See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/348325203

Efficient Attendance Management Using Multiple Face Recognition

Article · December 2019

CITATIONS READS

0 826

1 author:

Talari Tirupal
G PULLAIAH COLLEGE OF ENGINEERING AND TECHNOLOGY
45 PUBLICATIONS 237 CITATIONS

SEE PROFILE

All content following this page was uploaded by Talari Tirupal on 08 January 2021.

The user has requested enhancement of the downloaded file.

53899 T.Tirupal et al./ Elixir Comp. Engg. 137 (2019) 53899-53904
Available online at www.elixirpublishers.com (Elixir International Journal)

Computer Engineering
Elixir Comp. Engg. 137 (2019) 53899-53904

Efficient Attendance Management Using Multiple Face Recognition

T.Tirupal, EVN Sai Praneeth, P.Srihari, K.Sai Kiran and Md Kaleel Pasha
Department of ECE, G.Pullaiah College of Engineering and Technology, Kurnool, India.

ARTICLE INFO AB S T RA C T
Art i cl e h i sto ry : Nowadays, classroom attendance of a student is a very important task and if taken
Received: 01 November 2019; manually wastes a lot of time. There are many automatic methods available for this
Received in revised form: purpose i.e., biometric attendance. All these methods waste time, waiting in a queue to
26 November 2019; detect their thumb, palm, face or iris in front of the scanning device. This paper describes
Accepted: 5 December 2019; the efficient algorithm that automatically marks attendance without any human
intervention and waste of time. The algorithm works with images of multiple faces.
K ey w o rd s Firstly, the images are captured by using a camera placed on the top of the blackboard.
Biometric, The captured images are processed for noise removal using a median filter. Then the
Attendance, faces are detected using the detection algorithm. Lastly, the detected face is recognized
Multiple faces, by comparing the faces in the database and the attendance is marked. A Graphical User
GUI. Interface (GUI) is designed for calculating the percentage of the match of faces.
© 2019 Elixir All rights reserved.

Introduction Although, in the identification scenario, the similarity

In recent years, biometric-based techniques [1, 2, 3, 4] between a given face image (i.e. probe) and all the face
have emerged as the most promising option for recognizing images in a large database (i.e. gallery) is computed.
individuals. These techniques examine an individual‘s Literature Review
physiological and behavioral characteristics to determine and Maintaining attendance is very important in all the
ascertain their identity instead of authenticating people and institutes for checking the performance of students. Every
granting them access to physical domains by using institute has its method in this regard. Some are taking
passwords, PINs, smart cards, plastic cards, tokens or keys. attendance manually using the old paper or file-based
Passwords and PINs are hard to remember and can be stolen approach and some have adopted methods of automatic
or guessed easily; cards, tokens, keys and the like can be attendance using some biometric techniques. But in these
misplaced, forgotten, purloined or duplicated; magnetic cards methods students have to wait for a long time in making a
can become corrupted and unreadable. However, an queue at the time they enter the classroom. Many biometric
individual‘s biological traits cannot be misplaced, forgotten, systems are available but the key authentication is the same is
stolen or forged. Face recognition is one of the least intrusive all the techniques. Every biometric system consists of the
and fastest biometrics compared with other techniques such enrolment process in which unique features of a person are
as finger print [5, 6] and iris recognition. For example, in stored in the database and then there are processes of
surveillance systems [7], instead of requiring people to place identification and verification. These two processes compare
their hands on a reader (finger printing) or precisely position the biometric feature of a person with a previously stored
their eyes in front of a scanner (iris recognition), face template captured at the time of enrollment. Biometric
recognition systems [8, 9] unobtrusively take pictures of templates can be of many types like Fingerprints, Eye Iris,
people‘s faces as they enter a defined area. There is no Face, Hand Geometry, Signature, Gait, and voice. Our system
intrusion or capture delay, and in most cases, the subjects are uses the face recognition approach for the automatic
entirely unaware of the process. People do not necessarily attendance of students in the classroom environment without
feel under surveillance or their privacy being invaded. students‘ intervention. Face recognition consists of two steps,
Owing to its use in several applications, face recognition in first step faces are detected in the image and then these
has received substantial attention from both research detected faces are compared with the database for
communities and the market, and there has been an emerging verification. Several methods have been proposed for face
demand for robust face recognition algorithms that can deal detection i.e. AdaBoost algorithm, the Float Boost algorithm,
with real-world facial images. A general statement of the Neural Networks [10], the S-AdaBoost algorithm, Support
automatic face recognition problem is simply formulated as Vector Machines (SVM) [11], and the Bayes.
follows: given still or video images of a scene, identify or Existing Methods for Face Detection [12]
verify one or more persons in the scene using a stored During the last decade, several promising face detection
database of faces. algorithms have been developed and published. Among these
In other words, the face recognition system generally three stands out because they are often referred to when
operates under one of two scenarios: verification (one-to-one) performance figures etc. are compared. This section briefly
or identification (one-to-many), wherein the verification presents the outline and main points of each of these
scenario, the similarity between two face images is measured algorithms.
and a determination of either match or non-match is made.

Tele: +919491236447
E-mail address: [email protected]
© 2019 Elixir All rights reserved.
53900 T.Tirupal et al./ Elixir Comp. Engg. 137 (2019) 53899-53904
1. Robust Real-Time Object Detection: By Paul Viola and light systems extract features from the captured images that
Michael J. Jones [8, 12]. This seems to be the first article do not change over time while avoiding superficial features
where Viola-Jones presents the coherent set of ideas that such as facial expressions or hair. Several approaches to
constitute the fundamentals of their face detection algorithm. modeling facial images in the visible spectrum are Principal
This algorithm only finds frontal upright faces but is in 2003 Component Analysis, Local Feature Analysis, neural
presented in a variant that also detects profile and rotated networks, elastic graph theory, and multi-resolution analysis.
views.
2. Neural Network-Based Face Detection: An image
pyramid is calculated to detect faces at multiple scales. A
fixed-size sub-window is moved through each image in the
pyramid. The content of a sub- window is corrected for non-
uniform lightning and subjected to histogram equalization.
The processed content is fed to several parallel neural
networks [10] that carry out the actual face detection. The
outputs are combined using logical AND, thus reducing the
number of false detections. In its first form this algorithm also
only detects frontal upright faces. Figure 1. Face Recognition
3. A Statistical method for 3d Object Detection [11] Fingerprint Recognition
applied to Faces and Cars: The basic mechanics of this Fingerprints are unique for each finger of a person
algorithm is also to calculate an image pyramid and scan a including identical twins. One of the most commercially
fixed size sub-window through each layer of this pyramid. available biometric technologies, fingerprint recognition
The content of the sub- window is subjected to a wavelet devices for desktop and laptop access are now widely
analysis [12, 13, 14, 15, 16, 17, 18, 19, 20] and histograms available from many different vendors at a low cost. With
are made for the different wavelet coefficients. These these devices, users no longer need to type passwords –
coefficients are fed to differently train parallel detectors that instead, only a touch provides instant access. Fingerprint
are sensitive to various orientations of the object. The systems can also be used in identification mode. Several
orientation of the object is determined by the detector that states check fingerprints for new applicants to social services
yields the highest output. Opposed to the basic Viola-Jones benefits to ensure recipients do not fraudulently obtain
algorithm and the algorithm presented by Rowley et al. this benefits under fake names.
algorithm also detects profile views.
Existing Methods for Recognition [21, 22]
1. Introduction to Biometrics: Biometrics is an automated
method of identifying a person or verifying the identity of a
person based on a physiological or behavioral characteristic.
Examples of physiological characteristics include hand or
finger images, facial characteristics. Biometric authentication
requires comparing a registered or enrolled biometric sample
(biometric template or identifier) against a newly captured
biometric sample (for example, captured image during a Figure 2. Fingerprint Recognition
login). During Enrollment, as shown in the picture below, a Palm Recognition
sample of the biometric trait is captured, processed by a Palm print recognition is a biometric authentication
computer, and store for later comparison. method based on the unique patterns of various characteristics
2. Biometric Authentication Technology: in the palms of people‘s hands. Palm prints and fingerprints
Biometric recognition can be used in Identification are often used together to enhance the accuracy of
mode, where the biometric system identifies a person [23] identification.
from the entire enrolled population by searching a database
for a match based solely on the biometric. Sometimes
identification is called ―one-to-many‖ matching. A system
can also be used in Verification mode, where the biometric
system [24, 25, 26] authenticates a person‘s claimed identity
from their previously enrolled pattern. This is also called
―one-to-one‖ matching. In most computer access or network
access environments, verification mode would be used.
3. Types of Biometrics: There are various types of biometric
techniques [27] that we observe in our daily life. Some of
them are shown below
Face Recognition
The identification of a person by their facial image can
be done in many different ways such as by capturing an Figure 3. Palm Recognition
image of the face in the visible spectrum using an inexpensive Gait Recognition
camera or by using the infrared patterns of facial heat The Machine Visions approach to gait recognition entails
emission. Facial recognition [28, 29, 30, 31] in visible light the acquisition of gait signals using one or more video-
typically model key features from the central portion of a cameras from a distance. Therefore, it requires an ambient
facial image. Using a wide assortment of cameras, the visible set-up. As a first common step, systems in this category use
53901 T.Tirupal et al./ Elixir Comp. Engg. 137 (2019) 53899-53904
techniques for video and image processing to detect the user's unique traits by which DNA recognition can identify or verify
image in a scene, to track the user's walk, and to extract gait the identification of a person.
features for user recognition. The cells that contain DNA share genetic material
(information) through chromosomes. Humans have 23
chromosomes that house a person‘s DNA and their genes. Of
the 46 total chromosomes, 23 come from each parent of an
offspring. 99.7% of an offspring‘s DNA is shared with their
parents. The remaining .3% of an individual‘s DNA is
variable repetitive coding unique to an individual. This
repetitive coding is the basis of DNA biometrics. DNA
recognition uses genetic profiling, also called genetic
fingerprinting, to isolate and identify these repetitive DNA
regions that are unique to each individual to either identify or
Figure 4. Gait Recognition
verify a person‘s identity.
Voice Recognition
Voice recognition is ―the technology by which sounds,
Start
words or phrases spoken by humans are converted into
electrical signals, and these signals are transformed into
coding patterns to which meaning has been assigned‖. While
Input Image
the concept could more generally be called ―sound
recognition‖, we focus here on the human voice because we
most often and most naturally use our voices to communicate Histogram Normalization
our ideas to others in our immediate surroundings. The Normal ization
difficulty in using voice as an input to a computer simulation
Noise Removal
lies in the fundamental differences between human speech
and the more traditional forms of computer input. While
computer programs are commonly designed to produce a Skin Classification
precise and well-defined response upon receiving the proper
(and equally precise) input, the human voice and spoken
words are anything but precise. Each human voice is Face Detection
different, and identical words can have different meanings if
spoken with different inflections or in different contexts.
Select Region of Interest
Iris Recognition
This recognition method uses the iris of the eye, which is
the colored area that surrounds the pupil. Iris patterns are
Graphic User Interface (GUI)
thought unique. The iris patterns are obtained through a
video-based image acquisition system. Iris scanning devices Face
have been used in personal authentication applications for Database
several years. Systems based on iris recognition have INPUT IMAGE DATABASE IMAGE
substantially decreased in price and this trend is expected to
continue. The technology works well in both verification and
identification modes.

MATCH

If it matches Stored in
NO RESULT
Attendance Attendance
is marked Database

Figure 5. Iris Recognition YES

DNA Recognition
Deoxyribonucleic acid (DNA) is the genetic material End
found in most organisms, including humans. DNA serves as a
genetic code that is unique to every organism, no two being
exactly alike; only identical twins are an exact DNA match.
In the case of human beings, there are about 3 million bases, Figure 6. Schematic diagram of the proposed method
99% of which are the same from person to person. The Proposed Method
Attendance Management System is software developed
variations found in the final 1% are how DNA becomes
for daily student attendance in schools, colleges, and
unique to each individual. The final 1% also serves as the
institutes. If facilitates to access the attendance information of
foundation for DNA biometrics, being the location of the
a particular student in a particular class.
53902 T.Tirupal et al./ Elixir Comp. Engg. 137 (2019) 53899-53904
The information is sorted by the operators, which will be median filter. Then the faces are detected using the detection
provided by the teacher for a particular class. This system algorithm. Lastly, the detected face is recognized by
will also help in evaluating the attendance eligibility criteria comparing the faces in the database and the attendance is
of a student. The proposed method for effective attendance marked. A Graphical User Interface (GUI) is designed for
management using multiple face recognition is shown in calculating the percentage of the match of faces.
figure 6 and its explanation is given below. For each step, an
independent image is acquired that is shown in figure 7.
Image Acquisition
Image is acquired from a high definition camera that is
connected above the blackboard.
Histogram Normalization
The captured images sometimes have brightness or
darkness in it which should be removed for good results.
First, the RGB image is converted to the grayscale image for
enhancement.
Noise Filtering (a) (b)
Many sources of noise may exist in the input image when
captured from the camera. There are many techniques for
noise removal. Low pass filtering in the frequency domain
may be a good choice but this also removes some important
information in the image. In our system median filtering is
used for noise removal in the histogram normalized image.
Skin Classification
This is used to increase the efficiency of the face
detection algorithm. Voila and Jones is used for detection and
its accuracy can be increased if the skin is classified before
the scanning procedure of faces.
Face Detection (c) (d)
In this section, faces are detected and shown in figure 7
by marking circles on the faces of students.
Region of Interest:
After the face detection step, the next is face recognition.
This can be achieved by cropping the first detected face from
the image and compare it with the database image. This is
called the selection of a region of interest. In this way faces of
students are verified one by one with the face database using
the Eigen Face method.
Graphical User Interface (GUI)
In this proposed method, we use the Graphical User (e)
Interface (GUI) for accurate verification of detected faces. Figure 6. Output images for the steps in proposed method
Here, the detected face images are compared with database a) Input Image b) Gray-scale Image c) Skin Classification
images and matched for recognition. If both the faces are d) Face Detection e) Region of Interest
matched then attendance is marked to that detected face
otherwise it selects other image and the process is repeated.
Experimental Results
Figure 8 represents the Graphic User Interface for
recognition of captured images with matched and unmatched
tabs. If the input image is matched with the already existing
image in the database then the image is said to be recognized E.V.N.Sai
P.Sri Hari
and attendance will be given to that particular face. If the Praneeth K.Sai Kiran
16AT5A0420
image is not matched with the database image then it is 15AT1A04G1 16AT5A0417
Attendance
displayed as unmatched and attendance will not be given to Attendance Attendance Marked
Marked
that face image. Marked Press any key to
Press any key to
Conclusion Press any key to continue
continue
This paper presents an efficient and accurate method of continue
attendance management in the classroom environment that
can replace the old manual methods. This method is secure
enough, reliable and available for use. There is a need to use
some algorithms that can recognize the faces in a veil to
improve the system performance. This proposed algorithm
works with images of multiple faces. The paper is described No matches found No matches found
in the following steps: Firstly, the images are captured by Press any key to continue Press any key to continue
using a camera placed on the top of the blackboard. The Figure 7. Different images stored in database.
captured images are processed for noise removal using a
53903 T.Tirupal et al./ Elixir Comp. Engg. 137 (2019) 53899-53904
Based Face Recognition," 2018 International Conference
on Biometrics (ICB), Gold Coast, QLD, 2018,
pp. 124 -131. doi: 10.1109/ICB2018.2018.00029
[12] T. Tirupal and B. Chandra Mohan, ―Pixel-Level
Multifocus Image Fusion based on Wavelet Transform &
Principal Component Analysis,‖ Journal of Innovation in
Electronics & Communication, vol. 2(2), pp. 60–64, 2012.
[13] E. Jagadeesh and T. Tirupal, ―A Multimodal Image
Fusion based on Dual Tree Complex 2D-Discrete Wavelet
(a) (b)
Transform Method,‖ International Journal for Technological
Figure 8. Graphic User Interface for recognition of
Research in Engineering, vol. 2(12), pp. 3151–3156, Aug-
captured images a) Matched Recognition b) Unmatched
2015.
Recognition.
[14] E. Gopi Krishna and T. Tirupal, ―Performance of Image
References
Fusion Techniques for Satellite Images,‖ International
[1] N.I. Zainal , K.A. Sidek, T.S. Gunawan, and H.M.M. Journal for Technological Research in Engineering, vol.
Kartiwi , ―Design and development of portable classroom 2(12), pp. 3184–3188, Aug-2015.
attendance system based on Arduino and fingerprint
[15] T. Tirupal, B. Chandra Mohan and S. Srinivas Kumar,
Biometric‖, IEEE International conference on information
―On the Use of Undecimated Discrete Wavelet Transform for
and communication Technology for the Muslim world, Nov.
Medical Image Fusion,‖ National Conference on ‗Present
17-18, 2014.
Scenario and Future Trends in Biomedical Engineering and
[2] M.P. Potadar, V.V. Marathe , A.S. Khose, and L.A. Healthcare Technologies‘, FTBH-2014, IIT BHU, Varanasi,
Kotkar, ―Biometric Attendance Recording and 17-18 October 2014, pp. 5-6.
Communication System‖, International journal of innovations
[16] T. Tirupal, B. Chandra Mohan and S. Srinivas Kumar,
in engineering and technology(IJIET) , Vol. 5, Issue 2, 2015,
―Image Fusion of Natural, Satellite, and Medical Images
pp. 230-234.
using Undecimated Discrete Wavelet Transform and Contrast
[3] Wang and Jingli , ―The Design of Teaching Management Visibility,‖ IEEE National Conference on Recent Advances
System in Universities Based on Biometrics Identification in Electronics & Computer Engineering (RAECE), IIT
and the Internet of Things Technology‖, IEEE 10th Roorkee, India, pp. 11–16, Feb 2015,
International Conference on Computer Science & Education doi:10.1109/RACE.2015.7510217.
(ICCSE), Cambridge University, UK July 22-24, 2015, pp.
[17] T. Tirupal, B. Chandra Mohan and S. Srinivas Kumar,
979-982.
―Multifocus Medical Image Fusion based on Fractional
[4] M.B. Srinidhi and R Roy , ―A Web Enabled Secured Lower Order Moments and Modified Spatial Frequency,‖
System for Attendance Monitoring and Real Time Location International Conference on Advances in Biotechnology,
Tracking Using Biometric and Radio Frequency BIOTECH-2015, IIT Kanpur, India, 13-15 March 2015,
Identification (RFID) Technology‖, IEEE International pp.145-154.
Conference on Computer Communication and Informatics
[18] T. Tirupal, B. Chandra Mohan and S. Srinivas Kumar,
(ICCCI), Coimbatore, India, Jan. 08- 10, 2015
―Multimodal Medical Image Fusion based on Yager‘s
[5] A. Jain, L. Hong , S. Pankanti, and R. Bolle , ―An Intuitionistic Fuzzy Sets,‖ Iranian Journal of Fuzzy Systems,
Identity Authentication System Using Fingerprints‖, vol. 16(1), 2019, pp. 33–48, doi:10.22111/IJFS.2019.4482.
Proceedings of the IEEE, Vol. 85, Issue 9, 1997, pp. 1365-
[19] T. Tirupal, B. Chandra Mohan and S. Srinivas Kumar,
1388.
―Multimodal Medical Image Fusion based on Sugeno‘s
[6] Syed Mohsin Saif et al ., Effectiveness Of Gaussian And Intuitionistic Fuzzy Sets,‖ ETRI Journal, vol. 39(2), pp.173–
Average Noise Reduction Filters On Ideal Fingerprint Image 180, 2017, doi.org/10.4218/etrij.17.0116.0568.
In Biometric Fingerprint Identification System. International
[20] T. Tirupal, B. Chandra Mohan, S. Srinivas Kumar,
Journal of Recent Scientific Research Vol. 6, Issue, 8,
―Multimodal Medical Image Fusion based on Fuzzy sets with
pp.5731-5736, August, 2015
Orthogonal Teaching–Learning-Based Optimization,‖ In:
[7] Erik Learned-Miller, Gary Huang, Aruni Verma N., Ghosh A. (eds) Computational Intelligence:
RoyChowdhury, Haoxiang Li, and Gang Hua, "Labeled Faces Theories, Applications and Future Directions – Volume II.
in the Wild: A Survey," in Advances in Face Detection and Advances in Intelligent Systems and Computing, vol 799,
Facial Image Analysis‖, Editted by Michal Kawulok, M. october 2018, Springer, Singapore, doi.org/10.1007/978-981-
Emre Celebi, Bogdan Smolka, Springer, 2015. 13-1135-2_37.
[8] Viola, P. and M. J. Jones (2004). "Robust Real-Time Face
[21] Y. Li, S. Gong, and H. Liddell. Support vector regression
Detection." International Journal of Computer Vision 57(2):
and classification based multi-view face detection and
137-154.
recognition. In IEEE International Conference on Automatic
and Statistical Face Detection." IEEE Transactions On Face and Gesture Recognition, March 2000.
Pattern Analysis And Machine Intelligence: 26 (9) 1112-
[22] Viola, P. and M. Jones (2001). Rapid Object Detection
1123.
Using a Boosted Cascade of Simple Features, in: Proc. IEEE
[10] Curran, K., X. Li, et al. (2005). "Neural network face Conf. Computer Vision and Pattern Recognition.
detection." Imaging Science Journal, the 53(2): 105-115.
[23] Anil K. Jain, Arun Ross and Salil Prabhakar,‖ An
[11] J. Kittler, P. Koppen, et. al, "Conformal Mapping of introduction to biometric recognition‖, Circuits and Systems
a 3D Face Representation onto a 2D Image for CNN
53904 T.Tirupal et al./ Elixir Comp. Engg. 137 (2019) 53899-53904
for Video Technology, IEEE Transactions on Volume 14, [28] M. N. S. Zainudin et al., ―Face Recognition using
Issue 1, Jan. 2004 Page(s):4 – 20. Principle Component Analysis (PCA) and Linear
[24] Taiping Zhang, Yuan Yan Tang, Bin Fang, Zhaowei Discriminant Analysis (LDA),‖ no. 05, 2012.
Shang, Xiaoyu Liu. ―Face Recognition Under Varying [29] Payal Maken, "An Elementary Study on Various
Illumination Using Gradient faces‖. In Image Processing, Techniques Involved in Face Recognition Systems: A
IEEE Transactions, Nov. 2009. Review", International Journal of Scientific Research in
[25] D. Saini and R. Kaur, ―Assembling of Human Beings in Computer Science, Engineering and Information Technology,
an Image to Detect a Group on the basis of Distance using Vol. 5(1), pp. 519-525, 2019.
Digital Image Processing,‖ International Journal of Current [30] Z. Lu, X. Jiang and A. Kot, "Enhance deep learning
Engineering and Technology, vol. 6, no. 2, pp. 462–466, performance in face recognition," 2017 2nd International
2016. Conference on Image, Vision and Computing (ICIVC),
[26] H. M. Z. Hewa Majeed Zangana, ―A New Algorithm for Chengdu, 2017, pp. 244-248.
Human Face Detection Using Skin Color Tone,‖ IOSR doi: 10.1109/ICIVC.2017.7984554
Journal of Computer Engineering, vol. 11, no. 6, pp. 31–38, [31] M. Arsenovic, S. Sladojevic, A. Anderla and D.
2013. Stefanovic, ―FaceTime — Deep learning based face
[27] A. Amine, S. Ghouzali, and M. Rziza, ―Face Detection in recognition attendance system,‖ 2017 IEEE 15th International
Still Color Images Using Skin Color Information,‖ Network, Symposium on Intelligent Systems and Informatics, Subotica,
vol. 24, no. July 2014, p. 4, 2002. 2017, pp. 53-58.

View publication stats