Proceedings of the International Conference on Smart Electronics and Communication (ICOSEC 2020)

IEEE Xplore Part Number: CFP20V90-ART; ISBN: 978-1-7281-5461-9

Student Attendance System using Face Recognition

Samridhi Dev Tushar Patnaik
Department of computer science Department of Computer Science
C-DAC, Noida C-DAC, Noida
Samridhi9927@g mail.co m tusharpatnaik1108@gmail.co m

Abstract-: Face recogniti on is among the most producti ve these digital methods such as biometric fingerprinting
image processing applicati ons and has a pi votal role in the techniques and card swapping techniques. however, these
technical fiel d. Recognition of the human face is an acti ve methods prove to be a statute of limitations as it subjects
issue for authentication purposes specifically in the context students to wait in a time-consuming queue. if the student fails
of attendance of students. Attendance system using face to bring his id card then he will not be able to get attendance.
recognition is a procedure of recognizing students by using evolving technologies have made many imp rovements in the
face bi ostatistics based on the high defini tion monitoring changing world.
and other computer technologies. The development of this The system of intelligent attendance is generally implemented
system is aimed to accomplish digitizati on of the traditi onal with bio metrics help. Recognition of face is one of the
system of taking attendance by calling names and Bio metric ways of improving this system. Face recognition
maintaining pen-paper records. Present strategies for proved to be a productive method for taking attendance. The
taking attendance are tedious and ti me -consuming. normative face recognition techniques and methodologies fail
Attendance records can be easily mani pulated by manual to tackle challenges like scaling, pose, illu mination, variat ions,
recording. The tradi tional process of making attendance rotation, and occlusions. The framework proposed is designed
and present biometric systems are vulnerable to proxies. to solve the drawbacks of current systems. there has been a lot
This paper is therefore proposed to tackle all these of advancement in face recognition but the vital steps are face
problems. The proposed system makes the use of Haar detection, feature extraction, and face recognition. firstly, two
classifiers, KNN, CNN, SVM, Generati ve adversarial or more cameras depend on the need, and the size of the
networks, and Gabor filters. After face recogni tion classroom has to be installed on the ceiling of the classroom
attendance reports will be generated and stored in excel fro m where it covers the entire area. image captured fro m these
format. The system is tested under various condi tions like cameras will be considered as an input to the system. There
illumination, head movements, the vari ati on of distance may be a possibility of getting image b lurred due to movements
between the student and cameras. After vig orous testing of students, for better efficacy image can be upgraded using
overall complexity and accuracy are calculated. The Generative Adversarial Net works. A newly generated
Proposed system proved to be an efficient and robust device ameliorated image will be passed to the system for face
for taking attendance in a cl assroom without any ti me detection. process of face detection is acco mpanied by feature
consumption and manual work. The system developed is extraction and face recognition these process makes the use of
cost-efficient and need less installation. Gabor filters. face recognition is done using the K-nearest
neighbor algorithm, Convolutional neural networks, and SVM
Keywords – KNN, SVM , VIOLA-JONES, HAAR classifiers, algorith m with their co mparat ive studies. post-complet ion of
CNN. face recognition, the system generates the name and
identification nu mber o f the students who are present and
I. INT RODUCT ION identified in the image. then attendance is marked in front of
the student names in the excel fo rmat with respective date and
Attendance being a very necessary side of ad min istration may subject of a lecture in an institution. It requires very few
normally beco me an arduous, redundant activity, pushing itself hardware resources hence it is a cost-friendly system.
to inaccuracies. The tradit ional approach of making ro ll calls
proves itself to be a statute of limitations as it is very difficult to
call names and maintain its record especially when the ratio of
students is high. Every organizat ion has its way of taking
measures for the Attendance of students. So me organizations
use document-oriented Approach and others have implemented

978-1-7281-5461-9/20/$31.00 ©2020 IEEE 90

Authorized licensed use limited to: Middlesex University. Downloaded on November 01,2020 at 18:41:07 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the International Conference on Smart Electronics and Communication (ICOSEC 2020)
IEEE Xplore Part Number: CFP20V90-ART; ISBN: 978-1-7281-5461-9

captured and matched with the database and SMS is sent to the
Above statement can be illustrated as figure 1 below authorized number. major steps performed in this approach for
generating the features are scale-space extreme detection,
keypoint localization, o rientation assignment, keypoint
descriptor. As soon as the face is recognized by the system
LED on the Arduino board will start blinking.

In [3] author studied [6] t wo-stage hybrid face detection

scheme wh ich uses the probability-based Face Mask Pre-
Filtering (PFMPF)and the pixel-based hierarchical Feature Ad
boosting (PBHFA) method. This approach is aimed to solve the
problem in Haar cascade. The author proposed a system with
two phases, training phase, and testing phase. In the training
phase, they presented two main steps first is face detection in
which they have used the viola jones algorithm. the second step
which is feature extract ion after detecting faces fro m a video
feature is ext racted using the PCA algorith m. in the testing
phase, the data set is partitioned into two parts named training
dataset and testing dataset.
Figure 1 operating process of attendance system
In [4] the author used the convolutional neural network (CNN)
to obtain low dimensional features as the pre-processed images
are too high dimensional for a classifier to take it as input
II. LIT ERAT URE REVIEW directly. For face detection, they have used the viola and jones
algorith m and then used correlation tracker to track face fro m
The primary aim of this paper is to study the different frame to frame. In this paper, the author has worked on several
approaches given by authors and to develop a real-time parameters like pose estimation, sharpness, resolution, and
attendance system wh ich overco mes the shortcomings of brightness. The head position is determined using three-angle
previous methods and to give the best solution. roll, yaw, and pitch. Then approach includes final score
calculation named face quality assessment by assigning weights
In [1] Yoh ie Kawaguchi et.al proposed a system based on to each of the normalized parameters.
continuous observation and using face recognition. The author
presented a system with an act ive student detecting method In [5] the author presented the system wh ich used the
(ASD) having two cameras placed on the wall in which one is a Eigenfaces approach for face recognition. They have performed
sensing camera which is used for estimating seat inside the face detection followed by a cropping of faces then worked on
class and the other is capturing camera wh ich is used for face background subtraction for greyscale images and binary
detection. They have proposed a shooting plan in wh ich one images. The author has used the Eigenface method due to its
seat is estimated fro m the seating area obtained by ASD and simplicity, speed, and learning capability.
then directs the capturing camera to the seat and captures an
image. The e xistence of students is estimated using background In [7] Sav itha et al proposed the system which uses skin
subtraction and inters frame subtraction. The author has solved detection technique for face detection. after the skin is detected
the linear sum assignment problem to give the correspondence skin pixels are taken and the rest of the p ixels in images will be
of students and seats. made black. Then these skin pixels will be used for face
detection authors have used two databases, the first database is
Paper proposed by [2] introduced an automated system based for storing faces of students and the second database is for
on convolutional neural networks . The author has used the storing data of students.
GSM modu le to send the generated attendance report to an
authorized person. The author proposes the modified
convolutional neural network by adding two normalization
operations to two of the layers. This operation provides the
batch normalization accelerat ion of the network. The face
recognition system is designed using the SIFT algorith m. This
system will take attendance using MATLAB. The image will be

978-1-7281-5461-9/20/$31.00 ©2020 IEEE 91

Authorized licensed use limited to: Middlesex University. Downloaded on November 01,2020 at 18:41:07 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the International Conference on Smart Electronics and Communication (ICOSEC 2020)
IEEE Xplore Part Number: CFP20V90-ART; ISBN: 978-1-7281-5461-9

III. PROPOSED SYST EM system. This system captures single image for a student for
training purpose.
A. Architecture With the aid of all the p ictures the student has stored in the
database, facial recognition for all of the students attending a
The proposed system is very simp le, effort less, and manageable lecture. It can be accomplished.
with lucid operations. It embraces a database of student's faces
and their details like name, enro lment number, course. two or Image amelioration
more cameras depending on the need and size of the classroom
are to be accommodated on the ceiling of the classroom Due to the movements of a student in a classroom, the image
covering the entire area. these cameras will capture images captured by the camera may get blurred. the image can be
several times during a lecture. this will increase the efficiency ameliorated using Generative Adversarial Networks. GANs
of the system because if the camera will not cover so me are known for their ability to retain texture information in
students then other cameras will capture their faces. there are images, create solutions similar to the actual range of
numerous expressions and poses possible which a student can aspects, and look perceptibly convincing.
perform. if at a part icular instance system fails to detect faces
due to unfavorable poses then the system can detect those faces ‫ ܤܫ‬ൌ ݇ሺ‫ܯ‬ሻ ‫ ݏܫ כ‬൅ ܰ
at another instance of image acquisition. Once the image
acquisition is done when the teacher triggers the system by where ‫ ܤܫ‬is a distorted image, k(M) is referred to as unknown
making a click on the start button thereafter system will blur kernels identified by mot ion field M. ‫ ݏܫ‬is the sharp latent
undergo face detection. after the faces are detected in an image
image and * symbolizes convolution whereas N denotes an
taken by all cameras at all g iven instances then detected faces
additive noise.
will be co mpared with stored images of the students in the
database. Once the face is matched then present is marked in
Face detection
front of its corresponding enrolment number and name in excel
format. though there are mu ltiple cameras and mu ltip le
For detection of faces 68 landmarks of faces are taken into
instances, there is a possibility of redundant faces. collaborated
account. with the help of these landmarks , faces are detected.
results will be generated by excluding redundant faces of the
For face detection, Haar classifiers have been used. It is an
same student so that single attendance is given to that student
approach based on mach ine learning in wh ich a cascade
during a lecture.
function is trained fro m many positive and negative images.
This is then used on other images to detect images. These
B. methodology
classifiers are simp ly the subtraction of the su m o f p ixels under
the black area fro m the sum of pixels under the white area.
Developing an intelligent attendance management system,
applying 6000 features on each window frame was found to be
some steps need to be followed to achieve this Successful task.
difficult. features were grouped into stages which are known as
The steps are definable as follows:
cascades of a classifier.
AdaBoost is used for removing redundant features and for
x Database creation selecting only appropriate features. These features are known as
x Image amelioration weak classifiers. A weighted combination of weak classifiers is
x Face detection used to detect faces. using the AdaBoost linear comb ination of
x Feature extraction weak classifiers is constructed known as a strong classifier.
x Face recognition
x Redundancy removal
ܲ ሺܻሻ ൌ ෍ሺܵ௜ ‫݌ כ‬௜ ሺܻሻ ሻ
x Report generation

Here, P(Y) is a strong classifier and sᵢ are corresponding weights

Database creation to each weak classifier ‫ ݅݌‬ሺܻሻ.

In the first step, the database will be created at the time of Feature extraction
enrollment of students. The database will store generic
informat ion of students like name, identification nu mber,
For feature extraction, Gabor filters are used to lay hold on
course, semester subjects. alongside the image of the student is
to be captured by the system for train ing of the proposed
facial features inclined at various angles. It is a very critical
step since it is believed that a successful feature extractor

978-1-7281-5461-9/20/$31.00 ©2020 IEEE 92

Authorized licensed use limited to: Middlesex University. Downloaded on November 01,2020 at 18:41:07 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the International Conference on Smart Electronics and Communication (ICOSEC 2020)
IEEE Xplore Part Number: CFP20V90-ART; ISBN: 978-1-7281-5461-9

selects a function that is not prone to occlusion, lighting, Redundancy removal

context, and pose variance. 2D Gabor filters are used to
resolve spatial distortions caused by position and lighting As the system encompasses multip le cameras. there might be a
variances. possibility of the presence of the face of a single student in
different images. redundant faces will be removed and single
௫మ ାఊమ ௬ మ faces will be considered to mark single attendance for a student
൬ି ൰ ‫ݔ‬ଶ
ܹ ሺ‫ݔ‬ǡ ‫ݕ‬ǡ ߠǡ ߣǡ ߮ǡ ߪǡ ߛሻ ൌ ݁ ଶఙమ …‘• ቆʹ ෑ ൅ ߮ቇ during a lecture.
ߣ
Report generation
‫ ݔ‬ᇱ ൌ ‫ ߠ •‘… ݔ‬൅ ‫ߠ ‹• ݕ‬
Trailing face recognition reports are generated by marking
‫ ݕ‬ᇱ ൌ െ‫ ߠ ‹• ݔ‬൅ ‫ߠ •‘… ݕ‬ present in front of the student name and enrollment nu mber in
excel format during a lecture.

here (x, y) defines the situation of a light impulse and μ, ϕ, γ, λ,

σ are parameters of the sinusoidal wavelet. IV. RESULT S

Face recognition the system was tested on three different algorithms out of wh ich
the KNN algorithm proved to be better with the accuracy of
For face recognition, the K-nearest neighbor algorith m have 99.27 %. The system was tested on various conditions which
been used, convolutional neural networks , and support vector include illu mination, head movements, expressions, the
mach ine. These three algorith ms are co mpared on the grounds distance of students from the camera. The system stands up to
of accuracy, robustness, time complexity. the expectations even when the image contains faces with
beards and spectacles and without beard and spectacles.
A. K-nearest neighbor algorithm proposed system evinced to be magnificent to recognize faces
having two years of difference. Being tested on these conditions
KNN is called memo ry-based or lazy learning since it only KNN proved to be better by achieving the overall accuracy of
preserves the interpretations of the training examp les as a result 97 %. when tested on conditions listed above CNN achieved
of the way it learns. the overall accuracy of 95 % and SVM achieved an accuracy of
The Euclidean d istance metric is often selected to Determine 88 %. viewing the aspect of time co mp lexity, CNN exposed to
the location of data points within KNN. An object is classified have low time co mple xity. It was found that SVM has the
according to the Voting done by the majority of its neighbors, highest time co mplexity among these three listed algorith ms.
with the object delegated to the most common class of its The proposed system is tested on 200 real-time images of a
nearest neighbors (k Is a positive integer). If k = 1, the object is classroom with a maximu m strength of 70 students. The
then it is allocated to his closest neighbor 's family. proposed system is robust enough to take attendance of 70
students in a classroom.
݀ሺ‫ݔ‬Ǥ ‫ݕ‬ሻ ൌ ඥሺ‫ ݔ‬ଵ െ ‫ݕ‬ଵ ሻଶ ǥ ǥ ǥ ሺ‫ ݔ‬௡ െ ‫ݕ‬௡ ሻଶ
The figure below shows the result of our proposed system
݀ሺ‫ݔ‬Ǥ ‫ݕ‬ሻ is the euclidean distance which is by default used by
KNN to find the nearest class.

B. Convolutional neural networks

Convolution Neural Networks allow us to derive fro m images a Figure 4 Proposed system
large variety of features. This concept of extracting the
functionality for face recognition can also be used. CNN uses
68 facial landmarks to generate 128-dimensional encodings V. COMPARISON
which are facial features encoded in RGB format. These
encodings are co mpared to match faces. The strictness of face In our proposed system Haar cascades are used for face
comparison can be manipulated by tolerance value. detection and generative adversarial networks for image
amelioration and for feature extraction Gabor filters were used.
For face recognition, different algorith ms were used. These

978-1-7281-5461-9/20/$31.00 ©2020 IEEE 93

Authorized licensed use limited to: Middlesex University. Downloaded on November 01,2020 at 18:41:07 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the International Conference on Smart Electronics and Communication (ICOSEC 2020)
IEEE Xplore Part Number: CFP20V90-ART; ISBN: 978-1-7281-5461-9

algorith ms have been compared on the grounds of time

variation of accuracy with different pitch angle using CNN
complexity accuracy in various conditions.
120
Table 1 shows the comparison of algorithms under listed 100

accuracy
80
conditions.
60
40
20
0
-30 -20 -10 0 10 20 30 40 50
pitch angle variation

Graph 2 Variation of accuracy with variation in pitch angle using CNN

variation of accuracy with different pitch angle using SVM

100
Table 1 Comparison 80

accuracy
60
40
A. Head movements 20
0
In a classroom head movements at different angles is possible -30 -20 -10 0 10 20 30 40 50

during a lecture. Head movements can be categorized into three pitch angle variation
categories measured in angles which are p itch, yaw, and Roll at
the respective x-axis, y-axis, and z-axis. Graph 3 Variation of accuracy with variation in pitch angle using SVM
Above listed algorithms were tested with different head
movements. Graphs below show the variat ion of algorith ms
accuracy with a variety of angles. 2) Yaw angle variation
For Ro ll and Yaw angles, the negative x-axis shows the head
movement in the left d irect ion and the positive x-axis shows the variation of accuracy with different yaw angle using KNN
head movement in the right direction.
120
For the Pitch angle, the negative x-axis shows the head 100
movement in the downward direction and the positive x-axis
accuracy

80
shows the head movement in the upward direction. 60
40
20
1) Pitch angle variation 0
-80 -60 -40 -20 0 20 40 60 80

variation of accuracy with different pitch angle using KNN yaw angle variation

120 Graph 4 Variation of accuracy with variation in Yaw angle using KNN
100
accuracy

80
60
variation of accuracy with different yaw angle using CNN
40
20 120
0 100
accuracy

-30 -20 -10 0 10 20 30 40 50 80

pitch angle variation 60
40
20
Graph 1 Variation of accuracy with variation in pitch angle using KNN
0
-80 -60 -40 -20 0 20 40 60 80
yaw angle variation

Graph 5 Variation of accuracy with variation in Yaw angle using CNN

978-1-7281-5461-9/20/$31.00 ©2020 IEEE 94

Authorized licensed use limited to: Middlesex University. Downloaded on November 01,2020 at 18:41:07 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the International Conference on Smart Electronics and Communication (ICOSEC 2020)
IEEE Xplore Part Number: CFP20V90-ART; ISBN: 978-1-7281-5461-9

variation of accuracy with different yaw angle using SVM

B. Different camera positions
100
80 Although cameras are to be fixed at the ceiling of classrooms,
accuracy

60
there may be a possibility of the varying distance between
40
20 student and camera as students can sit at different seats. A
0 system using three different algorith ms were tested under the
-80 -60 -40 -20 0 20 40 60 80 situation of the varying distance between student and cameras.
yaw angle variation The graph below shows the variation inaccuracy of the above-
mentioned algorithms.
Graph 6 Variation of accuracy with variation in yaw angle using SVM

VARIATION OF AC C U RAC IES U N D ER D IFFERN T

3) Roll angle variation C AM ERA D ISTAN C E
120

variation of accuracy with different roll angle using KNN 100

120
80
100

accuracy
accuracy

80
60 60
40
20 40
0
-80 -60 -40 -20 0 20 40 60 80
20
roll angle variation
0
1m 2m 3m 4m 5m 6m 7m
Graph 7 Variation of accuracy with variation in roll angle using KNN
KNN 100 100 99 98 97 90 70
CNN 100 100 98 97 96 88 65
variation of accuracy with different roll angle using CNN SVM 95 95 90 88 85 79 55

120 Graph 10 Variation of accuracy with variation in distance

100
accuracy

80
60 C. Overall result
40
20
0
Taking into account all the above-mentioned conditions and
-80 -60 -40 -20 0 20 40 60 80 situations overall accuracy, precision, recall, F1 score, and time
roll angle variation complexity of the algorithm are calculated.

Graph 8 Variation of accuracy with variation in roll angle using CNN The table listed below describes the above statement.

variation of accuracy with different roll angle using SVM

120
100
accuracy

80
60
40
20
0
-80 -60 -40 -20 0 20 40 60 80
roll angle variation

Graph 9 Variation of accuracy with variation in roll angle using SVM Table 2 result

978-1-7281-5461-9/20/$31.00 ©2020 IEEE 95

Authorized licensed use limited to: Middlesex University. Downloaded on November 01,2020 at 18:41:07 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the International Conference on Smart Electronics and Communication (ICOSEC 2020)
IEEE Xplore Part Number: CFP20V90-ART; ISBN: 978-1-7281-5461-9

VI. CONCLUSION [13] Wei Wu, Chuanchang Liu, Zhiyuan Su, “Novel Real-time Face
Recognition from Video Streams”, International Conference on Computer
Systems, Electronics and Control, 2017.
The proposed system meets the objective of achiev ing high
precision and less computational complexity. This system is [14] Changxing Ding, Dacheng T ao, “Trunk-Branch Ensemble Convolutional
Neural Networks for Video-Based Face Recognition”, IEEE transactions on
cost-efficient and less manual work is needed. Using Gabor pattern analysis and machine intelligence, 2018.
filters accuracy is highly improved. Fo r face recognition, Th ree
algorith ms have been used which are K-nearest neighbor, [15] Aziza Ahmedi, Dr Suvarna Nandyal, “An Automatic Attendance System
convolutional neural networks , and support vector mach ine, Using Image processing”, The International Journal of Engineering and
Science, 2015.
among these, the KNN algorith m proved to have the highest
accuracy of 99.27 %. Convolutional neural networks evinced to
have low co mputational complexity. SVM algorith m proved to
be less efficient

VII. REFERENCES

[1] Yohei Kawaguchi, Tetsuo Shoji, “Face Recognition-based Lecture

Attendance System”, “3 rd AERU…”, 2005.

[2] B. Kavinmathi, S.Hemalatha, “Attendance System for Face Recognition

using GSM module”, 4th International Conference on Signal Processing and
Integrated Networks”, 2018.

[3] Ketan N. Mahajan, Nagaraj V. Dharwadkar,” Classroom attendance system

using surveillance camera”, International Conference on Computer Systems,
Electronics and Control”,2017.

[4] Shubhobrata Bhattacharya, Gowtham Sandeep Nainala, Prosenjit Das,

Aurobinda Routray “Smart Attendance Monitoring System (SAMS): A Face
Recognition based Attendance System for Classroom Environment”, IEEE 18th
International Conference on Advanced Learning T echnologies, 2018.

[5] E.Varadharajan, R.Dharani, S.Jeevitha,., “Automatic attendance

management system using face detection”, 2017.

[6] Guo, Jing-Ming, “Complexity reduced face detection using probability-

based face mask prefiltering and pixel-based hierarchical-feature Ada
boosting”, Signal Processing Letters, IEEE 2011.
.
[7] K.Senthamil Selvi1, P.Chitrakala, A.Antony, Jenitha S, “face recognition
based attendance marking system”, International Journal of Computer Science
and Mobile Computing, 2014.

[8] Chen, Joy Iong Zong. "Smart Security System for Suspicious Activity
Detection in Volatile Areas." Journal of Information T echnology 2, 2020.

[9] Jacob, I. Jeena. "Capsule network based biometric recognition system."

Journal of Artificial Intelligence 1,2019.

[10] Kirtiraj Kadam, Manasi Jadhav, Shivam Mulay, T ushar Indalkar,

“Attendance Monitoring System Using Image Processing and Machine
Learning”, International Journal of Advance Engineering and Research
Development, 2017.

[11] Rajat Kumar Chauhan, Vivekanand Pandey, Lokanath M, “ Smart

Attendance System Using CNN”, International Journal of Pure and Applied
Mathematics,2018.

[12] Mayank Yadav, Anmol Aggarwal, “Motion based attendance system in

real time environment for multimedia application”, 2018.

978-1-7281-5461-9/20/$31.00 ©2020 IEEE 96

Authorized licensed use limited to: Middlesex University. Downloaded on November 01,2020 at 18:41:07 UTC from IEEE Xplore. Restrictions apply.