Semester Project Progress Report

on
OBJECT DETECTION FOR BLIND
PERSON
Submitted in partial fulfillment for award of
BACHELOR OF TECHNOLOGY
Degree
In
COMPUTER SCIENCE & ENGINEERING

2022-2023
Under the Guidance of: Submitted By:
Mr.Sanjay Srivastava Sakshi khandelwal(1900330100189)
Assistant Professor Yash Varshney (1900330100254)
Pradumn Singh ( 1900330100161)
Sadik Ali (2000330109009)

DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING

RAJ KUMAR GOEL INSTITUTE OF TECHNOLOGY
5th K.M. STONE DELHI-MEERUT ROAD, GHAZIABAD

Affiliated to Dr. A.P.J. Abdul Kalam Technical University, Lucknow

November 2022
 Raj Kumar Goel Institute of Technology, Ghaziabad
ISO 9001:2015 Certified
5th KM, STONE, DELHI-MEERUT ROAD, GHAZIABAD(U.P.)-201003
Department of Computer Science & Engineering

Project Progress Report

1.Course : Bachelor of Technology
2.Semester : VII
3.Branch : Computer Science & Engineering
4.Project Title : Object detection for blind person

5. Details of Students:
S.No. Univ. Roll No. Name Role Signature

1 1900330100189 Sakshi khandelwal Designer,Report

2 1900330100254 Yash Vashney Tester,Report

3 1900330100161 Pradumn Singh Tester

4 2000330109009 Sadik Ali Developer

6. Project Supervisor:
Mr.Sanjay Srivastava

Remarks from Project Supervisor:

………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………

v
 SYNOPSIS

Object detection is a well-known computer technology connected with computer vision and image
processing.With the advent of deep learning techniques, the accuracy for object detection has increased
drastically.It focuses on detecting objects or its instances of a certain class (such as humans, flowers,
animals) in digital images and videos.

Many problems in computer vision were saturating on their accuracy before a decade. However, with
the rise of deep learning techniques, the accuracy of these problems drastically improved. One of the
major problem was that of image classi cation, which is de ned as predicting the class of the image. A
slightly complicated problem is that of image localiza-tion, where the image contains a single object and
the system should predict the class of the location of the object in the image. The more complicated
problem (this project), of object detection involves both classi cation and localization.

An accurate and efficient object detection system has been developed which achieves comparable
metrics with the existing state-of-the-art system. This project uses recent techniques in the eld of
computer vision and deep learning. Custom dataset was created using labelling and the evaluation was
consistent. This can be used in real-time applications which require object detection for preprocessing
in their pipeline. An important scope would be to train the system on a video sequence for usage in
tracking applications. Addition of a temporally consistent network would enable smooth detection and
more optimal than per-frame detection.

The goal of the this project is to model an object detector to detect objects for visually impaired people
and other commercial purposes by recognizing the objects at a particular distance. Available old
techniques for object detection needed large training data it takes more time and its quite complicated
and it’s a difficult task. With the help of object detection system and navigation system the blind person
can easily detect the object through cameras or through some sensors.

Object detection algorithms typically leverage machine learning to produce meaningful results. When
humans look at images or video, we can recognize and locate objects of interest within a matter of
moments. Getting ito iuse imodern iobject idetection imethods iin iapplications iand isystems, ias iwell
ias ibuilding inew iapplications ibased ion ithese imethods iis inot ia istraight iforward itask.

vi
This iproject itook iuse iof iseveral isoftware ilibraries, ipackages iand iprograms ito iutilize imachine
ilearning. iPython iwas ithe ichoice iof iprogramming ilanguage, iand iTensorFlow iwas iused ifor ithe
ideep ilearning icomputations, iwhich iin iturn ihas ia ilist iof idependencies. iTensorFlow ioffers ia
iversion ifor iCPU iusage iand ianother ifor iGPU, ithis iproject iused ithe iGPU iversion. i

Computer vision is still a developing discipline, it has not been matured to that level where it can be
applied directly to real life problems. After few years computer vision and particularly the object
detection would not be any more futuristic and will be ubiquitous. For now, we can consider object
detection as a sub-branch of machine learning.

vii
 TABLE OF CONTENTS

SNO. TITLE PAGE NO.

SYNOPSIS ii
LIST OF TABLES vii
LIST OF FIGURES viii

1. INTRODUCTION 1
1.1 Problem Statement 1
1.2 Motivation 2
1.3 Objective 2
1.4 Scope 3
1.5 Tools
1.5.1 Visual Studio
1.5.2 MySQL 4
1.6 Background and related work
2. HARDWARE AND SOFTWARE REQUIREMENTS 8
2.1 Hardware Requirements 9
2.2 Software Requirements 9
3. SDLC METHODOLOGIES 10
3.1 Different SDLC Methodologies Used 11
3.2 How It Affects Our Model 13
4. SOFTWARE REQUIREMENT SPECIFICATION AND 14
ANALYSIS
4.1 Introduction 15
4.1.1 Functional Requirements 17
4.1.2 Non- Functional Requirements 18
19
4.1.3 Interface Requirements
4.1.4 Performance Requirements
4.2 Syste
4.3 Overall Description
4.3.1 Product Perspective
5. RISK ASSESSMENT 22
5.1 Outliers 22
5.2 Feature Selection 23
5.3 Risk Assessment Service 25

vii
i
6. Project 26
6.1 ER Diagram 26
7. PROJECT MODULES DESIGN/SOFTWARE FEATURES 27
7.1 Software Features 27
8. TESTING AND EVALUATION 30
9. PROJECT SNAPSHOTS 36
10. LIMITATION 38
11. FUTURE SCOPE 39
12. CONCLUSION 40
13. REFERENCES 42
14 RESEARCH PAPER

ix
 LIST OF TABLES
Chapter Table No. Title Page No.

1 Table 1.6 Comparison of various 4

methodology

i
 LIST OF FIGURES

Chapter No. Title Page No.

3 3.1 Waterfall Model 19
3 3.2 Spiral Model 20
3 3.3 RAD Model 21
3 3.4 V Model 21
3 3.5 Iterative Model 22
3 3.5.1 How It Affects our model 22
 CHAPTER 1
INTRODUCTION

One of the most difficult problems in image processing is detecting particular objects in an image.
For a human observer it is very easy to identify any object, however it is far more difficult for a
machine. Numerous methods exist for detecting objects of known type in a particular environment
or image.

Videos are the sequences of frames that run fast enough to give an effect of continuity as human
eye perceives the frame sequences moving with a particular speed as a video. However, in many
cases, the visual characteristics of the objects are unknown, or it is necessary to detect objects that
are very different from each other.

As far as object detection is concerned techniques of image processing are applied to the frames
in order to identify any change so as to state the motion of object detected through the change
observed in the two consecutive frames after attaining the results of the image processing
techniques used for identification purpose.

The main goal of object detection is to scan digital images or real-life scenarios to locate instances
of every object, separate them, and analyze their necessary features for real-time predictions.
Object detecting and tracking has a wide variety of applications in computer vision such as video
compression, video surveillance, vision-based control, human-computer interfaces, medical
imaging, augmented reality, military applications, traffic monitoring and robotics.

A surveillance system can be implemented in three steps. The first step consists of detecting the
objects in motion .Then tracking them and finally High-level interpretation of the ongoing events.

With the help of object detection system and navigation system the blind person can easily detect
the object through cameras or through some sensors. For example, it is very difficult for to find a
particular room in an unfamiliar environment. The basic procedure is to combine data-driven
knowledge in the likelihood terms with clinical knowledge in the prior terms to classify an image
into a pre-determined number of classes.
1.1 Problem Statement

The basic concept behind object detection in videos engrosses the verification of the presence of
an object in image sequences and possibly locating it in particular for recognition. This is done by
solving the temporal correspondence problem, the problem of matching the target region in
successive frames of a sequence of images taken at closely spaced time intervals. The general
problem of motion understanding and tracking of the moving objects is one of the most
interestingly used areas of computer vision

1.2 Motivation
The motive of object detection is to recognize and locate all known objects in a scene. Preferably
in 3D space, recovering pose of objects in 3D is very important for robotic control systems.

1.3 Objective
Moving object detection is important in many real-time image processing applications such as
autonomous robotics, traffic control, and driver assistance and surveillance systems. Usually high
resolution gray-scale images must be processed; since each image pixel may belong to a moving
object, pixel-wise processing is required. Now days, video surveillance is an important and
challenging field in computer vision for both indoor and outdoor environments .The project
Moving object tracking from video sequences is an attempt to study some algorithms, which are
robust for the tracking of mobile but non rigid objects from the image sequences precisely called
video.

The goal of the this project is to model an object detector to detect objects for visually impaired
people and other commercial purposes by recognizing the objects at a particular distance.
Available old techniques for object detection needed large training data it takes more time and its
quite complicated and it’s a difficult task. With the help of object detection system and navigation
system the blind person can easily detect the object through cameras or through some sensors.
It also includes pre and post processing sequences as well which includes noise removal, image
enhancement issues, organization and classification etc. but these issues are not of primary
importance as far as this project is concerned so are not addressed in detail in this project.
1.4 Scope

Object detection is a computer vision technique that allows us to identify and locate objects in an
image or video. With this kind of identification and localization, object detection can be used to
count objects in a scene and determine and track their precise locations, all while accurately
labeling them.The main scope of object detection is to scan digital images or real-life scenarios to
locate instances of every object, separate them, and analyze their necessary features for real-time
predictions. Object detection is a part of the overall data architecture of a company.

1.5 Tools
1.5.1 Visual Studio
Visual Studio is anintegrated development environment(IDE) fromMicrosoft. It is used to
developcomputer programsincludingwebsites,web apps,web servicesandmobile apps. Visual
Studio uses Microsoft software development platforms such asWindows API,Windows
Forms,Windows Presentation
Foundation,Windows StoreandMicrosoft Silverlight. It can produce bothnative codeandmanaged
code. Visual Studio supports 36 differentprogramming languagesand allows the code editor and
debugger to support nearly any programming language, provided a language-specific service
exists.

1.5.2 MySQL
MySQL is an open source relational database management system. For WordPress sites, that
means it helps you store all your blog posts, users, plugin information, etc. It stores that
information in separate “tables” and connects it with “keys”. MYSQL used for the application
database. Insertions, Deletions and changes of the data in the system can do straightforward via
the designed GUI without interacting with the table.

1.6 BACKGROUND AND RELATED WORKS

TABLE 2.1. Comparison of various methodology suggested by authors

S.No. Paper Name Author Name Year Methodology
1 Object Detection Dr. M Y Babu, 2017 In this Object Detection system with
system with voice Akash Jatavath, G voice alertwe uses image recognition
alert Yashwanth Kumar voice generation modules for the
Reddy, Pittala Arun development of the project
Kumar
2 Real time object Shifa Shaikh, 2020 Real time object detection for blind
detection for blind Vrushali Karale, person uses unified framework for
person Gaurav Tawde both single and cross camera
tracking with affinity constraints.

3 Object Detection Hanen Jabnoun, 2018 Object Detection and identification

and Identification Faouzi Benzarti, for blind person visual substitution
for Blind People in Hamid Amiri. devices accept input from the user’s
Video Scene surroundings, decipher it to extract
information about entities.

4 Object Detection Mayuresh Banne, 2020 Oject detection and translation in

and Translation for Rahul these we are using a system from
Blind People Vhatkar, Ruchita which blind person can place an
Using Deep Tatkare object into that system.
Learning

5 Oject detection for Rajeshvaree 2021 Object detection for blind person
blind person with Ravindra Karmarkar with voice guidance account but the
voice guidance whole image is divided into grids.

6 Blind Person Aniket Birambole , 2022 Blind person assistant using SSD
Assistant: Object Pooja algorithm and tensorflow APIs. Very
Detection. Bhagat , Bhavesh first step of image processing is
Mhatre , Prof. Aarti readthe training as well as testing
Abhyankar images

M Y Babu et. al [1] Suggest a prototype system of camera-based assistive object reading. This
system consists of three functional components, screen capture, image processing, audio output.
The screen capture component collects the objects of interest in the form of image or video.The
Data processing system is used for our proposed algorithms, including Region-of-interest to
collects the objects from the cluttered backgrounds or from other surroundings, image localization
to attain the object regions to containing grey scale and object recognition to transform the image-
based readable codes. The audio output component is to inform the visually impaired persons of
recognised text codes.The 3.5 mm audio jack is interfaced with the board which receives as a
speech.

S Shaikh et al. [2] Suggest two cameras placed on blind person's glasses, GPS free service, and
ultrasonic sensors are employed to provide information about the surrounding environment.
Object detection is used to find objects in the real world such as faces, bicycles, chairs, doors, or
tables that are common in the scenes of a blind. Here, GPS service is used to create groups of
objects based on their locations, and the sensor detects an obstacle at a medium to long distance.
The descriptor of the Speeded-Up Robust Features (SURF) method is optimized to perform the
recognition. The use of two cameras on glasses can be sophisticated.

H. Jabnoun et al [3] Suggest that visual substitution devices accept input from the user’s
surroundings, decipher it to extract information about entities in the user’s environment, and then
transmit that information to the subject via auditory or tactile means or some combination of these
two They are based on rules for converting images into data sensory substitution tactile or auditory
stimuli. These systems are efficient for mobility and localization of objects which is sometimes
with a lower precision.

M. Rahul et al [4] Suggest a smart vision whose objective is to move anywhere in the environment
through a user-friendly interface system. his project mainly focusses on computer vision module.
In these authors made a system which will able to find the obstacle which are near to his head
specially while entering from door.in short it is made to protected his head from getting injury.
This product is design to navigate blind person in any environment and it guides the user about
that object and provide information about that obstacle using buzzer and vibrater as a two-output
mode of the user.

R.R. Karmarkar et al [5] Suggest about mask region-based convolutional network (Mask-RCNN)
is used for utilization for multi-class object detection for sensing the images. Transfer learning,
fine tuning and data augmentation are used to overcome object scale variability, the density of
object. Also, adaptive Mask-RCNN was compared to deep object detection methods. MaskRCNN
is an extended version to faster RCNN that allow an accurate pixel-based segmentation it consist
of two stages feature pyramid network (FPN) and region proposal network (RPN)
A. Birambole et al [6] Suggest attempts about object detection from the image and represented it
by their name and speech. And it also helps the blind people in location and encoded the audio
into 2 channel audios with the help of 3D binaural sound. In these a video is capture with
portable camera device from client side and it is streamed to a server for real time image
recognition with object detection. which mean it identify and follow the same object in
sequence of video frames sometimes video may have some noise. to remove that noise from
frames noise reduction technique is used that improve the image quality and extraction of
object frame is used to detected the object based on color of the moving frame.
 CHAPTER 2
HARDWARE AND SOFTWARE REQUIREMENTS
All computer softwares requires the presence of specific hardware parts or additional software
resources. These prerequisites, also referred to as (computer) system requirements, are frequently
employed as a general norm rather than an unbreakable law. The majority of software specifies a
minimum and recommended set of system requirements. System requirements tend to rise over
time due to the rising demand for more processing power and resources in newer versions of
software.

3.1 HARDWARE REQUIREMENTS

The physical computer resources, usually referred to as hardware, are the most typical set of
specifications defined by any operating system or software programme. A hardware compatibility
list (HCL) is frequently included with a list of the necessary hardware, especially when operating
systems are involved. For a specific operating system or application, an HCL describes hardware
components that have been evaluated, are compatible, and occasionally are not.

• Pentium i3 Processor
• Hard Disk : 500 GB
• Ultrasonic sensor
• GSM

3.2 SOFTWARE REQUIREMENTS

Operating System: Windows 98 and Higher Version
Open CV
• Jaws..
• Anaconda
• Web Server: Apache 2.4.18
• Database Server: MySQL 5.7.20
• Cloud Service – Amazon Web Service(AWS) or Oracle or any other similar providers
• Tech Stack– MYSQL, ExpressJs, Angular, NodeJs
• Operating System – Linux
• Panel Software (on cloud) – aaPanel
• Cloud remote access – Putty
• Web Browser – Chrome or Firefox
 CHAPTER 3

SDLC METHODOLOGIES
A software development life cycle model (also termed process model) is a pictorial and
diagrammatic representation of the software life cycle. A life cycle model represents all the
methods required to make a software product transit through its life cycle stages. It also
captures the structure in which these methods are to be undertaken. In other words, a life cycle
model maps the various activities performed on a software product from its inception to
retirement. Different life cycle models may plan the necessary development activities to phases
in different ways. Thus, no element which life cycle model is followed, the essential activities
are contained in all life cycle models though the action may be carried out indistinct orders in
different life cycle models. During any life cycle stage, more than one activity may also be
carried out.

3.1 WATERFALL MODEL

The waterfall model is the oldest of all SDLC methodologies. It’s linear and straightforward
and requires development teams to finish one phase of the project completely before moving
on to the next. Each stage has a separate project plan and takes information from the previous
stage to avoid similar issues (if encountered). However, it is vulnerable to early delays and lead
to big problems arising for development teams later down the road.

Figure 3.1:Waterfall Model3

3.2 SPIRAL MODEL

The spiral model is a risk-driven process model. This SDLC model helps the group to adopt
elements of one or more process models like a waterfall, incremental, waterfall, etc. The spiral
technique is a combination of rapid prototyping and concurrency in design and development
activities. Each cycle in the spiral begins with the identification of objectives forthat cycle, the
different alternatives that are possible for achieving the goals, and the constraints that exist. This
is the first quadrant of the cycle (upper-left quadrant). The next step in the cycle is to evaluate
these different alternatives based on the objectives and constraints. The focus of evaluation in this
step is based on the risk perception for the project .The next step is to develop strategies that solve
uncertainties and risks. This step may involve activities such as benchmarking, simulation, and
prototyping.

Figure 3.2: Spiral Model1

3.3 RAD MODEL

RAD or Rapid Application Development process is an adoption of the waterfall model; it

targets developing software in a short period. The RAD model is based on the concept that
abetter system can be developed in lesser time by using focus groups to gather system
requirements.
• Business Modelling

• Data Modeling

• Process Modelling

• Application Generation

• Testing and Turnover

Figure 3.3. RAD Model 8

3.4 V- Model

In this type of SDLC model testing and the development, the step is planned in parallel. So, there
are verification phases on the side and the validation phase on the other side. V-Model joins by
Coding phase.
 Figure 3.4. V-Model4

3.5 ITERATIVE MODEL

For Our model, Iterative process starts with a simple implementation of a subset of the software
requirements and iteratively enhances the evolving versions until the full system is implemented.
At each iteration, design modifications are made and new functional capabilities are added. The
basic idea behind this method is to develop a system through repeated cycles (iterative) and in
smaller portions at a time (incremental).The following illustration is a representation of the
Iterative and Incremental model –

Figure 3.5. Iterative Mode6

3.5.1. HOW IT AFFECTS OUR MODEL

Deep learning-based detectors usually produce a redundant set of object bounding boxes
including many duplicate detections of the same object. These boxes are then filtered using
non-maximum suppression (NMS) in order to select exactly one bounding box per object of
interest. This greedy scheme is simple and provides sufficient accuracy for isolated objects but
often fails in crowded environments, since one needs to both preserve boxes for different
objects and suppress duplicate detections. In this work we develop an alternative iterative
scheme, where a new subset of objects is detected at each iteration. Detected boxes from the
previous iterations are passed to the network at the following iterations to ensure that the same
object would not be detected twice. This iterative scheme can be applied to both one-stage and
two-stage object detectors with just minor modifications of the training and inference
procedures.
 CHAPTER 4
SOFTWARE REQUIREMENT SPECIFICATION AND
ANALYSIS

4.1 Introduction
The objective of this software requirement specification and analysis is to provide a detailed
description of the requirements and specifications of an object detection system for blind
persons. The system is designed to assist the blind by detecting objects in the user's
surroundings using computer vision and image processing techniques. The system will also

provide the user with audio feedback on the detected objects.The requirements were

analyzed and classified into three categories:

 Functional requirements

 Non-functional requirements

 Interface requirements

 Performance requirements

4.1.1 Functional Requirements:

Functional requirements are the requirements which define what the system should do. In this
case, the functional requirements are as follows: -
 Object Detection: The system should be able to detect various objects in the user's
surroundings such as chairs, tables, doors, and walls.
 Audio Feedback: The system should provide audio feedback to the user about the
detected objects, such as the type of object and its distance from the user.
 User Interface: The system should have a user-friendly interface that can be easily used
by blind users.
 Object Recognition: The system should be able to recognize and differentiate between
different types of objects.
 Real-time Processing: The system should be able to process the video feed in real-time
to provide quick feedback to the user.
 4.1.2 Non - Functional Requirements:

Non-functional requirements are the requirements which define how the system should

work. In this case, the non-functional requirements are as follows:

Performance: The system should be able to detect objects accurately and quickly.
 Accessibility: The system should be easily accessible and usable by the blind.
 Reliability: The system should be reliable and provide consistent results.
 Security: The system should be secure and protect the user's privacy.
 Scalability: The system should be scalable to handle a large number of users.

4.1.3 Interface Requirements:-

 Compatibility with various devices: The software should be compatible with a range of devices
such as smartphones, tablets, and wearable devices. This would allow users to choose the device
that best suits their needs and preferences.
 Accessibility features: The software should be designed with accessibility in mind, including the
use of large fonts, high-contrast colors, and voice recognition. This would ensure that the
software is accessible to as many users as possible, including those with visual or motor
impairments.

4.1.4 Performance Requirements :-

 Fast and accurate detection: The software should have fast and accurate object detection
capabilities with a response time of no more than a few seconds. This would enable users to
navigate their environment safely and efficiently.
 Low-light functionality: The software should be able to function in low-light conditions
and compensate for image distortion or blur. This would ensure that users can use the software
in a variety of lighting conditions.
 Minimal resource consumption: The software should consume minimal resources and
not excessively drain the battery of the device. This would ensure that users can use the software
for extended periods without having to recharge their device frequently.

4.2 System Design:

 The system will consist of the following components:
 Camera: The camera will capture the video feed of the user's surroundings.
 Object Detection Algorithm: The algorithm will process the video feed to detect objects.
 Object Recognition Algorithm: The algorithm will recognize and differentiate between different
types of objects.
 Audio Feedback System: The audio feedback system will provide the user with audio feedback
on the detected objects.
 User Interface: The user interface will allow the user to interact with the system and receive
feedback.

4.3. Overall Description

Object detection for blind individuals refers to the use of technology to help visually impaired
people detect and recognize objects in their environment. This technology can use a variety of
sensors and interfaces to provide information about objects, such as their location, size, shape, and
color.Object detection systems can be based on different technologies, such as computer vision,
machine learning, and artificial intelligence. These systems can use cameras, sensors, and other
devices to capture images or other data from the environment and then analyze it to identify and
describe objects.

The information provided by an object detection system can be conveyed to the user through
various interfaces, such as audio and visual feedback. Audio interfaces might use text-to-speech
engines to describe objects or a series of beeps or tones to indicate an object's location and distance.
Visual interfaces might use a smartphone camera to capture an image of the object and then use
image recognition software to describe its properties.Object detection systems can be helpful for
blind individuals in a variety of settings, such as navigating through unfamiliar environments,
identifying items on store shelves, or recognizing people's faces. These systems can help improve
the independence and quality of life of visually impaired people by providing them with more
information about their surroundings.

4.3.1 Product Perspective

From a product perspective, object detection technology for blind individuals is a product
designed to meet the unique needs of visually impaired people. This technology uses a variety of
sensors and interfaces to help blind individuals detect and recognize objects in their environment,
providing them with more information about their surroundings and helping them navigate through
unfamiliar environments.

There are several system interfaces for object detection that can be helpful for blind individuals.
Here are some options:

 Audio-based interfaces: Audio-based interfaces use sound to convey information about

the objects around a person. For example, an object detection system could use a text-to-
speech engine to describe the object's properties, such as its color, size, and shape.
Alternatively, the system could use a series of beeps or tones to indicate the object's
location and distance.

 Visual interfaces: Visual interfaces use graphics or other visual cues to convey
information about the objects around a person. For example, a system could use a
smartphone camera to capture an image of the object and then use image recognition
software to describe the object's properties.

 Wearable interfaces: Wearable interfaces are devices that are worn on the body, such as
smart glasses or a smartwatch. These devices can use a combination of audio, haptic, and
visual feedback to provide information about the objects around a person.

Overall, the choice of interface will depend on the individual's needs and preferences. Some people
may prefer an audio-based interface, while others may prefer a haptic or visual interface. It's
important to consider factors such as ease of use, portability, and cost when choosing an object
detection system for a blind person.
 CHAPTER-5
RISK ASSESSMENT
Risk analysis in software testing is an approach to software testing where software risk is analysed
and measured. Traditional software testing normally looks at relatively straight-forward function
testing (e.g. 2 + 2 = 4). A software risk analysis looks at code violations that present a threat to the
stability, security, or performance of the code.
Software risk is measured during testing by using code analyzers that can assess the code for both
risks within the code itself and between units that must interact inside the application. The greatest
software risk presents itself in these interactions. Complex applications using multiple
frameworks and languages can present flaws that are extremely difficult to find and tend to cause
the largest software disruptions.

The main risks associated with this project are:-

 Accuracy of object detection: The accuracy of the object detection system is crucial. If
the system fails to detect an object or provides incorrect information, it could result in
potential hazards for the blind person.
 Speed of object detection: The system needs to detect objects quickly enough for the user
to avoid them in time. If the system is too slow, it could fail to provide timely information,
which could lead to accidents.
 Usability: The system needs to be user-friendly and easy to use for the blind person. It
should not require too much cognitive effort or physical movement, which could be
difficult for the user.
 Environmental factors: The system needs to be able to detect objects in various
environmental conditions, such as low light, bright sunlight, or crowded spaces.
 False positives: The system should not produce too many false positives, as this could lead
to confusion and mistrust of the system.

5.1 Outliers
In the context of object detection for blind individuals, risk outliers refer to unexpected events or
situations that may occur when using the technology that could potentially cause harm or negative
consequences. Here are some possible risk outliers for object detection technology for blind
individuals:

 Environmental factors: Object detection technology may not work well in certain
environments, such as low light conditions or areas with many obstacles. In these
situations, users may rely too heavily on the technology and be at risk of injury or accident.
 Device malfunctions: Like any technology, object detection devices may malfunction or
fail, leading to incorrect information or no information being provided to the user. This
could potentially put the user at risk if they are relying on the technology to navigate or
identify objects.
 Cybersecurity threats: Object detection technology that connects to the internet or other
networks may be vulnerable to cyber attacks or hacking attempts, which could compromise
user data or put the user at risk in other ways.
 Legal and regulatory risks: The use of object detection technology may be subject to legal
and regulatory restrictions or requirements, which could pose risks to the user or the
developer if not properly addressed.

To address these risk outliers, developers of object detection technology for blind individuals
should take a proactive approach to risk management. This may include implementing strong
cybersecurity measures, conducting thorough testing and quality control, staying up-to-date with
legal and regulatory requirements, and engaging with diverse stakeholders to address social and
cultural concerns..

5.2 Feature selection

When selecting features for object detection technology for blind individuals, it is important to
consider the specific needs and preferences of the target audience. Here are some possible features
to consider for object detection technology for blind individuals:

 Multimodal sensory inputs: Object detection technology could use a combination of

sensors, such as cameras, infrared sensors, and ultrasonic sensors, to provide users with
information about objects in their environment. This would allow users to access
information about objects through multiple sensory modalities, enhancing their ability to
understand their surroundings.
  Real-time feedback: Object detection technology should provide real-time feedback to
users, allowing them to quickly and accurately identify objects and navigate their
environment.
 Voice commands: Object detection technology could allow users to control the device
using voice commands, making it easier for users to interact with the technology without
needing to use their hands.
 Customizable settings: Object detection technology should allow users to customize
settings, such as the types of objects to detect or the level of detail provided about each
object, to meet their specific needs and preferences.
 Navigation support: Object detection technology could provide navigation support, such
as directions or alerts, to help users navigate through unfamiliar environments.

Overall, feature selection for object detection technology for blind individuals should be guided
by the principles of usability, accessibility, and user-centered design..

5.3 RISK ASSESSMENT SERVICE

A risk assessment service for object detection for blind persons would be a valuable tool to help
improve their mobility and safety. The service could be designed to use computer vision algorithms
to analyze video or images from a camera mounted on a user's cane or worn on their person. The
algorithms would identify and classify obstacles, such as curbs, stairs, or other potential hazards,
and then provide the user with audio or haptic feedback to warn them of the obstacle's presence
and location.

To develop such a service, a team of engineers, computer scientists, and accessibility experts
would need to work together. They would need to create and train a machine learning model to
recognize common obstacles that a blind person might encounter, as well as develop the software
and hardware necessary to provide the user with feedback.The service would also need to be tested
extensively to ensure it is accurate, reliable, and easy to use. It would also be important to consider
the privacy and security implications of such a service, to protect the user's data and prevent
unauthorized access to the device.

In summary, a risk assessment service for object detection for blind persons has the potential to
significantly improve the mobility and safety of blind individuals. However, developing such a
service would require a multidisciplinary team, careful testing, and consideration of privacy and
security concern

CHAPTER 6
APPLICATION ARCHITECTURE
6.1 E-R Diagram:
ER Diagram stands for Entity Relationship Diagram, also known as ERD is a diagram that displays
the relationship of entity sets stored in a database. In other words, ER diagrams help to explain the
logical structure of databases. ER diagrams are created based on three basic concepts: entities,
attributes and relationships.

Fig 6.1 ER Diagram

 CHAPTER 7
SOFTWARE FEATURES

7.1 SOFTWARE FEATURES

The software features for an object detection system designed for blind individuals could include:
● Real-time object detection: The system should be able to detect objects in real-time and provide
feedback to the user about the objects present in their surroundings.

● Accurate object recognition: The system should be able to recognize different types of objects
accurately, including people, animals, vehicles, and other obstacles.

● Audio feedback: The system should provide audio feedback to the user to inform them about
the objects present in their surroundings. The audio feedback could include information about
the size, shape, and location of the object.

● Voice commands: The system should be able to take voice commands from the user, allowing
them to ask questions about the objects in their surroundings.

● Customizable settings: The system should have customizable settings, allowing the user to
adjust the sensitivity and range of the object detection system to suit their preferences and
needs.

● Multiple input options: The system should have multiple input options, allowing the user to
interact with the system through touch, voice commands, or other inputs.

● Data privacy: The system should prioritize data privacy and ensure that the user's personal
information is protected.

● Continuous improvement: The system should have a continuous improvement process,

allowing for updates and improvements to be made to the software over time.
● Integration with other devices: The system should be able to integrate with other devices, such
as smartphones or wearable technology, to provide additional functionality to the user.

● User-friendly interface: The system should have a user-friendly interface that is easy to use
and understand, even for those who are not tech-savvy.
 CHAPTER 8
TESTING AND EVALUATION
Object detection for blind people is an important and challenging area of research, and testing and
evaluation play a crucial role in ensuring that these systems are accurate and effective.

There are several approaches to evaluating object detection systems for blind people, including:

 Performance Metrics: A common approach to evaluating object detection systems is to use

performance metrics such as accuracy, precision, recall, and F1 score. These metrics can help to
quantify how well the system is performing in terms of detecting objects correctly.

 User Studies: Another approach to evaluating object detection systems for blind people is to conduct
user studies. These studies involve asking blind people to use the system and provide feedback on
its effectiveness, ease of use, and other important factors.

 Real-world Testing: Testing object detection systems in real-world environments is also important.
This can involve testing the system in different lighting conditions, with different types of objects,
and in various locations.

 Comparison with Existing Systems: Another important approach to evaluating object detection
systems for blind people is to compare them with existing systems. This can help to identify
strengths and weaknesses and highlight areas for improvement.

In addition to these approaches, it is also important to involve blind people in the development and
testing of these systems. By involving blind people in the process, developers can gain valuable
insights into how these systems can be improved to better meet the needs of their intended users.
 CHAPTER 9
PROJECT SNAPSHOTS

Figure. 8.1 File location of Assets

Figure. 8.2 Launching and Acessing Anaconda

Figure. 8.3 Launching Page

 Figure. 8.4 Detection of Objects -Chair

Figure. 8.5 Detection of Objects - Chair and Mobile Phone

Figure. 8.6 Detection of Objects - Chair and Mouse

Figure. 8.7 Detection of Human Being

Figure. 8.8 Detection of Water Bottle

Figure. 8.9 Detection of Cow

 Figure. 8.10 Detection of Bag

Figure. 8.11 Detection of Cat and person

 CHAPTER 10
LIMITATION

Object detection is a promising technology that has the potential to enhance the independence of blind
persons. However, there are several limitations of object detection technology that should be considered
when developing a project for blind persons. Here are some of the most important limitations:
 Accuracy: Object detection algorithms are not always accurate in identifying objects, particularly
when objects are partially obscured or located in complex environments. Inaccuracies can lead to
false positives or false negatives, which can create confusion or danger for blind persons.
 Cost: The cost of object detection technology can be high, making it difficult for many blind
persons to afford. Additionally, some object detection systems require specialized hardware, such
as cameras or sensors, which can further increase costs.
 Accessibility: Object detection technology may not be accessible to all blind persons, particularly
those who are not proficient with technology. It is important to ensure that any project utilizing
object detection technology is designed with accessibility in mind.
 Dependence: Blind persons may become overly dependent on object detection technology, which
can limit their ability to develop other skills and strategies for navigating the world. For example,
if they become reliant on object detection systems for navigation, they may not develop other
orientation and mobility skills.
 Privacy: Object detection technology can raise privacy concerns, particularly if cameras or sensors
are used to collect data about a person's movements or activities. It is important to ensure that any
project utilizing object detection technology is designed with privacy in mind.
 Cultural Differences: It is important to consider cultural differences in the use of object detection
technology. Some cultures may view the use of cameras or sensors as intrusive, which can limit
the effectiveness of object detection technology in those contexts.
 Training Data Bias: Object detection algorithms rely on large amounts of training data to identify
and classify objects accurately. However, if the training data is biased towards certain types of
objects or certain environments, the object detection system may perform poorly in other
environments or with other objects.
Overall, while object detection technology can be a valuable tool for enhancing the independence of blind
persons, it is important to consider these limitations when designing a project. By taking these limitations
into account, developers can create projects that are more effective, accessible, and respectful of the
privacy and cultural sensitivities of blind persons
 CHAPTER 11
FUTURE SCOPE
The future scope of object detection technology for blind persons is promising, as there is a lot of potential
for further development and improvement. Here are some of the potential future directions for object
detection technology for blind persons:
 Improved Accuracy: Researchers are constantly working on improving the accuracy of object
detection algorithms, particularly in challenging environments. This could help reduce false
positives and false negatives and make object detection technology more reliable.
 More Affordable Technology: As the technology advances and becomes more mainstream, it is
possible that the cost of object detection technology will decrease, making it more accessible to
blind persons.
 Improved Accessibility: Developers are working on improving the accessibility of object detection
technology for blind persons, including through the use of audio and haptic feedback.
 Integration with Other Technologies: Object detection technology could be integrated with other
technologies, such as artificial intelligence, to provide more advanced features and capabilities.
 Increased Privacy Protections: As concerns around privacy continue to grow, developers are
working on improving the privacy protections of object detection technology, such as by
minimizing the collection and storage of personal data.
 Adoption by Industry: Object detection technology is being adopted by industries such as retail
and transportation to improve accessibility for customers and passengers. As these industries
continue to embrace the technology, there may be further opportunities for the technology to be
utilized for blind persons.
 Overall, the future scope of object detection technology for blind persons is bright. With continued
research and development, it is possible that object detection technology could become an even
more effective and accessible tool for enhancing the independence of blind persons.
 Conclusion
In conclusion, object detection technology holds great promise for enhancing the independence of blind
persons. However, as with any technology, there are limitations that must be considered when developing
projects for blind persons. These limitations include accuracy, cost, accessibility, dependence, privacy,
cultural differences, and training data bias. Despite these limitations, the future scope of object detection
technology for blind persons is promising. Researchers are working to improve accuracy and accessibility,
integrate the technology with other systems, and increase privacy protections. As the technology continues
to develop and become more widely adopted, it has the potential to revolutionize the way that blind
persons navigate and interact with the world around them. Ultimately, it is important to continue to
develop and refine object detection technology in a way that is respectful of the unique needs and
perspectives of blind persons.
 REFERENCES

[1] Dr. M Y Babu, Akash Jatavath “Object Detection system with voice alert” ,International
research journal of engineering Vol 9 S(1), DOI: 10.17485/ijst/2016/v9iS(1)/108446
ref.December 2021.

[2] Shifa Shaikh, Vrushali Karale, Gaurav “Real time object detection for blind person” in Issue
09,September-2022Vision,,Vol. 9905, pp. 21-37, Sept. 2018

[3] Hanen Jabnoun, Faouzi Benzarti, Hamid Amiri “Object Detection and identification for blind
people” in vedio scene,international conference of computational intelligence,pp 27-34.

[4] Mayuresh Banne , Rahul Vhatkar, Ruchita Tatkare “Object Detection and Translation for

Blind People Using Deep Learning”, international Research Journal of Engineering and
Technology .(IRJET) e-ISSN: 2395-0056

[5] Rajeshwari Ravindra Karmarkar “Object detection system for blind person with voice
guidance” International Journal of Engineering Applied Sciences and Technology, 2021 Vol.
6, Issue 2, ISSN No. (2455-2143).

[6] Aniket Birambole, Pooja Bhagat, Bhavesh Mhatre, Prof. Aarti Abhyankar “Indian Journal of
Marine science” 38, (2009):

[7] Nada Alzahrani, Heyam-H Al Baity “ Object Recognition for the visually Impaired” ref fig
Information Technology Department, Collage of Computer and Information Sciences, King Saud
University, Riyadh P.O. Box 145111, Saudi Arabia.

[8] Manjula S. , Laxmi Krishnamurthy “ Study on Object detection” December 2016 International
journal of pharmacy and technology ISSN: 0975-766X.

[9] Tausif dewaan, Anirudh, Jitendra V. Tembhurne “ Object detection using YOLO ”
Department of Computer Science & Engineering, Indian Institute of Information Technology,
Nagpur, India. 2017.

[10] Udulie H.D.I., Lakshini Kuganandarmurthy “ Real time Object detection Using YOLO”
Approx Algorithms Metaheuristics, pp. 1-1432, 2007

[11] Daniel D, Rekha M S “ Facial Emotion and Object Detection for visually impaired and
Blind person” International Journal of engineering research and technology Volume 10, Issue 09(
September 2021) .
 OBJECT DETECTION FOR BLIND
PERSON USING ML

Sanjay Srivastav sir 1, Sakshi Khandelwal 2, Pradumn Singh3, Sadik Ali4, Yash Vashney5

Assistant Professor1,Department of CSE, Final Year Students2345,RKGIT Ghaziabad

201003,India

most difficult tasks in image processing is finding

certain objects in an image. A human observer can
easily identify any thing, but a machine has a far
more difficult time. There are numerous
Abstract: Different technologies have been approaches for recognising items of known type
integrated into autonomous vehicles, robots, and in a given environment or image.
industrial facilities as object recognition Videos are sequences of frames that run quickly
technology has recently advanced. The visually enough to give the illusion of continuity as the
impaired, who need these technologies the most, human eye perceives the frame sequences moving
do not, however, receive their benefits. This study at a specific speed as a video. However, in many
suggested a deep learning-based object detection cases, the visual characteristics of the objects are
system for visually impaired people. To make it unknown, or it is necessary to detect objects that
simpler for the blind to learn about items, the are very different from one another. In order to
object recognition deep learning model makes use state the motion of an object detected through the
of the You Only Look Once (YOLO) algorithm change observed in two consecutive frames after
and text-to-speech (TTS) voice announcement. attaining the results of the image processing
As a result, it incorporates an effective object- techniques used for identification, techniques of
detection system that aids the blind in finding image processing are applied to the frames in
objects in a particular environment without the aid order to identify any change.The blind individual
of others, and the system's effectiveness is can readily identify objects using cameras or other
examined through experiments. . Our technology sensors with the aid of object detection systems
comprises of a box containing a portable camera and navigation systems.
and a system for processing the image. Images are
collected with a portable camera device and
recognised in real time using existing object For instance, it can be exceedingly challenging to
detection algorithms. locate a specific room in a strange place. To
categorise an image into a predetermined number
1. Introduction of classes, the primary process is to blend clinical
Object detection is a well-known computer information in the previous words with data-
technique that is closely related to computer driven knowledge in the likelihood terms.. The
vision and image processing.The accuracy of following is a brief overview of each of these
object detection has grown dramatically with the processes.
introduction of deep learning techniques.It is
concerned with detecting objects or instances of a
particular class (such as persons, flowers, and Read picture :- Reading training and test images
animals) in digital photos and videos. One of the is the initial stage in image processing. The image
should be placed in the current working directory
or else we need to provide the absoluate path.
Grayscale conversion of a colour image: A
grayscale image may only have one channel while
a colour image has three channels (Red, Blue, and
Green). Colour is not necessary to recognise and
identify an image. - transforming an image to
grayscale minimises the processing time and
memory space requirements for the image.
Standardise Images :-Images must be resized to
2. Related Work
basic heights and widths before being input to the
algorithm. This is done in the case of picture pre
processing and scaling. P. S. Udigirkar, S. Sarokar, and S. Gore "Object
Detection and translation for blind person " This
Advantages of Object Detection for blind person research suggests presented a smart vision
have several advantages. Some of the significant whose goal is to go anywhere in the
advantages include: surroundings using a user-friendly interface
With the help of object detection system and technology. This project is primarily concerned
navigation system the blind person can easily with the computer vision module. These writers
detect the object through cameras or through created a system that can detect obstacles
some sensors. For example, it is very difficult for around his head, particularly while entering via
to find a particular room in an unfamiliar a door.
environment.
By F.F. Alkhalid , B.K. Oleiwi , "Online Blind
It is quite simple and inexpensive pose refiner,It
Assistive System Using Object Detection." This
also shows robustness when the scene is
study introduces a The purpose of computer
extremely clusttered.
vision is to accurately recognise indoor objects.
The network converts the object detection The visually handicapped can be helped by
problem into a classification problem, traversing the CNN framework's
significantly improving accuracy. purposes.4,5,14 To begin identifying distinct
items, we must first detect the pixels present in
In terms of object detection, image processing
the photos.
techniques are applied to the frames in order to
identify any change in order to state the motion of
Using S.Lee, M.Kang a Deep learning
the object detected through the change observed
technology were used to develop an object
in two consecutive frames after obtaining the
detecting system for the blind. The authors
results of the image processing techniques used
employed voice recognition technology to
for identification.
determine what objects a blind person desired
Limitations of Object Detection for blind person and then used object detection to locate the
Despite the advantages, One of the most difficult objects.
aspects of object recognition is that an object
might appear radically different from different R. Raied Mahmood, Dr . M.D. Younus, and Dr.
angles. They contain very little information about E.A. Khalal "Real time object detection for
self-velocity items, which are necessary for visually impaired person Based on a system can
travel. The navigation systems are expensive, and benefit blind persons by providing a low-cost
the average blind person cannot afford the portable solution. It comprised of shoes equipped
with ultrasonic sensors that surveyed the scene.
The camera module is triggered when the sensors
detect an object in front of the person. After
detecting the object, the camera module takes a systems, to announce street names when
picture of the object in front of the people. providing instructions.

Proposed system on Object Detection for visually Word Recognition: Voice recognition software
impaired System by S. Ketkale, M. Mujawar, and (also known as speech to textual content
S. Londhe suggest the proposal in this study software) allows users to use their voice instead
focuses on providing information on the various of typing on a computer. Voice recognition can
types of barriers in front of the person, their be used to dictate text into a computer or to issue
length, and their distance from the person. The commands to the computer. Voice recognition
camcorder is used to capture video. Following software enables a quick method of writing on a
that, video processing methods are used.. computer. It is also useful for people with
disabilities who find it difficult to operate a
3. Proposed Method keyboard.

A Object Detection model is suggested a lot of

work has been done on object recognition in
movies utilising so-called feature extraction and
detection along the video or by extraction of
visually conspicuous regions that are assumed to
contain object of interest.

Because of its application in a variety of

recognition issues, feature matching with 4. Result Discussion
invariant features has grown in popularity. Such
techniques have allowed us to match images that In this study, we investigated various ways for
have undergone diverse geometric and detecting and recognising objects have been
photometric modifications. discovered, and they employ various sorts of
data as input for their methodology. Following
Object Recognition: The suggested device helps a survey of several ways, it was discovered that
those who are blind or visually challenged employing SSD Architecture model that was
identify objects that they would normally be trained on COCO datasets is the easiest method
unable to. that can be simply deployed and appropriate in
all scenarios.
Object Detection: Object Recognition and
Detection The software system uses object In the second phase, It examined all of the things
detection to recognise, track down, and identify that have been done thus far, including whether or
items in a given image or video. A specific image not they have been successful. Based on all of the
or video is used to identify, discover, and fine- information that has been analysed, there may be
tune objects on the software programme device hope for what may be done in the future. The
using the distinctive object detection feature. reverse may be used for accurate object detection
to detect objects through form and colour sample
Speech to Text: Textual material that has been popularity lards, people, and vehicles. This
recognised can be preserved in a document. We version is appropriate for real-time usage due to
are developing this on the Android platform with velocity, according to journals like "Object
the help of the Eclipse painting bench. Our Detection Using Convolutional Neural
speech-to-text computer collects and converts Networks."
speech to text directly.

Text to Speech: Text-to-speech is also utilised

on handheld devices, such as portable GPS
 individual using wireless audio assistive devices..
An increase in the training data either had a
favourable effect or had no effect in each step.

Illumination Problem: An illumination problem

occurs when a change in lighting causes a face to
appear different despite being the same. Systems
based on comparing images misclassify input
photos because the variations caused by
illumination are frequently greater than the
differences between people.

The Object detection Problem: It is not

Launching and Accessing Anaconda unexpected that when there are major position
fluctuations in the input photos, The main idea
behind object detection in movies is to verify the
presence of an object in image sequences and
maybe locate it specifically for recognition. This
is accomplished by solving the temporal
correspondence issue, which involves matching
the target region in successive frames of an image
series captured at tightly spaced time intervals.
Detection of Objects - Chair
Distance Problem: Our Object detection system
for blind person fails if the distance between the
camera and the object exceeds more.

Camera Issue: A camera's resolution is crucial

for correctly identifying an image. The more
accurate the camera, the better it will be at
recognising object
Detection of Objects - Chair and Mouse

REFERENCES

[1] Dr. M Y Babu, Akash Jatavath Object

Detection system with voice alert , Vol 9 S(1),
DOI: 10.17485/ijst/2016/v9iS(1)/108446
December 2021.

[2] Shifa Shaikh, Vrushali Karale, Gaurav Tawde

Detection of Water Bottle
Design and implementation of smart shoes for
blind and visually impaired Real time object
The laptop's system will test it using its APIs and detection for blind person in Issue 09, September-
SSD ALGORITHM, and it will determine the 2022.ref figure5.1Vision,,Vol. 9905, pp. 21-37,
image's confidence accuracy. For several classes, Sept. 2018.
such books, mugs, and remotes, we achieved 98%
accuracy. Following the testing of the photos, we
are producing an output on a laptop-based system, [3] Hanen Jabnoun, Faouzi Benzarti, Hamid
Amiri Object Detection and identification for
and our forecast is converted into voice using
voice modules and communicated to the blind
blind people in vedio scene: Signal, Images et
Technologies de l’Information (LR-SITI-ENIT).

[4] Mayuresh Banne , Rahul Vhatkar, Ruchita

Tatkare Object Detection and Translation for
Blind People Using Deep Learning, international
Research Journal of Engineering and Technology
.(IRJET) e-ISSN: 2395-0056
[5] Rajeshwari Ravindra Karmarkar object
detection system for blind person with voice
guidance International Journal of Engineering
Applied Sciences and Technology, 2021 Vol. 6,
Issue 2, ISSN No. 2455-2143.

[6] Aniket Birambole, Pooja Bhagat,

Bhavesh Mhatre, Prof. Aarti Abhyankar ” Indian
Journal of Marine science 38, no. 3 (2009): 324-
331.