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

net/publication/339078481

An Android-based Face Recognition System for Class Attendance and

Malpractice Control

Article  in  International Journal of Computer Science and Information Security, · January 2020

CITATIONS READS

0 130

3 authors, including:

Folasade O. Isinkaye Jumoke Soyemi

Ekiti State University, Ado Ekiti The Federal Polytechnic, Ilaro
21 PUBLICATIONS   398 CITATIONS    38 PUBLICATIONS   139 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

A Scalable Intelligent Solution to Big Data problem in Recommender Systems View project

Computer in Education View project

All content following this page was uploaded by Jumoke Soyemi on 06 February 2020.

The user has requested enhancement of the downloaded file.

 International Journal of Computer Science and Information Security (IJCSIS),
Vol. 18, No. 1, January 2020

An Android-based Face Recognition System for

Class Attendance and Malpractice Control
*Folasade O. Isinkaye Jumoke Soyemi Olamide I. Arowosegbe
Department of Computer Science Department of Computer Science Department of Computer Science
Ekiti State University, Ado-Ekiti, The Federal Polytechnic, Ilaro, Ogun State, Ekiti State University, Ado-Ekiti,
Nigeria Nigeria Nigeria
[email protected] [email protected] [email protected]

Abstract—Over time, examination malpractice in form of resources. In addition, the menace called examination impersonation
students indulging in impersonation has been a serious setback to can be eliminated from the institutions to provide an atmosphere with
academic growth in several Educational institutions. free and fair examination. This in turn makes the award of degrees
Conventional methods of keeping student’s attendance records from institutions much more credible.
has proved inefficient as there are so many cases of data loss and
mismanagement. Several systems developed to solve this problem II. RELATED WORKS
are either immobile or costly to implement. In order to solve this There have been various techniques used to develop Attendance
problem of cost and rigidity as well as remove the problem of management system (AMS). Some of these techniques used Radio
examination impersonation, a mobile system running on the Frequency Identification (RFID) [7,8], fingerprint [9,10], iris [11,12],
Android Operating System was developed using the Viola-Jones palm print [13,14], voice [15,16] etc. For instance, a wireless
object detection framework and Eigen faces to carry out Facial attendance management system that used Iris identification was
Recognition of students and take record of attendance in classes proposed by [17]. The system consisted of three modules: Iris
in a user-friendly and secure manner. The facial recognition verification and identification module, Iris management module and
ability of the system was tested with 95% accuracy while that of Wireless communication module. The implementation of the system
Facial Detection ability gave an accuracy of 78%. The optimum was carried out with Daugman. Local Binary Pattern Histogram
security performance achieved by the system was based on its algorithm was used to design an application for both facial detection
strong backend and its distinct modular structure. and identification of students by [18]. The algorithm identifies face
by matching some parameters in which the algorithm was trained. An
Keywords-Impersonation; Examination Malpractice; Student’s Embedded Computer-based Lecture Attendance Management
Attendance Record; Android Operating System; Facial System based on a single chip computer was used to capture the
Recognition; Viola-Jones; Eigen Faces. lecture attendance of different students. The identity of each student
was validated through a card reader interface with a computer system
[19]. An efficient management system based biometric design was
I. INTRODUCTION proposed by [20] . The system takes attendance electronically with
Taking and monitoring student’s attendance in classes have been the aid of a fingerprint devise. The attendance stored in a database
a long-standing issue amongst lecturers and the school management was marked after the identification of students. [21] proposed a facial
as a whole because of the numerous deficiencies associated with the recognition approach that used Eigen face technique. The Eigen face
conventional paper and pen approach of taking attendance [1]. It is a recognition [22] utilized information from the raw pixel image for
very stressful method of keeping record with important data and training and classifying image identity. He suggested that the system
information getting lost easily, thus making the process inefficient can be used as the basis for the development of android applications
for monitoring student’s attendance in classes [2,3]. Furthermore, such as android mobile security application and as an archive for the
most schools tie the students’ class attendance with their examination, recognition of human identity. [23] proposed a RFID employee
using it as a criteria for admitting students into halls of examination attendance system that was incorporated into a database system. The
by determining whether the student is a member of the class meet up RFID attendance system was developed using components such as
with the number of required classes that qualifies him/her to sit for tags which was used as a replacement of ID cards and a reader device
the examination. This helps in fiercely combating examination that could read the information related to an employee attendance.
malpractice in form of impersonation since a non-student member of The system has the ability to store the information of all employee.
the class cannot be admitted to the venue to write such examination. [24] also used Radio Frequency Identification (RFID) technology for
The conventional approach has over time proven inefficient as the attendance management system developed. Their experiment was
students can easily write attendance for their absent friends, and the conducted on a sample of 60 students, enrolled in a particular course.
fact that records can be misplaced or destroyed easily [4,5]. There is Based on experiment, the total time taken to record the attendance of
also the problem of impersonation amongst students, false identities a class of 60 students by manual entry took about 10 minutes while
being used in getting into the examination hall to partake in 120 seconds was taken to take the attendance of 60 students using
examinations [6]. This has terribly affected the credibility of barcode and RFID technology.
examinations and the suitability of graduates being employed into Attendance Management System could be made smarter by using
job positions after school. facial recognition technique based on Viola–Jones and Eigen faces
This study develop a system that eliminates the issues algorithms developed on Android-based device.
encountered with the use of conventional method of keeping record
of student’s attendance in classes using facial recognition system to
identify and distinguish the genuine and fake students, hence making
examination conduct less stressful, saving valuable time and

79 https://sites.google.com/site/ijcsis/
ISSN 1947-5500
 International Journal of Computer Science and Information Security (IJCSIS),
Vol. 18, No. 1, January 2020

III. PROPOSED APPROACH

The design of Android-based face recognition system for class
attendance and examination malpractice control involved the use of
two different modules working together to meet the stated objective.
These modules are the ‘Registration Module’ and
‘Attendance/Verification Module’. These modules actively interact
with each other to query the database, which serves as a reservoir of
information and the central backend for AFRS as data storing and
retrieval activities are carried out in the database. Fig. 1 shows the
architecture of the system.

Fig. 2. Phases of Viola-Jones algorithm

B. Eigenface for Face Recognition

Eigenface algorithm is regularly employed in identifying

different features in presented facial images [28, 29, 22]. The
algorithm used Principal Component Analysis as the fundamental
idea for face identification [30]. It operates by forming an eigenface
from the specified face image. Then the Euclidian distances between
the eigenface and the earlier stored eigenfaces are computed. The
eigenface with the minimum Euclidian distance represents the one
that best identified a person. That is, a new face A is projected into a
Fig. 1. System architecture
face space by , where denotes the set of
major eigenvectors. Also, the weight vector B represents the new
A. Viola-Jones Algorithm for Detection face in face space. In order to decide which face class A fits into, the
The proposed technique employed Viola-Jones Euclidean distance is minimized as shown in equation 1.
algorithm which is used to identify faces in digital images ϵk=∥B−Bk∥ 1
[25, 26]. The algorithm has four main parts which
enhances its detection of images in real time, they include; Bk denotes the weight vector that represents the kth face class. The
face A is assumed as belonging to class k if the minimum ϵk is
Haar feature selection, integral image formation,
smaller than some predefined threshold θϵ, otherwise, it is
AdaBoost training, and cascade classifiers [27] as shown categorized as not known.
in Fig. 2. Haar feature selection was used to determine if a Therefore, to effectively mark a student’s attendance, the input
face presented in an image was existing or not. The image must be run through this component which in turn compares
essence of integral image formation phase was to change the facial features (length of nose, depth of eye sockets, shape of
the original image into integral image which helps to cheekbones, etc.) to the dataset of stored images in the database
speedily estimate the sum of values in a rectangle subset of which were collected in the facial detection phase. If the
a pixel grid. Also, the Viola-Jones used AdaBoost to corresponding image fails to match any of the stored images, the
choose a specific group of features to enable the training facial recognition step fails and the student’s attendance will not be
taken by the Attendance module. The attendance component of the
of the classifiers to be used in image detection process.
Attendance/Verification Module is responsible for marking student’s
Finally, the cascaded classifier speed up the removal of attendance. When the facial detection and recognition phases are
incorrect image rapidly by removing images that are not successful, the Attendance component is invoked to mark the student
classified as faces at the initial phase. When an image is as either present in the class or not. Similarly, when any of the facial
classified as a true face, it is sent to the next phase of the detection or recognition phase fail, the attendance is not marked. The
cascade which is more complex than the earlier one. A following condition are satisfied before the system can marks a
face is successfully identified if it scaled through all the student’s attendance as present:
necessary phases. 1. Student must appear in person at the lecture location to
take attendance.
2. Student must have completed the registration stage
successfully.
3. Student’s image must pass the attendance test during facial
recognition and also matches the image presented during
registration.
Students will be confirmed absent by the system until the all
conditions stated above are fulfilled. That is, by default, students are
marked absent until they pass the Facial Recognition phase whose
requirement is to complete the Registration Module correctly.

80 https://sites.google.com/site/ijcsis/
ISSN 1947-5500
 International Journal of Computer Science and Information Security (IJCSIS),
Vol. 18, No. 1, January 2020

IV. SYSTEM IMPLEMENTATION

The proposed system was built to be a very robust, modular,
scalable and secured. This was achieved using Android Studio,
JAVA (Android), SQLite Database and XML. Android Studio makes
it possible to create applications that work on the Android OS. JAVA
was employed because of its powerful libraries (Shane et al, 2010).
XML in Android development was used to create the user interface
of the application. It was used to style the application layout as well
as the various elements of the layout and to pass data around the
various elements of the Android system. SQLite is a relational
database management system that was embedded into the system to
make it versatile and easily accessible. The following are the output
of the implementation of the system.
The Home Screen is the first screen that is loaded and is by
default, the Main Activity of the system. It contains the user
interaction modules needed to navigate to other activities (screens)
within the system and can be considered generally as the entrance to
using the application. This is shown in Fig. 3 and 4. Fig. 5. The Attendance Screen

Fig. 6 and Fig. 7 constitute the entire Registration process.

Fig. 3. Home Screen Navigation Drawer

Fig. 6 Face Detection Phase

Fig. 4. Home Screen

Fig.7 List of Registered Students
Fig. 5 shows the attendance screen. The system marks the
student’s attendance and perform the general face recognition
activity. Fig. 8 is the module for viewing student’s data. The
registered students are displayed on this screen and a user can easily

81 https://sites.google.com/site/ijcsis/
ISSN 1947-5500
 International Journal of Computer Science and Information Security (IJCSIS),
Vol. 18, No. 1, January 2020

view detailed information regarding how frequent a student attends

classes.

Fig. 8 List of Registered Students

Fig. 9 Visualization graph for 50 live facial data
It is also possible for users to view every student that registered
in a class. The system can remove a student from a class based on
misconduct and also verify if a certain student is a member of a class For static general facial data:
and if they meet the minimum required number of attendance to sit TABLE 2
for tests or examinations.
RESULTS OF TESTING 20 STATIC FACIAL DATA

V. SYSTEM EVALUATION
The system was tested using facial data which included live feed
images of 50 students and 20 static general human faces and we
tested the % Success rate, %Failure rate and % of false positives of
the Facial detection and Facial recognition processes. Although for
testing static human images only the Facial detection process was
tested because it is impractical to detect the faces that are without
dynamic features such as pose, texture, and lightning as these are
static images that do not change. Sources of these static images
include; books, billboards and posters.
The result of testing the system with the aforementioned
parameters are detailed in Table 1 and Table 2, for student’s live
facial data:
TABLE 1
RESULTS OF TESTING 50 LIVE FACIAL DATA

Fig. 10 Visualization graph for 20 static facial data

From the data presented above, it is clear from the
visualization charts in figures 9 that the system handled Facial
detection activities very well with 5.3% false positives rate detected
while testing live data was as a result of inadequacies in the Viola-
Jones algorithm being used for the detection process. Also, as seen in
figure 10, we recorded a 6.7% failure rate while detecting the faces
from static sources because most of the faces in these sources were
not full frontal facing images, most of them contained images that
have only the side view of faces and the system was unable to detect
faces that were not front facing in images.
Facial Recognition activities on the other hand witnessed a major
drop in performance as opposed to Facial Detection. This is as a
result of unavailability of large sets of images for a face which would
be used to train the engine for Facial Recognition. To obtain optimal

82 https://sites.google.com/site/ijcsis/
ISSN 1947-5500
 International Journal of Computer Science and Information Security (IJCSIS),
Vol. 18, No. 1, January 2020

results and performance boost required that set(s) of training images International Conference on Electro-Technology for National
are provided for Facial Recognition. The larger the set of training Development (NIGERCON), (pp. 464-471). IEEE, 2017.
images for an image, the better the accuracy of the Recognition [14] G. Jaswal, R. Nath, and A. Kaul, “Texture based palm Print
recognition using 2-D Gabor filter and sub space approaches”, In
exercise.
2015 International Conference on Signal Processing, Computing
and Control (ISPCC), (pp. 344-349). IEEE, 2015.
VI. CONCLUSION [15] B. Soewito, F. L. Gaol, E. Simanjuntak, andF. E. Gunawan,
The implemented system proved to be a great step in the right “Smart mobile attendance system using voice recognition and
direction, as it provides a means of adequately and properly fingerprint on smartphone”, In 2016 International Seminar on
managing student’s attendance as well as controlling examination Intelligent Technology and Its Applications (ISITIA), (pp. 175-
180). IEEE, 2016.
malpractices. The system provides a secure avenue to store data and [16] M. N. Uddin, M. M. Rashid, M. G. Mostafa,M. W. Rahman, and
generally makes it easier for students and lecturers to work together S. Halder, “Development of Voice Recognition for Student
seamlessly removing the challenges faced by the conventional Attendance”, Global Journal of Human-Social Science Research:
method of paper-pen attendance taking. With the possession of a G Linguistics & Education, 16(1), 2016.
robust backend as well as a modular framework, it is easier for the [17] S.Kadry and M. Smaili, “Wireless attendance management system
system to attain its security goals and also provides optimum based on iris recognition”, Scientific Research and essays,
performance. 5(12):1428-1435, 2013.
[18] A. Pabbaraju and S. Puchakayala, “Face recognition in mobile
devices”, Electrical Engineering and Computer Science,
REFERENCES University of Michigan. 2010.
[1] K. Shrivastava, S. Manda, P. S. Chavan, T. B. Patil, and S. T. [19] O. S. Shoewu and A. L. Lawson, “Embedded Computer-Based
Sawant-Patil, “Conceptual Model for Proficient Automated Lecture Attendance Based Lecture Attendance Management
Attendance System based on Face Recognition and Gender System Management System”, African Journal of Computing &
Classification using Haar-Cascade, LBPH Algorithm along with ICT, 4(3):27-36, 2011.
LDA Model”, International Journal of Applied Engineering [20] O. Shoewu and O. A. Idowu, “Development of attendance
Research, 13(10):8075-80, 2018 management system using biometrics”, The Pacific Journal of
[2] M. S. Uddin, S. M. Allayear, N. C. Das, and F. A. Talukder, “A Science and Technology, 13(1):300-307, 2012.
location based time and attendance system”, International journal [21] M. A. Bhuiyan, “Towards face recognition using eigenface”, Int J
of computer theory and engineering, 6(1):36, 2014. Adv Comput Sci Appl. 7(5):25-31, 2016.
[3] M. A. Mohandes, “Class attendance management system using [22] P. Wagh, R. Thakare, J. Chaudhari, and S. Patil, “Attendance
NFC mobile devices”, Intelligent Automation & Soft Computing, system based on face recognition using eigen face and PCA
23(2):251-9, 2017. algorithms”, In 2015 International Conference on Green
[4] M. Sajid, R. Hussain, and M. A. Usman, “conceptual model for Computing and Internet of Things (ICGCIoT), (pp. 303-308).
automated attendance marking system using facial recognition”. In IEEE, 2015.
Ninth International Conference on Digital Information [23] G.D. Maramis and P. T. Rompas, “Radio frequency identification
Management (ICDIM 2014), (pp. 7-10). IEEE, 2014. (rfid) based employee attendance management system”, In IOP
[5] Kumar NA, Swarnalatha P, Chowdary P, Naidu JK, and Kumar Conference Series: Materials Science and Engineering, (306 (1):
KS. Smart Attendance Marking System using Facial Recognition. 012045). IOP Publishing, 2018.
Research Journal of Science and Technology, 11(2):101-108, [24] S. Nainan, R. Parekh, and T. Shah, “RFID technology based
2019. attendance management system”, International Journal of
[6] K. Puthea, R. Hartanto, and R. Hidayat, “A review paper on Computer Science, 10(1):1-6, 2013.
attendance marking system based on face recognition”, In 2017 [25] N. T. Deshpande and S. Ravishankar, “Face Detection and
2nd International conferences on Information Technology, Recognition using Viola-Jones algorithm and Fusion of PCA and
Information Systems and Electrical Engineering (ICITISEE), (pp. ANN”, Advances in Computational Sciences and Technology,
304-309). IEEE, 2017. 10(5):1173-89, 2017.
[7] A. Agrawal and A. Bansal, “Online Attendance Management [26] Y.Q. Wang, “An analysis of the Viola-Jones face detection
System Using RFID with Object Counter”, International Journal algorithm”, Image Processing on Line, 26(4):128-48, 2014.
of Information and Computation Technology, 3(3):131-138, 2013. [27] P. Viola and M. Jones, “Rapid object detection using a boosted
[8] V. J. Babu, T. K. Balaji, G. B. Gangadhara, and P. Jayanthi, “A cascade of simple features”, CVPR (1), 1(511-518):3, 2001.
Corruption less fee distribution system integrated with students [28] R. Boda, M. J. Priyadarsini, and J. Pemeena, “Face detection and
attendances in Educational Institutions”, Compusoft, 2(3):73, tracking using KLT and Viola Jones”, ARPN Journal of
2013. Engineering and Applied Sciences, 11(23):13472-13476, 2016.
[9] S. Rahman, M. Rahman, and M. M. Rahman, “Automated student [29] E. Rekha, and P. Ramaprasad, “An efficient automated attendance
attendance system using fingerprint recognition”, Edelweiss management system based on Eigen Face recognition”, In 2017
Applied Science and Technology, 1(2):90-94, 2018. 7th International Conference on Cloud Computing, Data Science
[10] G. Talaviya, R. Ramteke, and A. K. Shete, “Wireless fingerprint & Engineering-Confluence, (pp. 605-608). IEEE, 2017.
based college attendance system using Zigbee technology”, Int. J. [30] P. Singh and A. Sharma, “Face recognition using principal
Eng. Adv. Technology.(IJEAT), 2249:8958, 2013. component analysis in MATLAB”, International Journal of
[11] T. W. Hsiung and S. S. Mohamed, “Performance of iris Scientific Research in Computer Science and Engineering,
recognition using low resolution iris image for attendance 3(1):1-5, 2015.
monitoring’ In 2011 IEEE International Conference on Computer
Applications and Industrial Electronics (ICCAIE), (pp. 612-617).
IEEE, 2011.
[12] A. .Khatun, A. F. Haque, S. Ahmed, and M. M.Rahman,
“Design and implementation of iris recognition based attendance
management system”, In 2015 International Conference on
Electrical Engineering and Information Communication
Technology (ICEEICT), (pp. 1-6). IEEE, 2015.
[13] A. Charity, K. Okokpujie, and N. O. Etinosa, “A bimodal
biometrie student attendance system”, In 2017 IEEE 3rd

83 https://sites.google.com/site/ijcsis/
ISSN 1947-5500
View publication stats