less etching depth. The energy of reactive species increases as temperature increases.
The collision of species on surfaces increases by increasing in time, at less time less
species reacts and more time more species react to the Surface.
The etching time also has a significant effect on depth of etch, at minimum time,
the interaction of molecules on the surface is less which results in minimum depth of
etch. Similarly, at maximum time, the interaction of etchant species on the surface
is high which results in higher depth of etch. Figure 3 compared etching depth for
different temperature and time. At the highest temperature the reaction between
etchant species on the surface happens rapidly compared to other temperatures. The
etch rate of any material is depending on the properties of material such hardness,
conductivity and composition of material. The etchant solution is different for soft
and hard material.

4.2 Comparison Between Etching Temperatures and Surface

Roughness

The surface roughness of Cartridge brass material is higher at highest etching temper-
ature and time. The value of roughness changes by change in process parameters, it
increases by increase in temperature and time. The below fig shows the roughness
of cartridge material over a time and temperature. The surface roughness increases
from 0.312 to 0.555 microns. The increase in roughness caused by uneven etching of
surface as time and temperature increases. The energy of etchant molecules depends
on concentration and temperature of etching solution.
The reaction between etchant species on the surface increases with time which
gives maximum etch rate and depth of etch. As time increases, the uneven collisions
of etchant species increases which results in higher roughness value. The reaction of
or etch rate of material is depending on the different properties of material such as
hardness, conductivity and composition of material. The less etch rate obtained for
harder materials. Depending on the characteristic of material, the chemical solution
will be prepared. Figure 4 shows the relationship between surface roughness, temper-
ature and etching time. The surface roughness is higher i.e. 555 µm at temperature
49 °C and time 9 min.
1002 Bandu. A. Kamble et al.

5 Conclusions

The present study evaluate etching mechanism of cartridge brass material at different
temperatures and time. In this study, the effect of process parameters like etching
temperature and time on an etching depth has been studied. The temperature has a
significant effect on etching depth and surface roughness. The higher temperature
results in more etching depth and surface roughness, due to the rate collision of
etchant species on the surface of Cartridge brass is more at higher temperature.
Similarly at low temperature, the etchant molecules are less reactive and less in
number which gives less depth of etch. Depth of etching and roughness goes on
increasing by increase in temperature and etching time. The more etching depth and
surface roughness obtained at 49 °C temperature i.e. 371 microns and 0.555 microns
in 9 min. The lowest depth and roughness obtained at 40 °C temperature i.e. 20
microns and 0.312 micronsin 3 min.

Acknowledgements Authors are extending thanks to the Principal Collaborator Dr. Balasubrama-
niam (BARC) for providing facilities for experimentation funded by BRNS (DAE).

References

1. Davis PJ, Overturf GE III (1986) Chemical machining as a precision material removal process.
Precis Eng 8(2):67–71
2. Wangikar SS, Patowari PK, Rahul DM (2017) Effect of process parameters and optimization for
photochemical machining of brass and german silver. Mater Manuf Process 32(15):1747–1755
3. Agrawal D, Kamble D (2019) Optimization of photochemical machining process parameters
for manufacturing microfluidic channel. Mater Manuf Process 34(1):1–7
4. Jang DM, Ryu C, Lee KY, Cho BH, Kim J, Oh TS, Yu J et al (May 2007) Development
and evaluation of 3-D SiP with vertically interconnected through silicon vias (TSV). In 2007
proceedings 57th electronic components and technology conference. IEEE, pp 847–852
5. Sadaiah M, Patil DH (8–12 Jun 2015) Some investigations on surface texturing on monel 400
using photochemical machining. In proceedings of the ASME 2015 international manufacturing
science and engineering conference MSEC2015, Charlotte, North Carolina, USA. https://doi.
org/10.1115/msec2015-9294
6. Misal ND, Sadaiah M (2017) “Investigation on surface roughness of inconel 718 in photo
chemical machining”. Adv Mater Sci Eng
7. Misal ND, Saraf AR, Sadaiah M (2017) Experimental investigation of surface topography in
photo chemical machining of Inconel 718. Mater Manuf Process 32(15):1756–1763
8. Saraf AR, Misal ND, Sadaiah M (2012) Mathematical modelling and optimization of photo
chemical machining. Adv Mater Res 548:617–622
9. Cakir O (2006) Copper etching with cupric chloride and regeneration of waste etchant. J Mater
Process Technol 175:63–68
10. Allen DM, Almond HJ (2004) Characterisation of aqueous ferric chloride etchants used in
industrial photochemical machining. J Mater Process Technol 149(1–3):238–245
11. Bruzzone AAG, Reverberi AP (2010) An experimental evaluation of an etching simulation
model for photochemical machining. CIRP Ann 59(1):255–258
12. Kamble B, Utpat A, Misal ND, Ronge BP (2019) 3D Photochemical machining of copper
by using colored phototools. Int J New Technol Res (IJNTR) ISSN: 2454–4116, July 2019,
5(7):28–32
Effect of Process Parameters on Response Measures of Cartridge … 1003

13. Jadhav Saurabh M, Karatkar Onkar V, Bangale Kamesh N, Choudhari Deepak B, Utpat AA,
Kamble Banduraj K (2019) Etching depth variation of brass material for different operating
conditions. Int J New Technol Res (IJNTR) ISSN: 2454–4116, April 2019, 5(4):93–96
14. Mazarbhuiya RM, Rahang M (2020) Parametric study of photochemical machining of
aluminium using taguchi approach. Advances in mechanical engineering. Springer, Singapore,
pp 497–504
15. Manoharan S, Summerville S, Freiberg L, Coblyn M, Touma JG, Jovanovic G, Paul BK
(2020) Manufacturing process design of a micro-scale liquid-liquid extractor and multi-phase
separator. J Manuf Process
16. Das SS, Tilekar SD, Wangikar SS, Patowari PK (2017) Numerical and experimental study of
passive fluids mixing in micro-channels of different configurations. Micros Techno 23:5977–
5988
17. Wangikar SS, Patowari PK, Misra RD (2018) Numerical and experimental investigations on
the performance of a serpentine microchannel with semicircular obstacles. Micros Techno
24:3307–3320
18. Gidde RR, Pawar PM, Ronge BP, Shinde AB, Misal ND, Wangikar SS (2019) Flow field analysis
of a passive wavy micromixer with CSAR and ESAR elements. Micros Techno 25:1017–1030
19. Wangikar SS, Patowari PK, Misra RD (2016) Parametric optimization for photochemical
machining of copper using grey relational method. In 1st techno-societal, international
conference on advanced technologies for societal applications. Springer, pp. 933–943
20. Wangikar SS, Patowari PK, Misra RD (2018) Parametric optimization for photochemical
machining of copper using overall evaluation criteria. Mater Tod Proceed 5:4736–4742
21. Wangikar SS, Patowari PK, Misra RD, Misal ND (2019) Photochemical machining: a less
explored non-conventional machining process. In: Non-conventional machining in modern
manufacturing systems, IGI Global, pp 188–201
22. Kulkarni HD, Rasal AB, Bidkar OH, Mali VH, Atkale SA, Wangikar SS, Shinde AB (2019)
Fabrication of micro-textures on conical shape hydrodynamic journal bearing. Int J Tre Engg
Techno 36:37–41
23. Raut MA, Kale SS, Pangavkar PV, Shinde SJ, Wangikar SS, Jadhav SV, Kashid DT (2019)
Fabrication of micro channel heat sink by using photo chemical machining. Int J New Techno
Res 5:72–75
24. Patil PK, Kulkarni AM, Bansode AA, Patil MK, Mulani AA, Wangikar SS (2020) Fabrication
of logos on copper material employing photochemical machining. Nov MIR Res J 5(6):70–73
25. Wangikar SS, Patowari PK, Misra RD, Gidde RR, Bhosale SB, Parkhe AK (2020) Optimization
of photochemical machining process for fabrication of microchannels with obstacles. Mater
Manuf Proces. https://doi.org/10.1080/10426914.2020.1843674
Thermal Performance of Two Phase
Closed Thermosyphon with Acetone
as Working Fluid

Shrikant V. Pawar and Abhimanyu K. Chandgude

Abstract The circular two-phase closed thermosyphon with acetone as working

fluid is analyzed in this study. TPCT pipe is made up of Aluminium. The filling ratio
is 30%. The objective of this study is analyzing aluminium and acetone combination
at increasing heat input. The vertical position TPCT is analyzed for six different heat
input. Thermal efficiency and mean temperature difference between evaporator and
condenser is determined and plot. Thermal performance of TPCT is increasing with
an increase in input heat supply.

Keywords Acetone · Aluminium · Condenser · Evaporator

1 Introduction

Two-phase closed thermosyphon is most reliable heat transmission device. It trans-

mits heat from the high-temperature region to the low-temperature region by the
mechanism of phase change. The Working fluid absorbs the heat and gets converted
into the vapour state. This vapour goes to condenser section where it converts into
liquid by rejecting heat to the surrounding environment. It is a passive device of
heat transmission. Technology is developing day by day. Modern world technology
is more compact, fast, and more efficient. These high performing devices generate
more heat. So it’s needed for modern technology to develop a device which absorbs
more heat, compact in structure and more reliable. Therefore thermosyphon attract
the attention of the researcher.
In the past few years, the researcher tries to enhance the heat transmission perfor-
mance of TPCT. Shabgard and Xiao [1] analyzes closed thermosyphon under various
working fluid and filling ratio. The filling ratio can be defined as a ratio of the volume
of working fluid in the TPCT pipe to the volume of TPCT pipe. In this analysis,
finite volume model is to develop and validate by using experimental data. The
model can predict the optimum filling ratio. Nano fluid-based working fluid is used

S. V. Pawar (B) · A. K. Chandgude

School of Mechanical and Civil Engineering, MIT Academy of Engineering, Alandi, Pune, India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 1005
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_100
1006 S. V. Pawar and A. K. Chandgude

to improve Thermal properties. Kamyar et al. [2] use Al2 O3 and TiSiO4 Nanoparti-
cles suspended in water. He concluded that thermal resistance of TPCT decreases,
evaporator wall temperature decreases and an increase in heat transfer coefficient.
The surface of evaporator and condenser modified to improve Thermal performance
of TPCT. Rahimi et al. [3] analyzes the effect of surface modification on the evapo-
rator and condenser surface. It decreases the thermal resistance of TPCT. Tsai et al.
[4] analyze the TPCT for electronic cooling. He analyzes the effect of parameters
such as fill ratio, evaporator surface structure, heating power on the performance of
TPCT. Brusly Solomon et al. [5] analyzed anodised and non anodised TPCT with
acetone as working fluid. Simple porous surface is prepared by anodising technique.
Anodised surface has a significant effect on evaporator’s thermal resistance and heat
transfer coefficient and the same is negligible on the condenser. Ong and Haider-
E-Alahi [6] filled Refrigerant R-134a in TPCT and analyze for fill ratio, coolant
mass flow rates, the mean temperature difference between evaporator and condenser.
Nano-fluid cause fouling on the internal surface of TPCT. Sarafraz et al. [7] analyzes
the effect of fouling caused by Nano-fluid on an internal surface of TPCT. Fouling
increases with an increase in the mass of Nano-fluid. Fouling causes deterioration in
thermal performance with time. Noie [8] analyzes the TPCT made up of copper tube
filled with distilled water. In his study, three parameters are considered: heat transfer
rates, working fluid filling ratio, evaporator length. Problem with Nanofluid is that it
causes fouling on the internal surface of TPCT. Sarafraz et al. [7] analyze Effect of
fouling [2]. Never mentioned anything about fouling and it’s an effect on heat trans-
mission. Dry out at higher is one of the limitations of TPCT at higher heat input. [4,
6, 8] analyze the effect of filling ratio on the performance of TPCT. Ong and Tong
[9] analyzes the effect filling ratio and inclination. Water is used as a working fluid.
Effect of filling ratio is determined. Jiao and Qui [10] analyzes vertical TPCT filled
with nitrogen. Parameters like filling ratio, heat input, operating pressure, geome-
tries are analyzed. Ordaz-Flores [11] analyze acetone and methanol filled two-phase
flat plate solar system. Experimental result shows that partial vacuumed condition at
starting to improve Thermal performance of phase change system.
From the literature survey, it is clear that different method implemented to improve
the thermal performance of TPCT. The researcher tries a different combination of
pipe material and working fluid. In this study, aluminium and acetone combination
is used and it’s thermal performance evaluated for 30% filling ratio.

2 Experimental Setup

The experimental set-up is presented in Fig. 1. The cross-section of TPCT pipe is

circular. It fills with acetone as a working fluid. The length of TPCT is 350 mm and
the inner diameter is 17.5 mm and wall thickness is 1 mm. TPCT is made up of
aluminium. Acetone is fill up to 30% of the total volume of TPCT (fill ratio). TPCT
pipe is a divide in three sections adiabatic, evaporator, and condenser having the
length of 100 mm, 100 mm, and 150 mm respectively. Main components of the testing
Thermal Performance of Two Phase … 1007

Fig. 1 Diagrammatic representation of experimental setup

unit are Pump, Control panel, Thermocouple, Rotameter, tilting bracket, heating coil.
TPCT pipe placed in two aluminium block at two ends called as evaporator block and
condenser block. Middle section made up of glass wool is called adiabatic section.
The evaporator is 100 mm block with a slot at the centre for TPCT pipe. Four
holes drilled for heating coils around TPCT slot. Three thermocouples are placed at
equidistance to measure the surface temperature of TPCT. The similar block is used
for a condenser with two openings for water inlet and outlet. Water is used to carry the
heat reject by TPCT pipe. Three thermocouples are placed over the condenser section.
The adiabatic section made up of glass wool so that no heat transfer should take place.
Three number of a thermocouple is placed over there. Two number of a thermocouple
is used for measuring the inlet temperature (T11) and outlet temperature (T12) of
water. One thermocouple measures atmosphere temperature (T10). Rotameter is used
to adjust the flow of cooling water (LPH). Control unit is used to vary the heat input.
Acetone is used as a working fluid. Most of working fluids used have lower latent
heat of vaporization and lower boiling point at atmospheric pressure. But in case
of acetone latent heat of vaporization and boiling point, both are high as compare
to refrigerant like R134a. Also, acetone shows good compatibility with aluminium
metal. Due to the lower amount of working fluid in the pipe, dry out situation may
arise. So to deal with this phenomenon 30/c/o of filling ratio is used in TPCT.
Properties of acetone
Boiling point 329.2 k
The heat of evaporation 534 kJ/kg k
Specific heat capacity, Cp 2.14 kJ/kg k
Specific heat capacity, CV 1.55 kJ/kg k
1008 S. V. Pawar and A. K. Chandgude

3 Results

The acetone filled thermosyphon is analyzed in this paper. The vertical thermosyphon
having 30% filling ratio analyze for increasing heat input.
Heat input can be calculated as,

Qin = VI

Heat output can be calculated as,

Qout = mc (T12 − T11)

The efficiency of TPCT is

η = Qout /Qin

m—mass flow rate in kg/s.

c—specific heat of water, 4.187 kJ/kg k.
T11—water inlet temp.
T12—Water outlet temp.
V—voltage in volt.
I—current in Ampere.
Figure 2 shows the plot of surface temperature at different heat input. The temper-
ature of TPCT increases with increase in heat input. T1, T2, T3 are the nearly constant
evaporator temperature and appear as a straight line. So it can be concluded that
heat addition is an isothermal process. Adiabatic section temperature is continu-
ously decreased and appeared as a downward incline line. It signifies that vapour
of acetone rejects the heat to liquid acetone coming from condenser. Condenser
temperature T7, T8 and T9 are constant and follow a straight horizontal line. So
like heat addition, heat rejection is an isothermal process. Maximum temperature of
evaporator and condenser observe during testing are 90.5 °C and 43.2 °C. Figure 3 is
the plot of (Te-Tc) versus heat input. It shows that the mean temperature difference
between evaporator and condenser is gradually increased with an increase in heat
input. (Te-Tc) for 50, 100, 150, 200, 250 and 300 W is 12.7, 20.34, 26.6, 31.64,
38.43 and 45.4 °C respectively. As the heat input increases the average temperature
of the evaporator is increases and the same for a condenser. There is a linear rela-
tionship between (Te-Tc) versus heat input. Figure 4 is the plot of heat input versus
efficiency. The efficiency of TPCT is increasing with an increase in heat input. The
average efficiency of TPCT is 87.7%. The efficiency of TPCT is higher at high heat
input. Range of efficiency is 80–93%.
Thermal Performance of Two Phase … 1009

Fig. 2 Plot of temperature profile

Fig. 3 Plot of heat input versus (Te-Tc)

4 Conclusion

In this paper, the thermal performance of two-phase closed thermosyphon is analyzed

with acetone as a working fluid. The mean temperature difference between evaporator
and condenser and thermal efficiency of TPCT is determined and plot.
There is a linear relationship between heat input and the mean temperature differ-
ence between evaporator and condenser. The thermal efficiency of TPCT is increasing
1010 S. V. Pawar and A. K. Chandgude

Fig. 4 Plot of Heat input versus efficiency

with the increase in heat input. The percentage Change in efficiency when heat input
increases from 50 W to 300 W is 17.67%. Thermal performance of TPCT is best at
high heat input.

References

1. Shabgard H, Xiao B (2014) Thermal characteristics of a closed thermosyphon under various

filling conditions. Int J Heat Mass Transf 70:91–102
2. Kamyar A, Ong KS, Saidur R (2013) Effects of nanofluids on heat transfer characteristics of
a two-phase closed thermosyphon. Int J Heat Mass Transf 65:610–618
3. Rahimi M, Asgary K, Jesri S (2010) Thermal characteristics of a resurfaced condenser and
evaporator closed two-phase thermosyphon. Int Commun Heat Mass Transf 37:703–710
4. Tsai T-E, Wu H-H, Chang C-C, Chen S-L (2010) Two-phase closed thermosyphon vapour-
chamber system for electronic cooling. Int Commun Heat Mass Transf 37:484–489
5. Brusly Solomon A, Mathew A, Ramachandran K (2013) Thermal performance of anodized
two-phase closed thermosyphon (TPCT). Exp Therm Fluid Sci 48:49–97
6. Ong KS, Haider-E-Alahi M (2003) Performance of an R-134a-filled thermosyphon. Appl
Therm Eng 23:2373–2381
7. Sarafraz MM, Hormozi F, Peyghambarzadeh SM (2015) Role of nanofluid fouling on the
thermal performance of a thermosyphon: Are nanofluids reliable working fluid? Appl Therm
Eng 82:212–224
8. Noie SH (2005) Heat transfer characteristics of a two-phase closed thermosyphon. Appl Therm
Eng 25:495–506
9. Ong KS, Tong WL (2011) Inclination and fill ratio effect on water-filled TPCT, IHPS
Thermal Performance of Two Phase … 1011

10. Jiao B, Qui LM (2008) Investigation on the effect of filling ratio on steady-state heat transfer
performance of a vertical TPCT. Appl Therm Eng 28:1417–1426
11. Ordaz-Flores A (2012) Finding to improve the performance of a two-phase flat plate solar
system using acetone and methanol as working fluid. Solar Energy 86:1089–1098
Effect of Slip on Vortex Formation Near
Two-Part Cylinder with Same Sign Zeta
Potential in a Plane Microchannel

Souvik Pabi, Sumit Kumar Mehta, and Sukumar Pati

Abstract We investigate the effect of slip on the formation of recirculation zone near
the two-part cylinder with the same sign zeta potential placed in a microchannel. The
external electric field is used to actuate the electroosmotic flow (EOF). The governing
transport equations are solved using a finite element based numerical solver. The
vortex formation takes place near the upstream part of the cylinder. The strength of
the vortex is analyzed in terms of the maximum magnitude of reversed flow velocity
(UR ). It is found that the extent of the recirculation zone is smaller for the slip case as
compared to the no-slip case. The magnitude of UR increases with the slip coefficient
(β) for smaller values of β. Also there is a decrement in UR at larger values of slip
coefficient and the decrement is amplified at higher values of zeta potential. The flow
rate monotonically increases with the slip coefficient and zeta potential.

Keywords Electroosmotic flow · Interfacial slip · Recirculation zone

1 Introduction

Microfluidics is a science of controlling the behavior of fluids at the micro-scale level

and finds a wide range of applications in the field of biomedical diagnostics, drug
delivery, DNA analysis, microsensors, micro-electro-mechanical-system (MEMS),
microactuators [1–4]. Lab on a chip devices are favorable for microfluidics applica-
tions due to its low cost, rapid sample analysis, minimal consumptions of the sample,
automation, and portability [5, 6]. There are various flow actuation mechanisms in
microfluidics for pumping samples such as surface-tension driven flow, electroki-
netic flow. Electroosmotic flow (EOF) is a type of electrokinetic flow in which an
applied electric field generates a net force on the electric double-layer (EDL) due
to the movement of ionized electrolyte solution over a stationary charge surface [7,
8]. In microchannel, it is quite difficult to get rapid mixing due to low flow velocity
and smaller channel dimensions which makes the flow laminar with Re  1. In

S. Pabi (B) · S. K. Mehta · S. Pati

Department of Mechanical Engineering, National Institute of Technology Silchar, Silchar 788010,
Assam, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 1013
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_101
1014 S. Pabi et al.

an effort to enhance and control mixing, different methods are employed such as
active and passive mixing techniques. In active mixers, an external energy other than
flow energy is applied to perturb the flow and creates vortices [9–11]. On the other
hand, a passive mixer uses confined geometries in the form of rivets, grooves in the
channel to create vortices and enhance its vorticity for proper mixing [12–14]. In
passive micromixers, no external energy source is applied and it requires only the
pumping energy for flow actuation. On the other hand, active micromixers are more
efficient in micromixing although quite difficult to fabricate [15]. Over the years
researchers have been working on the formation of vortices in microfluidic devices.
Wang et al. [6] investigated numerically the formation of vortices near a two-part
cylinder with the same sign of zeta potential but different in magnitudes under a
DC electric field. Mondal et al. [15] investigated numerically the mixing character-
istics in raccoon and serpentine micromixers. Soleymani et al. [16] investigated the
fluid mixing in T-type micromixers and showed that the formation of the vortices is
essential to enhance better mixing performances and the vortex formation strongly
depends on the geometrical parameters of the micromixer. Erickson and Li [17–19]
carried out a study to analyze the formation of vortices in a straight microchannel
wall under electroosmotic flow with different sign or polarity of zeta potential. Song
et al. [20] investigated both numerically and experimentally the flow field in a two-
section straight microchannel of the same polarity and different magnitude of zeta
potential. As in macroscopic level no-slip condition is valid, for microscopic flow
regime, studies exhibit the existence of slip boundary condition for hydrophobic
PDMS surfaces [21]. From the literature, it is seen that there is no work on the
analysis of the formation of vortices due to the same polarity of zeta potential and
interfacial slip. Therefore, we set our objective to investigate the effect of slip on the
vortex formation in a microchannel with the same sign zeta potential of the cylinder,
placing inside it. In the present study, the slip boundary conditions on the wall in the
microchannel [22–29] and the cylinder are considered in order to determine the flow
rate and recirculation reversed flow velocity near the cylinder surface accounting the
effect of vortices.

2 Theoretical Formulation

We consider flow of symmetric (z : z) electrolytic (Newtonian) through a straight

microchannel consisting of a two-part cylinder with zeta potential of same sign [6]
placed in the center as shown in Fig. 1. The downstream and upstream zeta potentials
are ζD∗ and ζU∗ , respectively. The inlet channel height is H and the total channel length
is L = 10H . The diameter of the cylinder is taken as d = 0.5H . The flow is laminar,
steady, and incompressible with constant properties. The channel width is much
larger than H and therefore, flow is assumed as two-dimensional [6]. The effect of
Joule heating is neglected as well as the charge is assumed as a point. The order of
ionic Peclet number is very small (1) and therefore, the charge distribution in EDL
is independent of the convection field [20]. Apart from that the thickness of EDL is
Effect of Slip on Vortex Formation … 1015

Fig. 1 Physical model and the coordinate system

smaller than the radius of the cylinder (λD /r  1), therefore the curvature effect
on the charge distribution is neglected. Here, λD is the characteristic EDL thickness
and r is the radius of the cylinder. The dimensionless governing equations (Laplace,
Poisson-Boltzmann, Continuity and Momentum) can be written as [6, 20]:

∇2ϕ = 0 (1)

∇ 2 ψ = κ 2 sinh ψ (2)

∇ ·U=0 (3)

    
Re(U · ∇)U = −∇P + ∇ 2 U + κ 2 sinh ψ ∇ (ψ Ω) + ϕ (4)

Here,
 ϕ is
 the
 dimensionless external electric potential, normalized with the scale
∗
ϕref = V L ; V is the external potential difference between the channel ends;
∗
  
ψ is the normalized EDL potential normalized with the scale ψref = kB To ze ;
kB , To , z and e are the Boltzmann constant,reference temperature,
  valance
 on ions
and charge on the electron, respectively; U = U î + V ĵ = uî + v ĵ uref is the
   
dimensionless velocity vector,  u ref is the reference
  velocity, ∇ = ∂ ∂X , ∂ ∂Y ;
  ∗
X = x H ; Y = y H ; uref = εo εr Eo ψref μ is the reference velocity; Eo =
      
V L; P is dimensionless
  pressure scaled by μu ref H ; κ = H λD is the Debye
 −1/2
parameter and λD = 2n◦ z 2 e2 εo εr kB T is the characteristic EDL thickness.
Here, n◦ is the nominal ionic bulk concentration, εr is the relative permittivity
 of the
medium and εo is the absolute permittivity of the free space; Re = ρuref H μ is
1016 S. Pabi et al.

the Reynolds number; Ω is the ratio of applied potential difference to the reference
EDL potential and taken as 100. It corresponds to the external potential difference
of 25 V and reference EDL potential 0.025 V, respectively.
The following boundary conditions are imposed.
At the inlet

ϕ = 10; n · (∇ψ) = 0; PG = 0 (5)

At the wall

n · (∇ϕ) = 0; ψ = ζD and ζU for downstream and upstream part of cylinder;

n · (∇ψ) = 0 for other wall (6a)

U = β(∇U) (6b)

Here, ζU = −ζ and ζD = −ζ  5. ζ is the upstream side dimensionless zeta
∗
potential scaled by ψref , β = β ∗ H is the slip coefficient and β ∗ is the slip length.
At the outlet

ϕ = 0; n · (∇ψ) = 0; PG = 0 (7)

The dimensionless flow rate at the outlet can be calculated as:

0.5
Q= Udy (8)
−0.5

3 Numerical Method and Validation

Finite element based numerical technique is used to solve the governing equations
[10]. The relative convergence criterion taken in this simulation is set to 10−6 . An
extensive grid independence test is conducted and the selected mesh system is M3
with a relative difference in reverse flow velocity less than 1% (Table 1).
Validation is made with the results of Zhao et al. [30] as presented in Fig. 2 at
κ = 10. The close agreement of the present results with reported in [30] establish
the accuracy of the current results.
Effect of Slip on Vortex Formation … 1017

Table 1 Grid independence

Mesh system (M) Number of UR Difference (%)
test at different mesh system
elements
(M) by calculating the
characteristics recirculation M1 21826 2.5165 3.66
zone velocity at ζ = 4, M2 50352 2.5593 2.025
β = 0.1 and κ = 100
M3 77070 2.6092 0.11
M4 100048 2.6122 –

Fig. 2 Comparison of
dimensionless flow velocity
with the results of Zhao et al.
[30] for the electroosmotic
flow through a plane
microchannel at κ = 10

4 Results and Discussion

The objective of this work is to analyze the effect of interfacial slip on the forma-
tion of vortex near a two-part cylinder with zeta potential of same sign in a plane
microchannel. The values of Reynolds number (Re) and Debye parameter (κ) are
fixed as Re = 0.01, κ = 100, respectively [6, 20–22]. The dimensionless zeta poten-
tial (ζ ) is varied in the range of 1 ≤ ζ ≤ 4, which is equivalent to in the range
of 25–100 mV [6, 20–22]. The slip length varies in the order of 1 nm to 1 μm
[7], and therefore, the value of β is varied in the range of 0–0.1 for the microflu-
idic devices. The results are analyzed in terms of dimensionless flow velocity (U ),
streamlines, recirculation reversed
  flow velocity
  (Q), recirculation
(UR ), flow rate 
reversed flow velocity ratio UR UR(β=0) and flow rate ratio Q Qβ=0 .
Figure 3a and b shows the streamlines and dimensionless flow velocity (U) at
β = 0 and β = 0.1, respectively. The vortex is formed near the upstream side of
the cylinder. It is because of the larger upstream zeta potential allows larger EOF
actuation along with the higher local electric field intensity near the channel wall and
upper part (or lower tip) of the cylinder. Therefore, to follow the mass conservation,
reversed flow is induced near the microchannel wall in upstream. It can be noted that
the effect of slip enhances the flow velocity and reversed flow velocity magnitude
(see Fig. 3a, b). The corresponding flow velocity profile is presented in Fig. 3c and d
1018 S. Pabi et al.

Fig. 3 Dimensionless flow velocity and streamlines contours at a β = 0 and b β = 0.1 with ζ = 4.
Variation of the dimensionless flow velocity profile at X = 4.7 for different zeta potential, near the
cylinder at c β = 0 and d β = 0.1

at different ζ for β = 0 and 0.1, respectively at the section X = 4.7. It can be noted
that the flow velocity and reversed flow velocity magnitude enhances with ζ and
β due to the increase in EOF actuation and flow resistance due to hydrophobicity,
respectively. It is also noted that the extent of the recirculation zone (vertical length
of negative flow velocity) is larger for the no-slip case as found Y = 0.156 and 0.119,
respectively for β = 0 and 0.1. It is because of the enhancement in primary flow for
the slip case and thus suppresses the extent of the flow reversal zone. It can be noted
that this length is fixed for any values of ζ .
Figure 4 shows the variation of reversed flow velocity (UR ) with β for different
ζ . The value of UR increases rapidly with β at its smaller values for all ζ . While
for higher values of β, a reduction in UR is observed, and the reduction is more
significant for higher values of ζ . It can be explained as follows. For smaller values
of β, UR enhances due to the presence of slip and for larger values of β and ζ , the
amplification in primary flow velocity suppresses the strength of UR . The percentage
increases in UR for slip case with β = 0.1 as compared to the no-slip case are found as
207.54%, 245.94%, 309.08%, and 394.56%, for ζ =  1,2, 3, 4, respectively.
 Figure 5
shows the variation of reversed flow velocity ratio UR UR(β=0) with β at different
values of ζ . It can
 be observed that this
 ratio is greater than unity beyond the critical
slip coefficient βCri ≈ 2.97 × 10−4 . The value of this ratio increases with β for
Effect of Slip on Vortex Formation … 1019

Fig. 4 Variation of UR with

β at different ζ

Fig.5 Variation of
UR UR(β=0) with β at
different ζ

β > βCri up to β ∼ 0.08 and beyond that limit the reverse flow velocity ratio
decreases with β.
Figure 6 shows the variation of flow rate (Q) with β at different values of ζ .
The value of Q increases with β and ζ due to the increase
  in primary flow velocity
(see Fig. 3c, d). Figure
 7 shows the variation of Q Qβ=0 with β for different
ζ . The value of Q  Qβ=0 is greater than unity beyond a critical slip coefficient
βCri ≈ 4.8 × 10−4 . The value of Q Qβ=0 monotonically enhances with β for β >
βCri at all values of ζ .
1020 S. Pabi et al.

Fig. 6 Variation of Q with β

at different ζ


Fig. 7 Variation of Q Qβ=0
with β at the different ζ

5 Conclusions

The present paper investigates the effect of slip on the formation of vortex near a
two-part cylinder with same sign of the zeta potential placed in a microchannel. The
results are presented in terms of flow velocity, flow rate, and reverse flow velocity
by varying the slip coefficient and zeta potential. The key findings of this study are
as follows:
• The extent of the recirculation zone is smaller for the slip case as compared to the
no-slip case.
Effect of Slip on Vortex Formation … 1021

• The value of characteristic reverse flow velocity (UR ) increases with the slip
coefficient (β) for its smaller values. While a decrement in UR is observed at larger
values of β. This decrement is amplified at the higher values of zeta potential.
• The flow rate monotonically increases with the slip coefficient and zeta potential.

References

1. Zhao G, Jian Y, Li F (2016) Streaming potential and heat transfer of nanofluids in microchannels
in the presence of magnetic field. J Magn Magn Mater 407:1–25
2. Ohno K, Tachikawa K, Manz A (2008) Microfluidics: applications for analytical purposes in
chemistry and biochemistry. Electrophoresis 29(22):4443–4453
3. Becker H, Gartner C (2000) Polymer microfabrication methods for microfluidic analytical
applications. Electrophoresis 21(1):12–26
4. Nandy K, Chaudhuri S, Ganguly R, Puri IK (2008) Analytical model for the magnetophoretic
capture of magnetic spheres in microfluidic devices. J Magn Magn Mater 320:1398–1405
5. Ng TN, Chen XQ, Yeung KL (2015) Direct manipulation of particle size and morphology of
ordered mesoporous silica by flow synthesis. RSC Adv 5(18):13331–13340
6. Wang C, Song Y, Pan X (2020) Electrokinetic-vortex formation near a two-part cylinder with
same-sign zeta potentials in a straight microchannel. Electrophoresis 00:1–9
7. Liu Y, Jian Y (2019) The effects of finite ionic sizes and wall slip on entropy generation in
electroosmotic flows in a soft nanochannel. J Heat Transf 141:1–11
8. Chakraborty J, Pati S, Som SK, Chakraborty S (2012) Consistent description of electrohydro-
dynamics in narrow fluidic confinements in presence of hydrophobic interactions. Phys Rev E
85:046305
9. Mehta SK, Pati S (2021) Thermo-hydraulic and entropy generation analysis for magnetohydro-
dynamic pressure driven flow of nanofluid through an asymmetric wavy channel. Int J Numer
Method Heat Fluid Flow 31(4):1190–1213
10. Mehta SK, Pati S (2019) Analysis of thermo-hydraulic performance and entropy generation
characteristics for laminar flow through triangular corrugated channel. J Therm Anal Calorim
136(1):49–62
11. Mehta SK, Pati S (2020) Numerical study of thermo-hydraulic characteristics for forced
convective flow through wavy channel at different Prandtl numbers. J Therm Anal Calorim
141:2429–2451
12. Nguyen NT, Wu Z (2005) Micromixers-a review. J Micromech Microeng 15(2):R1–R16
13. Maeng JS, Yoo K, Song S (2006) Modeling for fluid mixing in passive micromixers using the
vortex index. J Korean Phys Soc 48(5):902–907
14. Shang X, Huang X, Yang C (2015) Mixing enhancement by the vortex in a microfluidic mixer
with actuation. Exp Therm Fluid Sci 67:57–61
15. Mondal B, Mehta SK, Patowari PK, Pati S (2019) Numerical study of mixing in wavy
micromixers: comparison between raccoon and serpentine mixer. Chem Eng Process Process
Intensif 136:44–61
16. Soleymani A, Kolehmainen E, Turunen I (2008) Numerical and experimental investigations of
liquid mixing in T-type micromixers. Chem Eng J 135S:S219–S228
17. Erickson D, Li D (2003) Three-dimensional structure of electroosmotic flow over heteroge-
neous surfaces. J Phys Chem B 107(44):12212–12220
18. Erickson D, Li D (2002) Influence of surface heterogeneity on electrokinetically driven
microfluidic mixing. Langmuir 18(5):1883–1892
19. Erickson D, Li D (2002) Microchannel flow with patchwise and periodic surface heterogeneity.
Langmuir 18(23):8949–8959
1022 S. Pabi et al.

20. Song Y, Wang C, Li J, Li D (2019) Vortex generation in electroosmotic flow in a straight

polydimethylsiloxane microchannel with different polybrene modified-to-unmodified section
length ratios. Microfluid Nanofluid 23(2):23:24
21. Niavarani A, Priezjev NV (2009) The effective slip length and vortex formation in laminar flow
over a rough surface. Phys Fluids 21(5):1–10
22. Pati S, Som SK, Chakraborty S (2013) Thermodynamic performance of microscale swirling
flows with interfacial slip. Int J Heat Mass Transf 57:397–401
23. Kaushik P, Pati S, Som SK, Chakraborty S (2012) Hydrodynamic and thermal transport char-
acteristics of swirling flows through microchannels with interfacial slip. Int J Heat Mass Transf
55:4359–4365
24. Kaushik P, Pati S, Som SK, Chakraborty S (2012) Hydrodynamic swirl decay in microchannels
with interfacial slip. Nanoscale Microscale Thermophys Eng 16:133–143
25. Pati S, Som SK, Chakraborty S (2013) Slip-driven alteration in film condensation over vertical
surfaces. Int Commun Heat Mass Transf 46:37–41
26. Dwivedi R, Pati S, Singh PK (2020) Combined effects of wall slip and nanofluid on interfacial
transport from a thin-film evaporating meniscus in a microfluidic channel. Microfluid Nanofluid
24:84
27. Pati S, Kaushik P, Chakraborty S, Som SK (2014) Film condensation in presence of non-
condensable gases: interplay between variable radius of curvature and interfacial slip. Int
Commun Heat Mass Transf 56:31–36
28. Pati S, Kumar V (2019) Effects of temperature-dependent thermo-physical properties on
hydrodynamic swirl decay in microtubes. Proc Inst Mech Eng Part E: J Process Mech Eng
233(3):427–435
29. Pati S, Som SK, Chakraborty S (2013) Combined influences of electrostatic component of
disjoining pressure and interfacial slip on thin film evaporation in nanopores. Int J Heat Mass
Transf 64:304–312
30. Zhao C, Zholkovskij E, Masliyah JH, Yang C (2008) Analysis of electroosmotic flow of power-
law fluids in a slit microchannel. J Colloid Interface Sci 326(2):503–510
Optimization of WEDM Parameters
During Machining of Ni-75 Using
AHP-MOORA Method

S. A. Sonawane and S. S. Wangikar

Abstract The objective of this study is to perform multi-characteristic optimization

of WEDM parameters while machining Nimonic -75 (Ni-75) alloy using an integrated
approach of AHP-MOORA method. Taguchi’s L27 orthogonal matrix is used to
conduct the experiments. The cutting factors selected for this research work are the
pulse on time (Pon), pulse off time (Poff), gap voltage (GV), peak current (IP),
wire velocity (WV ) and wire tension (WT ) while the outcomes are MRR, surface
roughness, and kerf width. AHP method is used to find out weights for the chosen
quality characteristics, and MOORA technique is employed to determine the most
favorable cutting conditions. The optimum combination of explanatory factors are
Pon = 110 machine units (mu), Poff = 51 mu, GV = 40 V, IP = 230 Amp., WV =
5 m/min and WT = 8 mu. Finally, authentication trials are accomplished to validate
the results.

Keywords WEDM · Taguchi method · Optimization · MRR · Surface roughness ·

Kerf width

1 Introduction

Nimonic-75 (Ni-75) is an 80-20 Ni-Cr alloy with controlled additions of titanium and
carbon. It can be easily manufactured and welded. It possesses excellent corrosion
resistance, mechanical properties, and heat resistance. It finds extensive applications
in aerospace fasteners, heat treatment equipment, gas turbine engineering, nuclear
engineering, industrial furnace structural parts, etc.
There are many challenges faced during traditional machining of the Nimonic-75
alloy such as notching of tool nose, confrontation to forced deformation required for
chip formation, miniature cavities created on the rake face, dissolution and diffusion
wear of the rake face of the tool, abrasive wear of the tool, etc. [1, 2]. Wire electrical
discharge machining (WEDM) is one of the non-traditional machining processes
which can be effectively employed to machine such a challenging to cut alloys. It

S. A. Sonawane (B) · S. S. Wangikar

SVERI’s College of Engineering, Pandharpur 413304, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 1023
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_102
1024 S. A. Sonawane and S. S. Wangikar

is an electro-thermal non-established machining method, where electrical power is

applied between the tool (wire) and the workpiece to produce an electrical sparkle,
and material deletion essentially occurs due to the thermal force of the flash. It plays
a vital role in different industries such as aerospace, automobile, nuclear, computer
and electronics and tool and dies industries to cut multiple, intricate profiles.
Chakraborty [3] applied moora method to different problems of manufacturing and
found that this method is more flexible. Gadakh [4] applied moora method to various
milling processes and found that the results obtained by this method were similar
to that achieved by other researchers. Gadakh et al. [5] used moora technique to the
various welding processes and found comparatively comparable results with those
of the previous researchers. Rajesh et al. [6] applied moora based entropy method to
optimize the parameters during the drilling process of red mud reinforced Al MMCs.
Singaravel et al. [7] employed a combination of moora and entropy scheme for
optimization of factors whilst turning of EN25 steel. Saha and Mondal [8] used PCA
based moora system for the optimization of WEDM of Nano-structured hardfacing
alloy. Kalirasu et al. [9] applied moora method during AWJM of Jute/Polyester
composite. Majumdar and Maity [10] carried out WEDM of Ti Grade 6 and used PCA
based moora method for optimization. Anand babu et al. [11] performed optimization
using AHP-DENG’S method during WEDM of Al-SiC MMC. Mohapatra and Sahoo
[12] used moora means for the optimization of pitch error and MRR during gear
cutting process using WEDM. This paper addresses the machining study of Ni-75
using WEDM. Taguchi’s L27 orthogonal run is used to carry out the experiments, and
AHP based moora method is used for multi-features optimization. The performance
of the machining is evaluated concerning metal removal rate kerf width (KW ), surface
roughness (SR) and (MRR).

2 Experimentation

In the current research, experiments are performed on Electronica ePULSE sprint

cut wire EDM. Six explanatory variables, namely pulse on period (Pon), pulse off
period (Poff), gap voltage (GV), peak current (IP), wire feed velocity (WV) and
cable tension (WT) are considered for the study, and the outcomes are MRR, surface
roughness, and kerf width. The range of input variables is decided by performing the
preliminary tests. A 0.25 mm diameter brass wire is used to cut a block of size 10 mm
× 10 mm × 5 mm thick from a workpiece of Nimonic-75 of dimension 30 mm ×
30 mm × 5 mm thick. Figure 1 shows the workpiece after machining.
Taguchi’s L27 array is used to conduct the experiments. Each test is replicated
thrice. Table 1 indicates the Taguchi’s orthogonal matrix and Table 2 represents the
results of the orthogonal tests. The cut characteristics considered for this machining
study are MRR, surface roughness (Ra) and kerf width (KW). MRR is computed by
the volume of material removed concerning machining time. The surface roughness
(Ra) is measured with the help of Mitutoyo surface tester SJ-210 and kerf width by
using a profile projector of least count 0.1 microns.
Optimization of WEDM Parameters … 1025

Fig. 1 WEDMed workpiece

Table 1 Taguchi’s L27 array

Exp. No. Pon mu Poff mu GV V IP Amp WV m/min WT mu
1 110 41 20 170 1 2
2 110 41 20 170 3 5
3 110 41 20 170 5 8
4 110 46 30 200 1 2
5 110 46 30 200 3 5
6 110 46 30 200 5 8
7 110 51 40 230 1 2
8 110 51 40 230 3 5
9 110 51 40 230 5 8
10 114 41 30 230 1 5
11 114 41 30 230 3 8
12 114 41 30 230 5 2
13 114 46 40 170 1 5
14 114 46 40 170 3 8
15 114 46 40 170 5 2
16 114 51 20 200 1 5
17 114 51 20 200 3 8
18 114 51 20 200 5 2
19 118 41 40 200 1 8
20 118 41 40 200 3 2
21 118 41 40 200 5 5
22 118 46 20 230 1 8
23 118 46 20 230 3 2
24 118 46 20 230 5 5
25 118 51 30 170 1 8
26 118 51 30 170 3 2
27 118 51 30 170 5 5
1026 S. A. Sonawane and S. S. Wangikar

Table 2 Results of
Exp. No. MRR mm3 /min SR µm KW mm
orthogonal tests
1 33.30 1.82 0.285
2 27.60 1.68 0.269
3 26.60 1.6 0.265
4 31.16 1.76 0.279
5 28.43 1.7 0.27
6 26.64 1.61 0.266
7 32.25 1.79 0.28
8 25.67 1.58 0.26
9 24.53 1.52 0.254
10 55.00 2.32 0.307
11 48.63 2.19 0.297
12 33.70 1.85 0.286
13 52.37 2.28 0.305
14 32.92 1.79 0.285
15 37.62 2.01 0.289
16 34.00 1.95 0.287
17 39.20 2.06 0.293
18 36.60 1.97 0.288
19 58.70 2.51 0.327
20 38.90 2.04 0.291
21 56.34 2.41 0.313
22 44.37 2.1 0.295
23 55.20 2.36 0.309
24 50.90 2.25 0.299
25 57.60 2.5 0.322
26 54.50 2.31 0.305
27 29.20 1.3 0.275

3 Multi-feature Optimization by Using AHP-MOORA

Method

In this study, a multi-criterion assessment-making approach MOORA integrated with

AHP method is applied for the optimization of various performance features while
WEDM of Ni-75. A range of steps entailed in this approach are as follows:
Optimization of WEDM Parameters … 1027

3.1 Multi-goal Optimization on the Foundation of Proportion

Analysis (MOORA)

A variety of stages occupied in the MOORA method are as below:

Step 1 Define the problem.
This step is concerned with defining the intent and to organize each option
and its features.
Step 2 Creation of decision array (Zij) for various options and its features.
Step 3 Standardization of performance features.
The different quality characteristics in the conclusion matrix are stabilized
by using Eq. 1. The standardized decision matrix (Nij) is given by

Zij
Nij = √  (1)
(Zij)2

Step 4 For multi-characteristic optimization, the standardized quality features are

inserted in case of maximization (for favorable objectives) and deducted
in case of minimization (for non-favorable objectives). Then the overall
assessment value (Yi) is given by Eq. 2.


g

n
Yi = Nij − Nij (2)
j=1 j=g+1

where g indicates the number characteristics to be increased, (n-g) is the

number of features to be diminished, and Yi is the standardized appraisal
value of the ith option concerning all the features. Generally, it is found
that some features are more significant than the others. Therefore, to give
more significance to the characteristic, it is multiplied by its related weight
[13]. When these specific weights are taken into account, Eq. 2 becomes:


g

n
Yi = WjNij − WjNij (3)
j=1 j=g+1

where, Wj is known as the weight of the jth characteristic, which is deter-

mined by applying the Analytical Hierarchy Process (AHP) method in this
study.
Step 5 The Yi values are graded, and the highest value of Yi indicates most
excellent option whereas the lowest worth stands for the most awful.
1028 S. A. Sonawane and S. S. Wangikar

3.2 Analytical Hierarchy Process (AHP)

The AHP is derived from the knowledge acquired by its inventor, Saaty [13] while
steering research assignments in the US Arms Control and Disarmament organiza-
tion. The AHP presents a means of disintegrating the objective into a hierarchy of
associate difficulties which can more easily be figured out and subjectively estimated.
The subjective appraisals are renovated into numerical assessments and processed to
grade each choice on a numerical degree. The scheme of the AHP can be elucidated
in the subsequent steps:
Step 1 The hitch is decayed into a chain of command of objectives, criterion,
sub-criteria and options.
Step 2 Devise the pronouncement matrix.
The verdict matrix is created containing all the quality characteristics
indicated in quantitative or numerical value.
Step 3 Engender pair wise matrices
A couple wise association matrix is created using a level of comparative
importance as shown in Table 2. The oblique elements of the array are 1.
The measure in the ith row is better than the measure in the jth column if
the rate of the aspect (i, j) is in excess of 1; otherwise, the measure in the
jth line is superior than that in the ith chain. The (j, i) element of the array
is the mutual of the (i, j) constituent.
Step 4 Determine the eigenvectors using a geometric mean method.
Step 5 Calculate the weights by normalizing the eigenvector by the total sum of
eigenvectors.
By applying the above steps to the values of the output features as shown
in Table 2, the overall assessment value is calculated and is represented in
Table 3.
It is observed that Exp. No. 9 has the highest value of Yi. Thus, the optimum
combination of explanatory factors are Pon = 110 machine units (mu), Poff
= 51 mu, GV = 40 V, IP = 230 Amp., WV = 5 m/min and WT = 8 mu.
The evidence tests are conducted to validate the most favorable parameters
settings. The results of evidence tests are indicated in Table 4.

4 Conclusions

The subsequent conclusions can be drawn from the current learning:

1. The MOORA method integrated with AHP is used to determine most advanta-
geous settings of machining factors during WEDM of Ni-75.
2. The weights of the quality features are determined by applying AHP method
which depends upon individual judgment.
Optimization of WEDM Parameters … 1029

Table 3 Overall assessment

Exp. No. Overall assessment value (Yi) Grade
value
1 −0.1665 11
2 −0.1574 6
3 −0.1542 4
4 −0.163 8
5 −0.1581 7
6 −0.1548 5
7 −0.1637 9
8 −0.1516 2
9 −0.1478 1
10 −0.1813 23
11 −0.1756 19
12 −0.1675 12
13 −0.1803 21
14 −0.166 10
15 −0.1711 15
16 −0.1699 13
17 −0.1736 17
18 −0.1701 14
19 −0.1939 27
20 −0.1723 16
21 −0.1857 25
22 −0.174 18
23 −0.183 24
24 −0.1771 20
25 −0.1915 26
26 −0.1803 22
27 −0.1524 3

Table 4 Results of evidence

Responses Optimal values
tests
Predicted value Experimental value
MRR 24.53 26.42
SR 1.52 1.48
KW 0.254 0.252
1030 S. A. Sonawane and S. S. Wangikar

3. The optimum permutation of input factors obtained are Pon = 110 machine
units (mu), Poff = 51 mu, GV = 40 V, IP = 230 Amp., WV = 5 m/min and WT
= 8 mu.

References

1. Ezugwu EO, Wang ZM, Machado AR (1999) The machinability of nickel-based alloys: a
review. J Mater Process Technol 86(1–3):1–16
2. Chaudhary IA, El-Baradie MA (1998) Machinability of nickel-based super alloys: a general
review. J Mater Process Technol 77(1–3):278–284
3. Chakraborty S (2011) Applications of the MOORA method for decision making in manufac-
turing environment. Int J Adv Manuf Technol 54(9–12):1155–1166
4. Gadakh VS (2011) Application of MOORA method for parametric optimization of milling
process. Int J Appl Eng Res 1(4):743–758
5. Gadakh VS, Shinde VB, Khemnar NS (2013) Optimization of welding process parameters
using MOORA method. Int J Adv Manuf Techno 69(9–12):2031–2039
6. Rajesh S, Pethuraj M, Thirumalai Kumaran S, Uthaykumar M, Rajini N (2015) Some studies
on drilling of red mud reinforced aluminium composite. Proc Inst Mech Eng Part L J Mater
Des Appl 231(4):382–393
7. Singaravel B, Selvaraj T, Vinodh S (2016) Multi-objective optimization of turning parameters
using the combined MOORA and Entropy method. Trans Can Soc Mech Engg 40(1):101–111
8. Saha A, Mondal SC (2017) Machining optimization of nano-structured hardfaced tool insert in
WEDM using MOORA method. In: Interanational conference on research into design. Springer
Publication, pp 905–917
9. Kalirasu S, Rajini N, Rajesh S, Winowlin Jappes JT, Karuppasamy K (2017) AWJM perfor-
mance of Jute/Polyester composite using MOORA and analytical models. Mater Manuf Process
32(15):1730–1739
10. Majumdar H, Maity K (2017) Optimization of machining condition in WEDM for Titanium
Grade 6 using MOORA coupled with PCA-A multivariate hybrid approach. J Adv Manuf Syst
16(2):81–99
11. Anand babu K, Venkataramaiah P, Dileep P (2017) AHP-DENG’S similarity based optimization
of WEDM process parameters of Al/SiCp composite. Am J Mater Sci Technol 6(1):1–14
12. Mohapatra KD, Sahoo SK (2018) Optimization of single pitch error and MRR in a WEDM gear
cutting process. In: Precision product-process design and optimization. Springer Publication,
pp 285–312
13. Satty TL (2007) Fundamentals of decision making and priority theory with AHP. RWS
Publications, Pittsburg
Structural Analysis of Novel Mini
Horizontal Axis Wind Turbine
(NMHAWT)

Pramod Magade, S. P. Chavan, Sushilkumar Magade, and Vikram Gaikwad

Abstract The wind energy is most of the promising renewable energy source. In
wind turbine technology, the turbine blades play an important role as it directly
comes in contact with the wind. Many researchers have concentrated on improving
the aerodynamic performance of wind turbine blade through testing and theoretical
studies. In general, moderate to high-speed winds, typically from 5 m/s to about
25 m/s are considered favourable for most wind turbines in India. But in rural areas,
wind speed is near about 3–9 m/s. The present investigation aims is to compare the
performance of eight blade novel mini Horizontal Axis Wind Turbine (NMHAWT)
blades of novel airfoils over National Advisory Committee for Aeronautics (NACA)
eight blade profiles in terms of loads and performance. Therefore blade design of
well known series of NACA airfoil was selected from Q blade and Mat-Lab software.
Analytical calculations are help out for the selection of NACA 4418 is suitable for
the comparison. The objective of this paper is to decide the novel profile of the blade
for the development. The comparison is done through software for predicting the
performance of a novel mini horizontal axis wind turbine.

Keywords Wind speed · NACA airfoil · Blade structural analysis

P. Magade (B) · V. Gaikwad

Department of Mechanical Engineering, Zeal College of Engineering and Research, Pune 411041,
India
e-mail: [email protected]
S. P. Chavan
Department of Mechanical Engineering, Annasaheb Dange College of Engineering and
Technology, Sangli 416301, India
S. Magade
Department of Civil Engineering, MIT College of Engineering, Pune 411038, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 1031
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_103
1032 P. Magade et al.

1 Introduction

Renewable energy of wind is one of the cheapest sources for providing electrical
or mechanical power. It can be found everywhere and every time. But wind cannot
produce required power if it has not sufficient strong velocity. The blowing wind
possesses kinetic energy and strikes on wind turbine blades. By this action kinetic
energy are lost and transmit to mechanical energy of wind turbine shaft [1]. The
maximum power that can be extracted is 59.3% of the power in wind. Wind turbine
shaft is coupling with generator. Therefore, wind turbine blade is an energy converter.
Besides the turbine blades’ profile also plays an important role to capture the wind
energy efficiently and effectively [2]. Energy generated from the wind depends on
direction of the blade profiles, airstream and angle of attack [3]. The optimized airfoil
profile is chosen from maximum lift to drag ratio at design Reynolds number and
angle of attack [4]. In order to achieve the maximum power coefficient design of blade
positioned properly [5]. The performance of blade is simulated in detailed by using
analysis [6]. Experimental results has been carried out for various factors like wind
speed, numbers of blade and size of blade. The eight blades model with is appropriate
for designing and it gives around 40 W power output [7]. Above researches gave an
idea of comparison of traditional and novel blade.

2 Design of Blade

The fundamental theory of design and operation of wind turbines is derived based
on a first principles approach using conservation of mass and conservation of energy
in a wind stream. The diameter of the blade decided by the generator specification
and Beltz limits. The wind power equation is shown as follows.

2.1 Power Equation

1
P= ρAv 3 CP (1)
2
where P is the power, ρ is the density of the air, A is the required area of the blade, v
is the velocity of air and Cp is power coefficient. Design calculation help out to find
the area A = 1.13 m2 and diameter D = 1.2 m of the blade.
Structural Analysis of Novel … 1033

2.2 Novel Blade Design Input Parameters

(a) Lift and Drag Force

Lift force is defined to be perpendicular to direction of the oncoming airflow.
The lift force is a consequence of the unequal pressure on the upper and lower
airfoil surfaces. Lift force can be written as
1
FL = ρAv 2 CL Therefore Lift Coefficient ∴ CL = 1.573 (2)
2
Drag force is defined to be parallel to the direction of oncoming airflow. The
drag force is due both to viscous friction forces at the surface of the airfoil and
to unequal pressure on the airfoil surfaces facing toward and away from the
oncoming flow. Drag Force can be written as

1
FD = ρAv 2 Cd Therefore Drag Coefficient ∴ Cd = 0.043 (3)
2
(b) Lift to Drag ratio: In order to obtain high efficiency, it is therefore essential to
use airfoil shaped rotor blades with a very high lift to drag ratio,

cl
Lift to Drag Ratio ∴ = 36.816 (4)
Cd
(c) Reynolds Number
Low Re airfoils is suited for mini wind turbine applications must be designed
to avoid high leading edge suction peak and high adverse pressure gradients
that leads to flow separation.

D×V
Reynolds Number Re = ∴ Re = 500000 (5)
v

Angle of Attack ∴ α = 15◦ (6)

(d) Tip speed ratio is the ratio between the tangential speed of the tip of a blade
and the actual speed of the wind. Higher tip speeds result in higher noise levels
and require stronger blades due to larger centrifugal forces. Also calculated by
λ = 4Πn
from the wind energy math calculation n is no of blade.

∴ λ = 1.5. (7)
1034 P. Magade et al.

Fig. 1 NACA 4418 airfoil

Fig. 2 Q blade screen

2.3 NACA 4418 Airfoil Selection

The optimum airfoil is selected from maximum lift to drag ratio at designed Reynold
number and angle of attack. Therefore Q blade software is used to select the suitable
airfoil based on calculations. For Cl/Cd = 36.819, Alpha = 15° and Reynold number
is 500000 as shown in Figs. 1 and 2.

2.4 Mat-lab Programming

The mat-lab programmed is used to calculate the position, chord length & twist angle
of blade. Input parameter require in mat-lab to calculate the chord length and twist
angle is number of blade’s, capacity of turbine, velocity of wind, Tip speed ratio,
Lift coefficient, Reynold number and Air density. Shown in Table 1.
Base on mat-lab calculation different chord length and twist angle are obtained
but at 15° angle of attack blade gives the maximum lift force. Therefore novel blade
and regular shape blade are designed for 15° angle of attack.

Table 1 Mat-lab calculations

Radius of blade (m) Chord length (m) AOA (°)
0.08700 0.2049533 16.25
0.14100 0.1851145 15.40
0.19500 0.1625817 14.66
0.24900 0.1058759 13.31
0.30300 0.0696264 12.54
0.41100 0.0101380 10.75
Structural Analysis of Novel … 1035

3 Comparison of Structural Design of NACA 4418 Profile

and Novel Mini Horizontal Axis Wind Turbine Blade

The design of structure which is necessary in turbine design. To find out the structural
analysis (von mises) stress and total deformation of 1200 mm diameter with eight
blade [7] geometry of two different blade profiles is compared with each other. The
analysis is done by using ANSYS workbench software.

3.1 Finite Element Analysis

The results of von mises stresses and total deformation of regular shape blade and
novel shape blade is compared by using ANSYS workbench software. The Von
mises stress is also known as equivalent stress. It is used in design work because any
arbitrary three-dimensional stress can be represented as a single positive stress value.
The Von mises stress is the part of the maximum equivalent stress failure theory used
to predict yielding in a ductile material. The von mises stress measurements is in
metric and to determine if the structure has started to yield at any point. In ANSYS
software material properties is necessary. Aluminium material is used for design
of blade and its Density is 2.77 × 10−6 kg /mm3 , Compressive yield strength (Syc )
is 280 MPa, Tensile yield strength (Syt ) is 280 MPa, Ultimate tensile strength is
310 MPa, Modulus of elasticity is 69 GPa and Poisons ratio is 0.34.

3.2 Force Exerted on a Blade by Wind

Now find out the von mises stress and total deformation by using ANSYS workbench
software with the load of 20 N. The following steps demonstrate how to create the
model geometry and find a solution by using ANSYS workbench.
1. Conduct the pre-processing by specifying the element type such as material
properties and meshing. The accuracy that can be obtained from any FEA model
is directly related to the finite element mesh that is used. The finite element mesh
is used to subdivide the CAD model into smaller domains called elements, over
which a set of equations are solved. These equations approx. represent the
governing equation of interest via a set of polynomial functions defined over
each element. As these elements are made smaller and smaller, as the mesh
is refined, the computed solution will approach the true solution. This process
of mesh refinement is a key step in validating any finite element model and
gaining confidence in the software, the model, and the results. The following
Fig. 3 shows the meshing of the blade.
1036 P. Magade et al.

Fig. 3 Meshing of blades

2. Specify the constraint and loads. The fixed support of the in middle side and
load are applied to the blade is 20 N. Following figure shows the stresses occurs
in NACA 4418 and NMHAWT.
The maximum amounts of stresses are developed in the root of blade as shown in
Fig. 4. For the load of 20 N NACA profile blade creates 26.9 MPa stress and for
novel profile it will reach up to 17.059 Mpa. Stresses occur in novel blade profile
is less than NACA blade profile in connection with this the ultimate strength of
aluminum is 40–50 Mpa, Hence the maximum von-Mises stress found in both
the blade is below the ultimate strength of aluminum therefore blade will not
fails under 20 N defined load. It can be conclude from the results that blade
design is safe for 20 N load.
3. Deformation is the change in the physical properties of a material. It can be clas-
sified into two categories i.e. plastic and elastic deformation. The deformation
that is permanent or which remains even after the removal of stress is known as
plastic deformation whereas the deformation that remains only until the force
is applied is known as elastic deformation. The following result shows the total
deformation at different load conditions.

Fig. 4 Equivalent stress at load 20 N

Structural Analysis of Novel … 1037

Fig. 5 Total deformation at load 20 N

Figure 5 shows the deformation occurs in NACA 4418 and NMHAWT for a load
of 20 N.
The maximum amounts of deformation are developed beyond swept area or tip
side of both the blade as shown in Fig. 5. For the load of 20 N NACA profile blade
deform 2.29 mm and for novel profile is up to 2.52 mm. Deformation occur in novel
blade profile is quite less in NACA blade profile.

4 Result and Discussion

The present paper describes the static structural analysis by ANSYS for uniformly
distributed loading. A parameter study is carried out where advances in this field
have allowed the price of wind energy to be competitive with more conventional
sources. The blades are an important component of wind turbines and much research
has been done in this area. Optimization of blade consider for analysis where wind
load of 20 N. Identify its suitability for its application on wind turbine blades and
good agreement is made between results. In two-dimensional aerofoil modeling, the
comparison between the two profiles of blades in structural analysis by ANSYS
results is a reliable benchmark. Stress and deformation of the blade performances
of different aerofoils have been compared. An airfoil of different profile blade for
a horizontal axis wind turbine is designed based on Structure, it is concluded that
the structural analysis is efficient in predicting rotor aerodynamic characteristics.
The structural analysis is performed to evaluate the proposed novel blade design
configuration is the suitable for further development. In this study,
1. Static structural analysis of three, five and eight blade with 1200 mm diameter
NACA4418 and Novel HAWT is checked under various uniform load conditions
as like 10 N, 15 N and 20 N.
2. The amount of stresses and deformation occurs in the three blades NACA 4418
for the load of 10 N is 37.34 Mpa and 4.92 mm.
1038 P. Magade et al.

Table 2 Comparisons of static analysis results of NACA 4418 and NMHAWT

Sr. No. Blade profile Number of Load applied(N) Total Equivalent
blades deformation stresses (MPa)
(mm)
1 NACA 4418 Three UDL of 10 N 4.92 37.34
Five UDL of 15 N 2.95 22.40
Eight UDL of 20 N 2.29 26.98
2 Novel HAWT Three UDL of 10 N 5.86 20.37
Five UDL of 15 N 2.72 14.46
Eight UDL of 20 N 2.52 17.05

3. The amount of stresses and deformation occurs in the five blades NACA 4418
for the load of 15 N is 22.40 Mpa and 2.98 mm.
4. The amount of stresses and deformation occurs in the eight blades NACA 4418
for the load of 20 N is 26.98 Mpa and 2.29 mm.
5. The amount of stresses and deformation occurs in the three blades NMHAWT
for the load of 10 N is 20.37 Mpa and 5.86 mm.
6. The amount of stresses and deformation occurs in the five blades NMHAWT
for the load of 15 N is 14.46 Mpa and 2.72 mm.
7. The amount of stresses and deformation occurs in the eight blades NMHAWT
for the load of 20 N is 17.05 Mpa and 2.52 mm.
8. It was observed that NMHAWT gives very less amount of stress generation
at the root of the blade and negligible amount of deformation of the blade as
compared to NACA 4418.
9. Newly developed NMHAWT gives better results as compared to conventional
blade profile as shown in Table 2. So NMHAWT can be proposed for further
development and testing.

References

1. Guo Y, Cao X, Wang N (2019) A hybridized design of pinwheel by coupling tribo electrification
and electromagnetic induction effects. Elsevier Nano Energy, vol 60, pp 641–648
2. Riyanto,Pambudi NA (2019) The performance of shrouded wind turbine at low wind speed
condition, (ICAE 2018), Hong Kong, china, pp 260–265
3. Tian W, Yang Z, Zhang Q (2017) Bionic design of wind turbine blade based on long-eared owls
airfoil
4. Thu AM, Aung YYH (2016) Design and performance testing of a small-scaled horizontal axis
wind turbine for low wind speed regions
5. Sinha DK, Tummala A (2016) A review on small scale wind turbines. Science direct
6. Elmagid WMA, Mekhali TA (2015) Experimental and CFD of designed small wind turbine
7. Magade Pramod B, Chavan Srirang P, Magade Sushilkumar B (2019) Experimentation on design
and development of mini wind turbine. Int J Innov Technol Explor Eng (IJITEE) 8(11). ISSN:
2278-3075
Fabrication of Tree Type Micro-Mixer
with Circular Baffles Using CO2 Laser
Machining

Sachin R. Gavali, Sandeep S. Wangikar, Avinash K. Parkhe,

and Prashant M. Pawar

Abstract Micro-mixer is a device which is used to carry the mixing of two or more
than two fluids. Any one dimension of it is in micrometer. Due to low Reynolds’
Number Micro-mixer design is challenge for the designers. To address this chal-
lenge, novel methods of mixing enhancement within micro-fluidic devices have been
explored for a variety of applications. For Passive Micro-mixer mixing will be carried
out with the help of geometry of the micro-mixer. In this study fabrication of a tree
type Micro-mixer with Circular baffles using CO2 Laser machining is discussed. The
Circular baffles are responsible for making diversion of flow of the fluid in Conver-
gent & Divergent manner due to which proper mixing of fluids become possible. This
type of micro-mixer is suitable for Bio-medical applications such as Urine testing,
Blood testing, in preparation & in testing of Drugs. This paper focuses on design
and fabrication of tree type Micro-mixer with Circular Baffles. The parametric study
for CO2 laser machining by considering the speed and power as the control param-
eters and the depth as a performance measure for engraving different width micro
channels is presented in this paper. The tree type Micro-mixer with Circular Baffles
fabricated by CO2 laser machining is found to be suitable for master molds used in
soft lithography process.

Keywords Micro-mixer · Reynolds number · Circular baffles · Micro channels

1 Introduction

A device having the size in microns which is employed for mixing the fluids is called
as Micro-mixer. parts used to mix fluids. This device represents a significant tech-
nology to fields like pharmaceutical industry, chemical industry, analytical chemistry,
high-throughput synthesis, and biochemical analysis, as it employs miniaturized parts
in turn the reduced quantities involved in the chemical and/or biochemical processes.
The Active and passive micro-mixers are the main to types of it. The external energy
source like electric or magnetic source is used by active micro-mixers for performing

S. R. Gavali (B) · S. S. Wangikar · A. K. Parkhe · P. M. Pawar

Department of Mechanical Engineering, SVERI’s College of Engineering, Pandharpur, MH, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 1039
P. M. Pawar et al. (eds.), Techno-Societal 2020,
https://doi.org/10.1007/978-3-030-69921-5_104
1040 S. R. Gavali et al.

the mixing of the fluids. There is no any power source utilized in the passive micro-
mixers and the pressure guides the flow of fluids. Mixing in micro-fluidic devices
presents a crucial challenge [1–5]. A lot of research has been done on micro-channels
with different configurations, both experimentally and numerically. Wong et al. [2]
fabricated micro-T-mixersfrom silicon substrate and tested it for investigating their
probability as quick mixing micro-mixers. Orsi et al. [3] has performed numer-
ical simulations for studying the mixing dynamics of two miscible liquids within
a T shaped micro-mixer by comparing the case where the two inlet fluids are both
water (W-W). Tsai et al. [4] investigated the mixing phenomena between ferro-nano
fluids (a water solution with suspended Fe3 O4 nano particles) and water in a Y-type
semi-active micro-mixer. Further, many researchers have reported the performance
analysis of different types of microchannels using various approaches [3–13]. The
photochemical machining process was also used by various researchers for fabrica-
tion of micro-mixers and molds for micro-mixers employing soft lithography process
[14–21]. The CO2 laser machining was employed for fabrication of micro features
and some straight channels by Chavan et al. [22].
In this study, fabrication of a tree type Micro-mixer with Circular baffles is
performed by CO2 laser machining. The Circular baffles are responsible for making
diversion of flow of the fluid in Convergent & Divergent manner due to which proper
mixing of fluids become possible. The design is suitable for not only two fluids but
also for four fluids. This type of micro-mixer is suitable for Bio-medical applications
such as Urine testing, Blood testing, in preparation & in testing of Drugs. This design
is also helpful in testing of liquids used for Food Processing, testing of Milk, alcohol,
wines etc. It is also suitable for Soil testing purposes. The parametric study for CO2
laser machining by considering the speed and power as the control parameters and
the depth as a performance measure for engraving different width microchannels is
presented in this paper.

2 Materials and Methods

2.1 Selection of Material

Acrylic is extensively known for its significance to the industry. It is employed for
making the different products like bath enclosures, shower doors, skylights, and
windows. It is preferred over the glass for different reasons. It is stronger than glass
which makes it impact resistant and hence the safer. Another, big advantage is that
acrylic is only half as heavy as glass which makes working with acrylic much easier
[5].
Fabrication of Tree Type Micro-Mixer … 1041

2.2 Methodology

1. Preparation of 2D-CAD Drawing

First of all designed the Micro-mixer in CAD software and prepared the 2D CAD
drawings in AutoCAD software 2018 version. The 2D drawing with different
configuration like with circular obstacle and without circular obstacle is prepared
later. Used Circular obstacles to get more mixing.
2. Input to CO2 LASER Machining
The CO2 laser machining takes input as a 2D CAD drawings and it analyzes the
path of channel which is given in CAD drawing. The drawings of the Tree type
of micro-mixer with different configurations like without circular baffles and
with circular baffles as per the Figs. 1 and 2 are provided.to the CO2 LASER
machining in.dxf format.
3. Material Feeding
After giving input to CO2 LASER machining the acrylic material is fed to the
machine.
4. Dry run
To check the weather the process is going correctly or wrong actual dry run
processes are carried out. The CO2 LASER Cutting & Engraving Machine is
presented in Fig. 3.
5. Machining Process
After the dry run process, the actual process has started to get the desired mixing
by varying speed and number of passes. The speed and number of passes are kept
constant to achieve the 0.5 mm depth approximately. The material is removed
by vaporization and ablation.

Fig. 1 Micro-mixer
(without circular baffles)
1042 S. R. Gavali et al.

Fig. 2 Micro-mixer (with

circular baffles)

Fig. 3 CO2 LASER cutting

and engraving machine

6. Fabrication of Micro-mixer (Mold)

After all the above steps actual micro-mixer mould is prepared which is
fabricated on CO2 LASER machine having width dimension 0.5 mm.
Fabrication of Tree Type Micro-Mixer … 1043

Fig. 4 Micro-mixer without

circular baffles

3 Results and Discussion

3.1 Fabrication of Tree Type Micro-Mixer

The micro-mixer without baffles and the micro-mixer with circular baffles have been
fabricated on acrylic material using CO2 laser machining and depicted in Figs. 4 and
5, respectively.

3.2 Parametric Analysis

While fabrication of Micro-mixers, the parameters presented in Table 1 are consid-

ered The Speed of LASER machine & its LASER power are plays important role
while fabrication in order get the micro-mixer with required depth. The effect of
power on the engraved depth is analyzed and it is observed that the higher depth is
noted with increased power.
1044 S. R. Gavali et al.

Fig. 5 Micro-mixer with

circular baffles

Table 1 Performance parameters used for micro-mixer fabrication

Width Speed No. of pass Power Average depth achieved
0.3 100 1 60 0.551
0.4 100 1 60
0.5 100 1 60
0.3 100 1 40 0.517
0.4 100 1 40
0.5 100 1 40

4 Conclusion

Micro-mixer is one of the essential components in integrated microfluidic systems

for chemical, biological and medical applications. In different features of fluid at the
microscale one of the most relevant to mixing applications laminar flow where mixing
can be dominantly accomplished by molecular diffusion. The fabrications of Tree
type micro-mixer with and without circular baffles have been carried out using CO2
laser cut machining. The effect of laser power and speed on the average engraving
depth of micro-mixers of different widths is analyzed. The depths recorded are as
Fabrication of Tree Type Micro-Mixer … 1045

0.551 mm and 0.517 mm. The influence of the power is observed to be prominent
and the depth is noted to be increasing with increase in laser power. This study shows
the suitability of CO2 laser cut machining for fabrication of micro-mixers or molds
for micro-mixers used in soft lithography process.

References

1. Klank H, Kutter JP, Geschke O (2000) CO2-Laser micromachining and back-end processing
for rapid Production of PMMA-based microfluidic systems. La Chi 2:242–246
2. Das SS, Tilekar SD, Wangikar SS, Patowari PK (2017) Numerical and experimental study of
passive fluids mixing in micro-channels of different configurations. Micros Technol 23:5977–
5988
3. Seck HW, Patrick B, Michael W, Christopher W (2013) Investigation of mixing in a cross-
shaped micromixer with static mixing elements for reaction kinetics studies. Sens Actua B
95:414–424
4. Gianni O, Mina R, Elisabetta B, Chiara G, Roberto M (2013) Water-ethanol mixing in T-shaped
microdevices. Chem Eng Sci 95:174–183
5. Tsung-Han T, Dar-Sun L, Long-Sheng K, Ping-Hei C (2009) Rapid mixing between ferro-
nanofluid and water in a semi-active Y-type micromixer. Sens Actua A 153:267–273
6. Wangikar SS, Patowari PK, Misra RD (2018) Numerical and experimental investigations on
the performance of a serpentine microchannel with semicircular obstacles. Micros Technol
24:3307–3320
7. Gidde RR, Pawar PM, Ronge BP, Misal ND, Kapurkar RB, Parkhe AK (2018) Evaluation of the
mixing performance in a planar passive micromixer with circular and square mixing chambers.
Micros Technol 24:2599–2610
8. Zonghuan L, Jay McMahon HM, Barnard D, Shaikh TR, Mannella CA, Wagenknecht T, Toh-
Ming L (2010) Passive microfluidic device for submilli second mixing. Sens Actua B 144:301–
309
9. Gidde RR, Shinde AB, Pawar PM, Ronge BP (2018) Design optimization of a rectangular
wave micromixer (RWM) using Taguchi based grey relational analysis (GRA). Micros Technol
24:3651–3666
10. Gidde RR, Pawar PM, Ronge BP, Shinde AB, Misal ND, Wangikar SS (2019) Flow field analysis
of a passive wavy micromixer with CSAR and ESAR elements. Micros Technol 25:1017–1030
11. Gerlach A, Knebel G, Guber AE, Heckele M, Herrmann D, Muslija AS (2002) Microfabrication
of single-use plastic microfluidic devices for high-throughput screening and DNA analysis.
Micros Technol 7:265–268
12. Arshad A, Kwang-Yong K (2012) Passive split and recombination micromixer with convergent-
divergent walls. Chem Eng J 203:182–192
13. Rotting O, Ropke W, Becker H, Gartner C (2010) Polymer microfabrication technologies.
Micros Technol 8:32–36
14. Wangikar SS, Patowari PK, Misra RD (2017) Effect of process parameters and optimization
for photochemical machining of brass and german silver. Mater Manuf Process 32:1747–1755
15. Wangikar SS, Patowari PK, Misra RD (2016) Parametric optimization for photochemical
machining of copper using grey relational method. In: 1st Techno-societal, international
conference on advanced technologies for societal applications. Springer, pp 933–943
16. Wangikar SS, Patowari PK, Misra RD (2018) Parametric optimization for photochemical
machining of copper using overall evaluation criteria. Mater Tod Proc 5:4736–4742
17. Wangikar SS, Patowari PK, Misra RD, Misal ND (2019) Photochemical machining: a less
explored non-conventional machining process. In: Non-conventional machining in modern
manufacturing systems. IGI Global, pp 188–201
1046 S. R. Gavali et al.

18. Kulkarni HD, Rasal AB, Bidkar OH, Mali VH, Atkale SA, Wangikar SS, Shinde AB (2019)
Fabrication of micro-textures on conical shape hydrodynamic journal bearing. Int J Tre Eng
Technol 36:37–41
19. Raut MA, Kale SS, Pangavkar PV, Shinde SJ, Wangikar SS, Jadhav SV, Kashid DT (2019)
Fabrication of micro channel heat sink by using photo chemical machining. Int J New Technol
Res 5:72–75
20. Patil PK, Kulkarni AM, Bansode AA, Patil MK, Mulani AA, Wangikar SS (2020) Fabrication
of logos on copper material employing photochemical machining. Nov MIR Res J 5(6):70–73
21. Wangikar SS, Patowari PK, Misra RD, Gidde RR, Bhosale SB, Parkhe AK (2020) Optimization
of photochemical machining process for fabrication of microchannels with obstacles. Mater
Manuf Process. https://doi.org/10.1080/10426914.2020.1843674
22. Chavan NV, Bhagwat RM, Gaikwad SS, Shete SS, Kashid DT, Wangikar SS (2019) Fabrication
& characterization of microfeatures on PMMA using CO2 laser machining. Int J Tre Eng
Technol 36:29–32