Automated Attendance
defined as “The action or state of going regularly to or
system with Multi-faces
using Convolution
Neural Network (CNN)
Zunaira Naveed
Department of Computer Science &IT
Institute of Southern Punjab, Multan, Pakistan
mailto:Zunairanaveed85@gmail.com
Abstract— Today, data provided by internet-
connected intelligent machines is growing gradually.
To process most of the
data produced by the Internet of Things(IoT), it requi
res an effective and efficient system. CNN is used to
classify images and perform image identification in
scenes. Deep learning is the developing technology
that is used to process huge amounts of information
efficiently. Authentication is the main problem in this
exclusive world of data systems.
Several methods have been used to solve this issue an
d various techniques have been actually employed to
solve this problem. Face recognition is one of the mo
st secure varieties. The only facial recognition system
offers automatic property. This article praises an
algorithm based on convolution neural networks
(CNN) for face detection and recognition that beats
ordinary methods.
A smart student attendance classroom is provided wit
h face recognition to validate the performance of the
suggested algorithm. The proposed system achieved
97.9 percent precision on the test information. This
recommended system for automatic attendance system
seeks to be less time-consuming compared to the
current system for marking attendance.
Keywords— Automated attendance system,
Attendance system, Multi-faces recognition,
CNN, Attendance system in classroom.
I. INTRODUCTION
The popular medium for verification of identity are
pins and passwords, used for security (Bae, Cho et al.
2015). According to Oxford dictionary attendance is
being present at a place or event". It is said that
students with poor attendance have poor academic
achievement(Ahmedi,Nandyaletal.2015). The teacher
shall keep a suitable record on a regular basis only for
registration that takes a long time and can be hectic
(Krishnan,Renukaetal.2016). After several years of
development, biometric testing is used. One of the
most popular biometric tests is facial verification.
Biometric factors need to recognize acceptance,
uniqueness, durability, measurability and efficiency
(Jogiji,Ghateetal.2017).In modern society, biometric
systems have played a huge part in the recognition
system. These systems can identify individuals from
their physiological or behavioral
features(YounisandMuhammad2018).
Over the previous few years, biometric face recogniti
on technology has received a lot of attention due to it
s broad spectrum of applicability in law enforcement
and institutes and organizations. The
face recognition model can also recognize the individ
ual's contact without touching the person.
Furthermore, the face recognition system also helps
to
prevent crime, since the captured image can be stored
in a repository and can be useful in many ways in ide
ntifying a person.
Face recognition applications are introduced on socia
l media pages such as Facebook, entrance to airports,
railway stations, bus stop, extremely fortified regions,
advertising, and health care. These applications
minimize criminal jobs and fake authentication.(Kha
n, Harous et al. 2019).
The main problem with this approach is that due to se
parate viewing perspectives, illumination or shadowi
ng it suffers from
an uncontrolled variation of face images. Scientists pr
oposed a range of local feature filtering alternatives t
o address this challenge to obtain solid face recogniti
on.
Deep learning is focused on a system of face recognit
ion that delivers outstanding efficiency in comparison
with current methods. Deep learning methods use ma
chine neural network architects to acquire and classif
y facial vectors from raw facial images(Zhiyao and Di
2019).
The complicated features exclusively recognize the di
fferent images by using different hurdles such as illu
mination, posture
and aging for face recognition reasons, and large data
sets are required.
III. METHODOLOGY
Over the previous few years, there has been an
Face recognition techniques use the convolutional
increase in facial recognition systems. With limited f
neural network (CNN) to train face data sets and
aces in the image, most facial recognition systems are provide high accuracy as it takes characteristics from
presently expected to perform well. faces automatically and learns these attributes. In the
Similarly, these methodologies provide suitable facial domain of computer vision, deep learning performs a
poses and images that are not blurred. In making the very significant role as face recognition, segmentation,
optical character recognition, and age assessment.
schools smart, face recognition also has its impacts. Neural network software inventing is capable of
The accuracy of the identification task has describing the picture or video events.
significantly improved for the sake of
deep learning algorithms
due to the availability of high computational energy.
The suggested algorithm uses the Convolution Neural Fig 1: Face recognition building blocks
Network, a deep learning methodology for attaining The following construction blocks are generally
efficiency when image has multi faces. comprised of facial identification schemes:
i. Facedetection: A face detector removes the
II. LITERATURE REVIEW facial characteristics in the picture if it finds
Kohonen presents that a simple neural network the facial place back with bounding box
could perform face recognition aligned and normalized coordinates.
face images. The network is used to calculate a face
classification by calculating the autocorrelation matrix ii. Face alignment: In face alignment face
of the face images, known as "eigenfaces". A feasible images are crop with same size on the basis of
effort was needed for specific alignment and landmarks, created by landmark detector. Fig
normalization(Zhiyao and Di 2019). Since we b and c shows the landmarks on faces.
understand that recognizing multiple faces is a very iii. Facematching: Templates are used to match
challenging job, to resolve this issue shallow training images in picture recognition, and other
and deep learning are well- known methods which rely methods such as HAAR characteristics and
on supervised learning. In order to achieve an estimated Gradient Histogram are used.
outcome, most of the methodologies use the Labeled
Faces in the Wild (LFW) for face recognition. The two-
sided image was intended in parallel with the front
image (Sivic, Everingham et al. 2009). The model is
practiced on a 92.4percent exact LFW dataset (Chen,
Cao et al. 2012).
86 percent accuracy was obtained by using the (a)
dataset and 93.1percent accuracy was achieved by
using LFW dataset. (Parkhi,Simonyanetal.2014). The
Hollywood Movies data collection was accurate to 90.7
percent using the spatial orientation field (Sivic,
Everingham et al. 2005). The convolution neural
network was used to acknowledge the image as it (b) (c)
deleted the features from the convolution neural
network's last inserted. 45 percent accuracy was
managed on LFW dataset(Sun, Wang et al. 2014).
Many researchers used interactive whiteboards and
other multimedia instruments to create intelligent
environments (Pacheco, Flores et al. 2018). To make
classrooms smarter deep learning methods are (a) Bounding boxes found by a face detector.
used(Winer, Cooperstock et al. 2002). The proposed (b) and
architecture used the Convolution Neural Network for
the purpose of detection and recognition. (c) Aligned faces and reference points.
 IV. PROPOSED METHODOLOGY:
The approach centered on face identification of my
suggested intelligent participation scheme. The first
high definition will take pictures of all learners and
make sure learners are not lacking. The scheme used
input picture with 128-dimensional embedding. For
testing Support Vector Machine (SVM) is used for
classification of cropped faces images and draw a
boxes on faces with name. After the classification all
student’s information is stored in Excel file and email
generate automatically to corresponding factuality
member. The implementation of proposed system and
it gave good results and showed all students and
marked the attendance.
Fig. 2. Smart attendance system implementation
V. CONCLUSION
In this article, we provided an automatic face-based
technique for student attendance. We trained a model
on different face images of students. This model
provides better results for all students and achieves
better accuracy. Results show that this model is better
for the attendance system and also fast for recognition.
Next, we will use CNN for detection, the literature
says its results are very good in terms of accuracy and
performance. This article proposes an algorithm for
face detection and recognition based on Convolution
Neural Networks (CNN). in this paper. In order to
validate the efficiency of the proposed algorithm. To
minimize human errors, an automatic attendance
system has been used. The primary obstacle of the
system is distance as a range obviously brings up the
image blurry so the system produces incorrect results
on the foggy faces in some situations.
REFERENCES
[1] Ahmedi, A., et al. (2015). "An Automatic Attendance System
Using Image processing." 4: 1-8
[2] Bae, M.-Y., et al. (2015). "Design and implementation of
automatic attendance check system using BLE beacon."
10(10): 177-186
[3] Jogiji, A., et al. (2017). "WSN Based Automatic Attendance
Monitoring System." 6(8).
[4] Khan, M. Z., et al. (2019). "Deep Unified Model for Face
Recognition based on Convolution Neural Network and Edge
Computing."
[5] T. Hassner, S. Harel, E. Paz, and R. Enbar, “Effective face
frontalization in unconstrained images,” in Proceedings of the
IEEE Conference on Computer Vision and Pattern
Recognition, pp. 4295–4304, 2015
[6] E. Zhou, Z. Cao, and Q. Yin, “Naive-deep face recognition:
Touching the limit of lfw benchmark or not?” arXiv preprint
arXiv:1501.04690, 2015.
[7] Parkhi, and A. Zisserman, “Vggface2: A dataset for
recognising faces across pose and age,” arXiv preprint
arXiv:1710.08092, 2017.
[8] Krishnan, R., et al. (2016). "Effective automatic attendance
marking system using face recognition with RFID." 158-162
[9] Nech and I. Kemelmacher-Shlizerman, “Level playing field
for million scale face recognition,” in 2017 IEEE Conference
on Computer Vision and Pattern Recognition (CVPR), pp.
3406–3415, IEEE, 2017.
[10] Younis, M. I. and R. S. J. W. P. C. Muhammad (2018). "IFRS:
An Indexed Face Recognition System Based on Face
Recognition and RFID Technologies." 101(4): 1939-1966.
[11] Zhiyao, Z. and B. M. Di (2019). Multiplet selection: a practical
study of multi-faces recognition for student attendance system.
Proceedings of the 2nd International Conference on Image and
Graphics Processing, ACM.
[12] A. Bansal, A. Nanduri, C. D. Castillo, R. Ranjan, and R.
Chellappa, “Umdfaces: An annotated face dataset for training
deep networks,” in Biometrics (IJCB), 2017 IEEE
[13] A. Bansal, C. Castillo, R. Ranjan, and R. Chellappa, “The
do’s and don’ts for cnn-based face verification,” arXiv preprint
arXiv:1705.07426, vol. 5, 2017.
[14] K. Simonyan and A. Zisserman, “Very deep convolutional
networks for large-scale image recognition,” arXiv preprint
arXiv:1409.1556, 2014.
[15] D. Erhan, V. Vanhoucke, A. Rabinovich, et al., “Going deeper
with convolutions,” Cvpr, 2015.
[16] D. Erhan, V. Vanhoucke, A. Rabinovich, et al., “Going deeper
with convolutions,” Cvpr, 2015
[17] K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning
for image recognition,” in Proceedings of the IEEE conference
 on computer vision and pattern recognition, pp. 770–778,
2016.
[18] R. Ranjan, C. D. Castillo, and R. Chellappa, “L2-constrained
softmax loss for discriminative face verification,” arXiv
preprint arXiv:1703.09507, 2017.
[19] W. Liu, Y. Wen, Z. Yu, M. Li, B. Raj, and L. Song, “Sphereface:
Deep hypersphere embedding for face recognition,” in The
IEEE Conference on Computer Vision and Pattern
Recognition (CVPR), vol. 1, 2017.
[20] Y. Wu, H. Liu, J. Li, and Y. Fu, “Deep face recognition with
center invariant loss,” in Proceedings of the on Thematic
Workshops of ACM Multimedia 2017, pp. 408–414, ACM,
2017
[21] H. Wang, Y. Wang, Z. Zhou, X. Ji, Z. Li, D. Gong, J. Zhou, and
W. Liu, “Cosface: Large margin cosine loss for deep face
recognition,” arXiv preprint arXiv:1801.09414, 2018.
[22] F. Wang, W. Liu, H. Liu, and J. Cheng, “Additive margin
softmax for face verification,” arXiv preprint
arXiv:1801.05599, 2018.
[23] J.Deng, J. Guo, and S. Zafeiriou, “Arcface: Additive angular
margin loss for deep face recognition,” arXiv preprint
arXiv:1801.07698, 2018.
[24] Y. Yamada, M. Iwamura, and K. Kise, “Deep pyramidal resid-
ual networks with separated stochastic depth,” arXiv preprint
arXiv:1612.01230, 2016.
[25] X. Wu, R. He, and Z. Sun, “A lightened cnn for deep face
representa- tion,” in 2015 IEEE Conference on IEEE
Computer Vision and Pattern Recognition (CVPR), vol. 4,
2015.
[26] Y. Sun, D. Liang, X. Wang, and X. Tang, “Deepid3: Face
recognition with very deep neural networks,” arXiv preprint
arXiv:1502.00873, 2015.
[27] S. Sankaranarayanan, A. Alavi, C. D. Castillo, and R.
Chellappa, “Triplet probabilistic embedding for face
verification and clustering,” in Biometrics Theory,
Applications and Systems (BTAS), 2016 IEEE 8th
International Conference on, pp. 1–8, IEEE, 2016.
[28] B. Kumar, G. Carneiro, I. Reid, et al., “Learning local image
descriptors with deep siamese and triplet convolutional
networks by minimising global loss functions,” in Proceedings
of the IEEE Conference on Computer Vision and Pattern
Recognition, pp. 5385–5394, 2016.
[29] D. S. Trigueros, L. Meng, and M. Hartnett, “Enhancing
convolutional neural networks for face recognition with
occlusion maps and batch triplet loss,” Image and Vision
Computing, 2018.
[30] Y. Wen, K. Zhang, Z. Li, and Y. Qiao, “A discriminative feature
learning approach for deep face recognition,” in European
Conference on Computer Vision, pp. 499–515, Springer,
2016.
[31] X. Zhang, Z. Fang, Y. Wen, Z. Li, and Y. Qiao, “Range loss
for deep face recognition with long-tail,” arXiv preprint
arXiv:1611.08976, 2016.
[32] A. Brock, T. Lim, J. M. Ritchie, and N. Weston, “Neural
photo editing with introspective adversarial networks,” arXiv
preprint arXiv:1609.07093, 2016
[33] W. Shen and R. Liu, “Learning residual images for face
attribute manipulation,” in 2017 IEEE Conference on
Computer Vision and Pattern Recognition (CVPR), pp. 1225–
1233, IEEE, 2017.
[34] Y. Lu, Y.-W. Tai, and C.-K. Tang, “Conditional cyclegan for
attribute guided face image generation,” arXiv preprint
arXiv:1705.09966, 2017.
[35] Y. Choi, M. Choi, M. Kim, J.-W. Ha, S. Kim, and J. Choo,
“Stargan: Unified generative adversarial networks for multi-
domain image-to- image translation,” arXiv preprint
arXiv:1711.09020, 2017.
[36] W. Yin, Y. Fu, L. Sigal, and X. Xue, “Semi-latent gan:
Learning to generate and modify facial images from
attributes,” arXiv preprint arXiv:1704.02166, 2017.
[37] D. Samaras, “Neural face editing with intrinsic image
disentangling,” in Computer Vision and Pattern Recognition
(CVPR), 2017 IEEE Conference on, pp. 5444–5453, IEEE,
2017.
[38] G. Lample, N. Zeghidour, N. Usunier, A. Bordes, L. Denoyer,
et al., “Fader networks: Manipulating images by sliding
attributes,” in Ad- vances in Neural Information Processing
Systems, pp. 5969–5978, 2017.
[39] Z. He, W. Zuo, M. Kan, S. Shan, and X. Chen, “Arbitrary
fa- cial attribute editing: Only change what you want,” arXiv
preprint arXiv:1711.10678, 2017.
[40] Y. Zhou and B. E. Shi, “Photorealistic facial expression
synthesis by the conditional difference adversarial
autoencoder,” arXiv preprint arXiv:1708.09126, 2017.
[41] H. Ding, K. Sricharan, and R. Chellappa, “Exprgan: Facial
expres- sion editing with controllable expression intensity,”
arXiv preprint arXiv:1709.03842, 2017.
[42] C. Donahue, A. Balsubramani, J. McAuley, and Z. C. Lipton,
“Se- mantically decomposing the latent spaces of generative
adversarial networks,” arXiv preprint arXiv:1705.07904,
2017.
[43] R. Huang, S. Zhang, T. Li, R. He, et al., “Beyond face rotation:
Global and local perception gan for photorealistic and identity
preserving frontal view synthesis,” arXiv preprint
arXiv:1704.04086, 2017.

[44] L. Tran, X. Yin, and X. Liu, “Representation learning by

rotating your faces,” arXiv preprint arXiv:1705.11136, 2017.

[45] G. Antipov, M. Baccouche, and J.-L. Dugelay, “Face aging

with conditional generative adversarial networks,” arXiv
preprint arXiv:1702.01983, 2017.

[46] Z. Zhang, Y. Song, and H. Qi, “Age progression/regression by

condi- tional adversarial autoencoder,” in The IEEE
Conference on Computer Vision and Pattern Recognition
(CVPR), vol. 2, 2017.

[47] T. Weyand, M. Andreetto, and H. Adam, “Mobilenets:

Efficient convo- lutional neural networks for mobile vision
applications,” arXiv preprint arXiv:1704.04861, 2017.
[48] M. Sandler, A. Howard, M. Zhu, A. Zhmoginov, and L.-C.
Chen, “Inverted residuals and linear bottlenecks: Mobile
networks for classifi- cation, detection and segmentation,”
arXiv preprint arXiv:1801.04381, 2018.

[49] S. Chen, Y. Liu, X. Gao, and Z. Han, “Mobilefacenets:

Efficient cnns for accurate real-time face verification on mobile
devices,” arXiv preprint arXiv:1804.07573, 2018.
Paper | Publisher: IEEE

[50] Nienten Lin, Sen Chen, Lihong Zhao et al., "Student-based

design of classroom roll-call system [J]", Modern
Education Technology, vol. 25, no. 7, pp. 113-
119, 2015.

[51] Y Wen, K Zhang, Z Li et al., A

Discriminative
Feature Learning Approach for Deep
Face Recognition [J], vol. 47, no. 9, pp. 499-515,
2016.