International Research Journal of Innovations in Engineering and Technology (IRJIET)

ISSN (online): 2581-3048

Volume 6, Issue 4, pp 100-103, April-2022
https://doi.org/10.47001/IRJIET/2022.604021

Facial Recognition Based Smart Attendance System

1
Saurabh Bhosale, 2Ajay Patil, 3Vilas Ovhal, 4Mayur Sonawane, 5Prof. Dr. Sushen Gulhane
1,2,3,4
Student, B.E., Information Technology, Dr D Y Patil College of Engineering Ambi, Pune, Maharashtra, India
5
Professor, B.E., Information Technology, Dr DY Patil College of Engineering Ambi, Pune, Maharashtra, India

Abstract - The main purpose of this project is to build a acquisition methodology, Face Recognition is divided into 3
face recognition-based attendance monitoring system for categories, namely Face Recognition Method based on image
educational institution to enhance and upgrade the current intensity (Intensity Image), a method that works based on
attendance system into more efficient and effective as video sequences and Faces Recognition Method which
compared to before. The current old system has a lot of requires other sensors such as 3D Information and infra-red
ambiguity that caused inaccurate and inefficient of image (infra-red imagery).
attendance taking. Many problems arise when the
authority is unable to enforce the regulation that exists in There are two categories in intensity mages -based Face
the old system. The technology working behind will be the Recognition, feature-based and holistic face recognition.
face recognition system. The human face is one of the Feature-based face recognition based is a face recognition
natural traits that can uniquely identify an individual. method that processes input images to identify and to extract
Therefore, it is used to trace identity as the possibilities for features of faces such as nose, mouth, eyes, etc. then calculate
a face to deviate or being duplicated is low. In this project, the geometric relationships between these face points thereby
face databases will be created to pump data into the reducing the face input images to geometric features vectors.
recognizer algorithm. Then, during the attendance taking The initial face recognition research based on feature-based
session, faces will be compared against the database to was conducted by, that study uses the Geometrical Features
seek for identity. When an individual is identified, its and Template Matching method in identifying a person's face.
attendance will be taken down automatically saving In that study using the integral projection feature extraction
necessary information into a excel sheet. At the end of the method and the classification method used is the Bayesian
day, the excel sheet containing attendance information Classifier.
regarding all individuals are mailed to the respective
II. METHODOLOGY
faculty.
According to the previous attendance management system,
Keywords: face recognition, attendance monitoring, and
the accuracy of the data collected is the biggest issue. This is
technology.
because the attendance might not be recorded personally by
I. INTRODUCTION the original person, in another word; the attendance of a
particular person can be taken by a third party without the
An attendance system using a camera-based on a face realization of the institution which violates the accuracy of the
recognition system has been developed by several researchers data. For example, student A is lazy to attend a particular
to produce a face attendance system that is accurate and able class, so student B helped him/her to sign for the attendance
to store a large-scale face image database. The design and which in fact student A didn’t attend the class, but the system
manufacture of facial attendance systems using cameras are overlooked this matter due to no enforcement practiced.
very useful effective and efficient to further improve the Supposing the institution establish enforcement, it might need
accuracy of user data and is useful for the high mobility of to waste a lot of human resource and time which in turn will
users who use it. Facial attendance system using a camera is not be practical at all. Thus, all the recorded attendance in the
very safe and accurate for detecting users because it has a previous system is not reliable for analysis usage. The
more accurate data process and high accuracy so that it can problem of the previous system is where it is too time
produce a system that is reliable and robust to identify human consuming. Assuming the time taken for a student to sign
faces to be used as a time attendance on attendance machines. his/her attendance on a 3-4 paged name list is approximately 1
Research on Face Recognition for human interests has been minute. In 1 hour, only approximately 60 students can sign
carried out, especially for various interests such as security their attendance which is obviously inefficient and time
systems, surveillance, General identity verification, Image consuming. The third issue is with the accessibility of that
database investigations, Criminal justice systems, "Smart information by the legitimate concerned party. For an
Card" applications, video indexing, Multi-media environments example, most of the parents are very concerned to track their
and Witness face reconstruction. According to, based on data child’s actual whereabouts to ensure their kid really attend the

© 2022-2017 IRJIET All Rights Reserved www.irjiet.com 100

 International Research Journal of Innovations in Engineering and Technology (IRJIET)

ISSN (online): 2581-3048

Volume 6, Issue 4, pp 100-103, April-2022
https://doi.org/10.47001/IRJIET/2022.604021

classes in college/school. However in the previous system, region of interest (RoI) coordinate point of the face image
there are no ways for the parents to access such information. which is the location point and the size of the RoI face image
Therefore, evolution is needed to be done to the previous produced. Normalization or preprocessing is a process that
system to improve efficiency, data accuracy and provides result a face image that has detected on process of face
accessibility to the information for that legitimate party. The detection.
main intention of this project is to solve the issues encountered
in the old attendance system while reproducing a brand new In this normalization phase, a combination of several face
innovative smart system that can provide convenience to the image processing models is used. We used the cropping
institution. In this project, an application will be developed method, resizing, RGB-Gray, and using histogram
which is capable of recognizing the identity of each equalization as a contrast-brightness adjustment to optimize
individuals and eventually record down the data into a the facial recognition. The preprocessing method is used to
database system. Apart from that, an excel sheet is created improve the sharpness of the image to anticipate several
which shows the students attendance. variations in illumination that commonly appear when
capturing facial images.

Convolutional Neural Networks Algorithm

In this research CNN algorithm is used to detect the

disease from given ECG image. We have used three layers of
the CNN these are Convolutional Layer, Pooling Layer and
fully connected layer. Here, convolutional layer and pooling
layer, both layer work together. First it represents the image
into three-dimension vector space then applies filters to
convert that image into 2 dimensions. Again, fully connected
layer applies the filters to recognize the image. After training
the system when we pass the image to test it, as per the
specified categories it classifies the image and detects the
disease accurately. After disease detection system suggests the
medicine on the detected disease.

The Conv layer is the core building block of a

Convolutional Network that does most of the computational
Figure 1: Architecture Diagram
heavy lifting. To summarize, the Conv Layer:
In this research, the development of a face recognition
Accepts a volume of size W1×H1×D1
model as a face attendance machine using a hybrid feature
extraction method using CNN-PCA was built using a Requires four hyper parameters:
combination of face detection and face recognition framework
model using real-time cameras that function as a face Number of filters K,
detection tool and human face identification. The stages of the
facial recognition process that will be carried out consist of the Their spatial extent F,
processes performed on data acquisition, face detection
The stride S,
process, preprocessing, feature extraction process, and
classification processes can be seen in Fig. 2. The camera that The amount of zero padding P.
used in face detection process is performed using face
detection based on the Viola-Jones from the OpenCV package. Produces a volume of size W2×H2×D2 where:
Face detection is done by a camera to take face images of
objects taken. The image taken from the camera lens is a raw W2=(W1−F+2P)/S+1
image containing a background image and a face image. In
H2=(H1−F+2P)/S+1 (i.e. width and height are computed
this face detection process is carried out the process of
equally by symmetry)
detecting and searching for facial features in the camera
image, which at this stage the system recognizes patterns as D2=K
faces or not. The face image detected in the background image
produced by each camera lens is marked by the position of the

© 2022-2017 IRJIET All Rights Reserved www.irjiet.com 101

 International Research Journal of Innovations in Engineering and Technology (IRJIET)

ISSN (online): 2581-3048

Volume 6, Issue 4, pp 100-103, April-2022
https://doi.org/10.47001/IRJIET/2022.604021

With parameter sharing, it introduces F⋅F⋅D1 weights per

filter, for a total of (F⋅F⋅D1)⋅K weights and K biases.

In the output volume, the d-th depth slice (of size W2×H2)
is the result of performing a valid convolution of the dd-th
filter over the input volume with a stride of S, and then offset
by dd-th bias.

Notations Description
𝑎𝑖𝑙 Output feature map of kernel.
Figure 3: Output Screen
𝑤𝑖𝑙 Is the Weight matrix.
𝑎𝑙−1 Represents the input feature-
map of the layer.

In this system, face databases will be created to pump data

into the recognizer algorithm. Then, during the attendance
taking session, faces will be compared against the database to
seek for identity. When an individual is identified, its
attendance will be taken down automatically saving necessary
information into a excel sheet. At the end of the day, the excel
sheet containing attendance information regarding all
individuals are mailed to the respective faculty.

Figure 4: Output Screen

III. RESULTS AND CONCLUSION

In this project, several studies on face recognition, there

are still many methods and algorithms that have not been
investigated specifically the use of reconstruction algorithms
from 2D images to 3D forms that are used as a database in
face recognition. In this study, a study of facial recognition
using an approach to the development of 2D to 3D image
reconstruction models using Convolutional Neural Network
Figure 2: Input data into system (CNN) and the use of PCA are used as the feature extraction
method. The CNN method is used to produce a 3D face image
from a 2D face image. The PCA method used as a feature
extraction method used as a classification method on the
proposed face recognition based attendance system can work
well. The proposed method can produce a face recognition
that has a high accuracy of up to 98%.

REFERENCES

[1] Harguess, J., Hu, C., Aggarwal, J. K., 2009, Fusing

Face Recognition from Multiple Cameras, 978-1-4244-
5498-3/09, IEEE.
[2] Kim, J., Choi, J., Yi, Y., 2004, ICA Based Face
Recognition Robust to Partial Occlusions and Local

© 2022-2017 IRJIET All Rights Reserved www.irjiet.com 102

 International Research Journal of Innovations in Engineering and Technology (IRJIET)

ISSN (online): 2581-3048

Volume 6, Issue 4, pp 100-103, April-2022
https://doi.org/10.47001/IRJIET/2022.604021

Distortions," International Conference on [11] “Attendance System Using NFC Technology with
Bioinformatics and its Applications. Fort Lauderdale, Embedded Camera on Mobile Device” (Bhise, Khichi,
Florida, USA, 2004, pp.147-154. Korde, Lokare, 2015).
[3] He, X., Yan, S. C., Hu, Y., Niyogi, P.,Zhang, H. J., [12] K.Senthamil Selvi, P.Chitrakala, A.Antony Jenitha,
2005, Face Recognition Using Laplacianfaces, IEEE "Face Recognition Based Attendance Marking
Transactions on Pattern Analysis and Machine System", IJCSMC, Vol. 3, Issue 2, February 2014.
Intelligence, Vol.27, pp.328-340, 2005. [13] “Fingerprint Based Attendance System Using
[4] Mazloom, M., Ayat, S., Combinational Method for Microcontroller and LabView” (Kumar Yadav, Singh,
Face Recognition: Wavelet, PCA and ANN, Digital Pujari, Mishra, 2015).
Image Computing: Techniques and Applications, [14] “RFID based Student Attendance System” (Hussain,
IEEE: 978-0-7695-3456- 5/08. Dugar, Deka, Hannan, 2014).
[5] Agarwal, M., Agrawal, H., 2010, Face Recognition
using Principle Component Analysis, Eigen face and AUTHORS BIOGRAPHY
Neural Network, International Conference on Signal
Saurabh Bhosale, Student, B.E.,
Acquisition and Processing, 978-0-7695-3960-7/10, Information Technology, Dr D Y
IEEE. Patil College of Engineering Ambi,
[6] Kim, J., M., Kang, M., A., 2010, A Study of Face Pune, Maharashtra, India.
Recognition using the PCA and Error Back-
Propagation, Second International Conference on
Intelligent Human-Machine Systems and Cybernetics. Ajay Patil, Student, B.E.,
Information Technology, Dr D Y
[7] Kelsey, R., Daniel, C., Jesús, O., 2011, A Face
Patil College of Engineering Ambi,
Recognition Algorithm using Eigen phases and Pune, Maharashtra, India.
Histogram Equalization,
INTERNATIONALJOURNAL OF COMPUTERS
Issue 1, Volume 5.
[8] X. Zhu, Z. Lei, X. Liu, H. Shi, and S. Z. Li, “Face
Vilas Ovhal, Student, B.E.,
alignment across large poses: A 3D solution”, In Proc. Information Technology, Dr D Y
IEEE Conference on Computer Vision and Pattern Patil College of Engineering Ambi,
Recognition, pages 146 – 155, Las Vegas, NV, June Pune, Maharashtra, India.
26-July 1 2016.
[9] A.Jourabloo and X. Liu, “Large-pose face alignment
via CNN-based dense 3D model fitting”, In Proc. IEEE
Con-ference on Computer Vision and Pattern Mayur Sonawane, Student, B.E.,
Recognition, pages 4188 – 4196, Las Vegas, NV, June Information Technology, Dr D Y
Patil College of Engineering Ambi,
26-July 1 2016.
Pune, Maharashtra, India.
[10] E. Richardson, M. Sela, and R. Kimmel, “3D face
reconstruction by learning from synthetic data”, In
Proc. International Conference on 3D Vision, pages
460–469, California, USA, October 25-28 2016.

Citation of this Article:

Saurabh Bhosale, Ajay Patil, Vilas Ovhal, Mayur Sonawane, Prof. Dr. Sushen Gulhane, “Facial Recognition Based Smart
Attendance System” Published in International Research Journal of Innovations in Engineering and Technology - IRJIET,
Volume 6, Issue 4, pp 100-103, April 2022. Article DOI https://doi.org/10.47001/IRJIET/2022.604021

*******

© 2022-2017 IRJIET All Rights Reserved www.irjiet.com 103

© 2022. Notwithstanding the ProQuest Terms and Conditions, you may use
this content in accordance with the associated terms available at
https://irjiet.com/about_open_access