ISSN 2278-3091

Achimba Terfa et al., International JournalVolume 11,Trends

of Advanced No.1,inJanuary
Computer-Science
February 2022
and Engineering, 11(1), January - February 2022, 26 – 33
International Journal of Advanced Trends in Computer Science and Engineering
Available Online at http://www.warse.org/IJATCSE/static/pdf/file/ijatcse061112022.pdf
https://doi.org/10.30534/ijatcse/2022/061112022

Biometric Based Authentication System for

Computer Based Assessment
Achimba Terfa1, Alaaga James2, Kwaghbee Sever3, Vanen Samuel4
1
Department of Mathematics and Computer Science, University of Mkar, Benue State, Nigeria, [email protected]
2
National Agency for Food and Drug Administration (NAFDAC), Abuja, Nigeria, [email protected]
3
Department of Mathematics and Computer Science, University of Mkar, Benue State, Nigeria, [email protected]
4
Department of Mathematics and Computer Science, University of Mkar, Benue State, Nigeria, vanensamuel2004 @gmail.com

Received Date : December 09, 2021 Accepted Date :January 11, 2022 Published Date : February 06, 2022

 acquisition, self-paced learning, easy management for large

ABSTRACT groups of students etc. These activities allow the students to
adapt their schedule and attend their preferred courses
Electronic-Learning has attracted a lot of attention in recent regardless from where they are offered [5].
years. This is evident by the increase in population shift from Over the years, the population of people that participate in
the traditional way of learning. An important aspect of electronic learning has increased in significant dimension,
Electronic Learning is the assessment of students. Since the indicating how electronic learning is replacing the
traditional way of student assessment is characterized by conventional (old or traditional) way of learning [6].
issues such as insecurity, they are replaced by Electronic One important aspect of electronic-learning is the assessment
means of assessment. Most Electronic assessment systems of students. Assessment for Learning is the process of
make use of the knowledge based and object based methods seeking and interpreting evidence for use by learners and
for authentication, which can be compromised,shared or their teachers to decide where the learners are in their
transferred, resulting into issues such as impersonation.This learning, where they need to go and how best to get there[7].
research focuses on developing a system that uses facial The learning capability of students is judged by student
biometric traits in authenticating candidtates for Electrronic assessment which is mostly through examinations. Therefore,
Examination.The research uses some mathematical models the need for examinations today in universities, schools,
based on the Principal Component Analysis(PCA) technique colleges and even companies for recruitment cannot be
for face recognition.The face is prefered for its ability to offer overemphasized.
continous authentication,which will eradicate isues such as The traditional way of assessment (paper based examinations)
impersonation. The system was tested in the department of has a lot of issues. Also, conducting traditional examinations
Mathematics and Computer Science, University of Mkar, during pandemic situations is another big challenge for most
Mkar, Nigeria and it offers a reliable platform for of the institutions [8].
authentication. In an electronic-learning environment, assessment is done
through an electronic means known as electronic assessment
Key words: E-assessment; authentication; Computer Based (e-assessment), which eradicates most of the problems with
Assessment; E-learning; examination the traditional form assessment.
Computer-Based Assessment, also known as e-assessment is
1 INTRODUCTION the use of information technology for any assessment- related
activity [3]. E-assessment is an integral and vital component
Information and Communication Technologies (ICTs) is of E-learning. It can improve the standards of student’s
growing at a rapid rate and has continued to change people’s assessment whereas the conventional way of assessment
lifestyle. The use of various ICT tools has impacted the requires more effort on the part of students and invigilators.
society by creating new ways to access, process and deliver E-assessment can be conducted for remote candidates at any
useful knowledge [1]. Information and Communication time and does not incur higher cost as compared to traditional
Technology has improved our lives and provided a new examinations, does not necessarily involve arrangement of
dimension of operation by bringing regular day-to-day tasks examination centers and doesn’t require invigilators [9].
to our doorsteps. Web Technology, an aspect of ICT is wide E-examination can be delivered through offline and online
spread in all aspect of life and has changed the mode of modes. The offline mode is delivered without the use of
communication and information for humans [2-4]. Internet, strictly operated on machine software and
The evolution of Information and Communication authentication is done on single Personal Computer (PC)
Technologies provide an improved and efficient way of while the online mode is delivered through an internal
delivering contents to remote students. Web based learning network (or the Internet) and its distributed authentication
activities offer decreased delivery costs, speed in knowledge servers to locally provide access to the question banks through
server-client interactions among computers in the network
[10-11]. E-examination is timely, has online real-time
26
Achimba Terfa et al., International Journal of Advanced Trends in Computer Science and Engineering, 11(1), January - February 2022, 26 – 33

capabilities, is flexible in scoring and has diversity in question 2 BIOMETRICS-BASED AUTHENTICATION IN

type or format. It is reliable, impartial, has storage efficiency, E-ASSESSMENT
enhanced question styles and provides safety of scripts [12].
Despite the numerous advantages offered by E-learning and The popularity of E-learning in the delivery of educational
E-assessment, there are several limitations/challenges. One of content and the unprecedented growth of electronic
the challenges of both traditional and E-assessment is examinations as means of checking learning effect on students
insecurity. Since the inception of e-learning, there have been has made it necessary to provide a more secure, faster and
security and reliability issues and these pose various threats more suitable method of user’s authentication. This can be
especially when examinations are held electronically. achieved by the introduction of biometrics technology.
The implementation of e-assessment in an e-learning Biometrics can be defined as human characteristics and trait
environment can only be successful only if trust, privacy and related metrics. Biometrics Authentication refers to the
security are secured and strengthened. One way to mitigate identification of humans by their physiological characteristics
security breach during electronic examination is to identify, (such as fingerprint, face, DNA and hand print) and
authenticate and monitor candidates during examinations [2]. behavioral characteristics (such as signature, gait, voice and
Authentication can be defined as a technique used by a typing rhythm). Some of the advantages include universality,
computer in verifying the identity of individuals who want to distinctiveness, permanence, collectability, performance,
have access to that systems’ resources. It is a method that acceptability and resistance to circumvention [12, 13].
prevents unauthorized users from accessing sensitive Several research works have been carried out in attempt to
information [13]. Authentication is widely used in electronic eradicate the deficiencies identifies with Electronic
examinations. Authenticating the students’ identity is a very examinations. A new approach of using facial recognition for
important practice because it helps to reduce academic enhanced verification of e-Exam’s candidates is presented in
malpractices and also ensures quality assurance purposes in [17]. It generates web Application Programming Interface
Education. Both the examinee and examiner should be (API) authentication, image, and video with feature extraction
authenticated in order to secure the E-assessment system. matching action. Even though the work attempts to ensure that
Authenticating the examiner is also an important aspect the candidates who take the Electronic Examination are the
because the examiner has access to the examinee registration right persons, it fails to solve the problems of Face recognition
data as well as the examinee assessment data [14]. Illumination Variation, Pose/Viewpoint Variation, Facial
There are three ways to carry out authentication: (i) Expression / Facial Occlusion. The authors of [5] proposed an
knowledge based, which requires a user to provide what he efficient client/server architecture for face recognition using
knows (PINS, Usernames, Passwords etc); (ii) object based online webcam in an internet environment. Even though the
which requires a user to show what he owns (mobile device, technique can be applied to confirm users’ presence
electronic and integrated circuit cards etc); (iii) biometrics throughout an examination, it requires complex image and
(fingerprints, faces etc) [15, 16]. Most authentication systems video processing to function. The authors of [6] developed an
make use of the knowledge based and object based methods. improved e-Learning Authentication and Monitoring System
This conventional way of authentication for electronic (ELAMS) using logistic Regression (LR) to address the
assessment, which is shown in figure 1, can be vulnerabilities of e-examination. The system combines
altered,compromised,shared or transferred, which can result Biometric authentication, surveillance (webcam) and machine
to impersonation. This poses a potential threat to institutions learning techniques to detect academic dishonesty in an
and administrators of Electronic based examination. e-learning environment, which improves authentication
accuracy but the methods used are complex.
The authors of [18] presented design that uses face
recognition for examination identity authentication. The
research involves advanced technologies, such as face image
capturing, human-machine interaction, data transmission, and
communication. The advantages of the system include low
cost and high scalability. However, the methods involved in
the design are complex.
In [19], the paper attempts to develop an automated face
detection and recognition system that can be used for
detecting impersonation of candidates in examination system.
However, the system fails to handle issues such as face
illumination variation. The authors of [7] investigated
methods used for cheating detection in online exams, through
continuous authentication and online proctors. Their system
uses fingerprint reader authenticator and eye tribe tracker in
examination session. Even though the research offers some
form of continuous authentication, the system is not
implemented over the web. In addition, the use of fingerprint
for authentication involves making contact with devices
Figure 1: A conventional e-test authentication system which might result to candidates contacting diseases. In [20],
(adopted from [12]).
27
Achimba Terfa et al., International Journal of Advanced Trends in Computer Science and Engineering, 11(1), January - February 2022, 26 – 33

the authors examined the impacts, challenges and security with the system while the registration module serves as the
lapses of the existing electronic-examination system. A new backbone for pre-registration of examination takers.
system that uses data encryption for protecting the questions
and a biometric fingerprint for authentication was developed. 3.1 Verification Module
Authentication using fingerprint occurs only at login time,
which still gives room for impersonation after the initial login The verification module determines the validity of a user in
phase. It is also possible for users to contact diseases through accessing the computer based assessment system. In this
the devices. The authors of [21] proposed a continuous phase, the facial trait of a candidate is verified using some
authentication system using a multimodal biometric sequence of activities.
framework. The proposed multimodal framework combines
face, mouse dynamics, and keystroke dynamics as the PCA technique is used to determine eigen faces. For a given D
biometric technologies for authentication. The framework has dimensional data set X:
potential for promoting security of electronic examinations,
but the reliability of the methods used for the biometrics First, the mean vector μ[d] of dimensions Dx1 is computed,
authentication has not been investigated. The study in [22] where N is the number of samples in the data set:
examined the existing authentication method for the
Computer-based test system of Air Force Institute of
Technology (AFIT), Kaduna, Nigeria and proposed the
fingerprint biometric technique for authentication. The
fingerprint used for authentication does not provide a
continuous form of authentication and might result to users Compute the DxD covariance matrix C as follows:
contacting diseases. In [12], the research focused on
improving security and control by developing a platform Subtract the mean vector μ[d] from each column of
based on fingerprint-based technology for authenticating the data matrix X resulting in matrix B of DxN
electronic based examination. The platform involves the use dimension.
mathematical models for fingerprint database, enhancement,
feature extraction and pattern matching. Fingerprint
technology does not offer continuous authentication and can
be a threat to the health of users. The research carried out by
[2] addresses the issue of accuracy by proposing an image Compute the covariance matrix C of D x D dimension:
enhancement approach that uses fingerprint in conjunction
with electronic learning environments to eradicate unethical
practices with electronic examination in a university
environment. It involves extracting minutiae points of
fingerprints using the Crossing Number (CN) Concept. The
research is limited by the fact that fingerprint technology does
not guarantee continuous authentication and is also risky The eigenvectors (ed) and eigenvalues (λd) will be extracted as
given the fact that users can contact diseases from the devices follows:
involved.
Eigenvectors: compute matrix V, formed by D
3 PROPOSED BIOMETRIC COMPUTER BASED columns vectors of D length each, which
ASSESSMENT AUTHENTICATION SYSTEM. diagonalizes the covariance matrix C.

The facial trait is preferred in this research because it can V-1 C V = ^ (4)
provide a form of continuous authentication. This will solve
the problem of impersonation as it ensures that the original V = [e1 … eD] (5)
candidate is present throughout the examination period [23].
Traits such as fingerprints provide authentication only at login Eigenvalues: the diagonal values of DxD matrix L:
time, which still gives room for impersonation after the initial
login. The use of facial trait for authentication is also less
expensive, as it doesn’t require extra devices for biometric
capture. The computer system’s camera is used to capture the
face, which offers the user a convenient form of
authentication. The facial trait also offers a contactless form
of authentication in which there is no risk of contacting
diseases in an attempt to provide biometric trait. The proposed
face biometric system for computer based assessment is
shown in Figure 2.It consists of user interface, registration,
verification, computer based assessment and database
modules. The interface module facilitates visual interaction (6)

28
Achimba Terfa et al., International Journal of Advanced Trends in Computer Science and Engineering, 11(1), January - February 2022, 26 – 33

Figure 2: Proposed facial biometric authentication system for Computer Based Assessment (CBA) (adopted from [12]).

The eigenvectors with the largest eigenvalues correspond to 3.2 Assessment Module
the directions of maximum variance. Then, the L largest
eigenvalues and their corresponding eigenvectors have to be This is the module used for assessment of candidates that are
chosen, assuming that the rest of the dimension space, D-L, is authenticated successful. This assessment is done through
irrelevant information. Some information will be lost, but the taking of examinations, where candidates are required to
smaller the eigenvalues, the smaller the lost. Once the provide solutions to the questions been displayed to them. The
components (eigenvectors) have been selected, an LxL matrix module computes the result which is used for student
A with the L eigenvectors as columns has to be formed. assessment. To ensure continuous authentication, the system
Finally, the transpose of matrix A has to be taken and camera takes the picture of candidates at some specified time
multiplied by the original dataset, having previously extracted intervals. If at any interval the picture taken does not match
its mean value. with the one stored in the database, the system detects an act
of impersonation, in which the candidate will be penalized
after some warnings. This is done to avoid the situation where
X' = At . (X - µ) (7)
a candidate will be impersonated after the initial login.
The inverse transform to get the original data back can be
3.3 Database Module
achieved by applying the following back-projection.
This is the backend for hidden and background operations. It
X'' = A.X' + µ (8) stores information about the candidates, houses the biometric
templates and provides querying capabilities.
A facial image can be projected onto N' dimensions by
computing 4 SYSTEM IMPLEMENTATION

Ω = [v1v2 ...vN']T (9) The system implementation involves the use of Php and
MySql which was used for the database design. The
Where vi = edTB. vi is the ith coordinate of the facial image in environment is characterized by Window 10 operating system
the new space [24]. on Intel dual Core T6400 processor with 2G RAM and 40GB
hard disk. The system was tested by selected students of the
The Euclidean distance is minimized to determine which face department of Mathematics and Computer Science,
class provides the good description for input face image as University of Mkar, Gboko, Nigeria. The system is discussed
follows: below.

Ck = ll(Ω−Ωk)ll (10) 4.1 Home page

This is the first page to be loaded into the browser when its
Where Ωk is a vector describing the kth face class. If Ck is less URL (Uniform Resource Locator) is typed into the address
than some predefined threshold θC, a face is classified as bar. The home page can be accessed from any page in the site.
belonging to the class k. The home page provides access to the registration,
instructions, and administrator modules. This is shown in
figure 3 below.

29
Achimba Terfa et al., International Journal of Advanced Trends in Computer Science and Engineering, 11(1), January - February 2022, 26 – 33

Figure 3: Home page

Figure 4: Registration page/Biometric Enrollment.

4.2 Registration/Biometric Enrollment 4.3 Authentication Page

For a user to be eligible for the e-exams, he/she needs to For a user to be granted access to the exam module, he needs
provide details for registration in the registration page as to provide his registration number and face biometric. If the
shown in figure 4. face provided does not match the one stored in the database
during registration, authentication fails. If the face biometric
provided matches the one stored in the database, the user is
authenticated and then he can proceed to write exams. This is
shown in figure 5 and figure 6.

30
Achimba Terfa et al., International Journal of Advanced Trends in Computer Science and Engineering, 11(1), January - February 2022, 26 – 33

Figure 5: Authentication failure because of wrong biometric

Figure 6: Successful authentication

Figure 7: Student assessment page

31
Achimba Terfa et al., International Journal of Advanced Trends in Computer Science and Engineering, 11(1), January - February 2022, 26 – 33

4.4 Assessment Page

[7] Razan B, Abdullah B, Anas F and Shehab G; “E-exam
After a successful authentication of the face biometric, the
Cheating Detection System”, International Journal of
user is then provided with the e-exam page where he/she can
Advanced Computer Science and Applications, Vol. 8, No. 4,
now take the examination. This is shown in figure 7.
2017.

5 . CONCLUSION [8] Amr J; “New Detection Cheating Method of

Online-Exams during COVID-19 Pandemic”, International
This research presents a system that makes use of facial
Journal of Computer Science and Network Security, VOL.21
biometric traits to authenticate candidates for electronic
No.4, April 2021
examinations. Although a face biometric trait is not as
accurate as fingerprint, it has a high level of acceptability and
[9] T. Ramu and T. Arivoli, “A framework of secure
provides a form of continuous authentication. With facial
biometric based online exam authentication: an alternative to
biometrics, the system camera captures the candidate’s face at
traditional exam”, International Journal of Scientific &
different intervals, to make sure that the facial biometric used
Engineering Research, Volume 4, Issue 11, November-2013
for the initial login is the face that appears throughout the
ISSN 2229-5518.
examination. This will solve the problem of impersonation in
electronic examination. The system was tested in the
[10] Mohammad AS and Mohammed I. “Challenges of online
department of Mathematics and Computer Science,
exam, performances and problems for online university
University of Mkar, Mkar, Nigeria and the users’
exam”. International Journal of Computer Science Issues.
acceptability, performances in terms of speed, accuracy, and
2013;10(1):439-443.
ease of use were established. Our further research will attempt
to make use of Multimodal biometrics for the authentication
[11] Peat M and Franklin S. “Use of online and offline
of examination takers in order to improve on system accuracy.
formative and summative assessment opportunities: Have
they had any impact on student learning?” Proceeding of
REFERENCES
ASCILITE; 2002.Available:www.ascilite.org/conferences/au
ckland02/proceedings/papers/019.pdf
[1] Fayyoumi, A and Zarrad, A (2014). “Novel Solution
Based on Face Recognition to Address Identity Theft and
[12] Iwasokun G. B., Akinyokun O. K., Akinyede R. O and
Cheating in Online Examination Systems”, Advances in
Udoh S. S; “Fingerprint-Based Authorization Platform for
Internet of Things, 4, 5-12.
Electronic-Based Examination”, Journal of Scientific
http://dx.doi.org/10.4236/ait.2014.42002.
Research & Reports 12(6): 1-10, 2016; Article
no.JSRR.29179 ISSN: 2320-0227.
[2] Mohammed B.I, Abubakar U.O, Olumide S.A and Bukola
F.B; “Design of a Fingerprint Biometric Authentication
[13] Achimba T, Alaaga J and Kwaghbee S (2021):
Technique for Electronic Examination”, International Journal
“Multi-Modal Biometrics Systems: Concepts, Strengths,
of Computer Science and Telecommunications [Volume 8,
Challenges and Solutions”, International Journal of
Issue 2, March 2017]
Advanced Trends in Computer Science and Engineering,
Volume 10, No3 May-June 2021.
[3] Achimba T (2014). “Design and Implementation of a
Computer Based Assessment System”, M.Tech dissertion,
[14] Saleh M and Hanif U, “Security Considerations and
Department of Computer Science, Federal University of
Recommendations in Computer-Based Testing”, The
Technology, Akure.
Scientific World Journal Volume 2014, Article ID 562787
[4] Alotaibi, S. (2010). “Using Biometrics Authentication via
[15] Xiaoling Z and Chenglong C; “Secure Online
Fingerprint Recognition in E-exams in E-Learning
Examination with Biometric Authentication and
Environment”. The 4th Saudi International Conference,
Blockchain-Based Framework”, Mathematical Problems in
Friday 30 and Saturday 31 July 2010, The University of
Engineering Volume 2021, Article ID 5058780.
Manchester, UK.
[16] Alexandra O, Denise W, Wayne H and Chris E.
[5] Bruno E.P and Aparecido N.M; “A video-based biometric
“e-Authentication for online assessment: A mixed-method
authentication for elearning web applications”, International
study”, British Journal of Educational Technology, Vol 50,
Conference on Enterprise Information Systems, 2009.
No 2, 2019.
[6] Adetoba, B.T, Awodele, O, Kuyoro S.O, Nwaocha O.V
[17] Samson I, Mohd F.R, Maheyzah S and Siti H.O, “A
“An Improved Authentication and Monitoring System for
Smart Approach of E-Exam Assessment Method Using Face
E-Learning Examination Using Supervised Machine Learning
Recognition to Address Identity Theft and Cheating”,
Algorithms”, International Journal of Scientific &
International Journal of Computer Science and Information
Engineering Research Volume 11, Issue 3, March-2020 ISSN
Security (IJCSIS), Vol. 14, No. 10, October 2016.)
2229-5518

32
Achimba Terfa et al., International Journal of Advanced Trends in Computer Science and Engineering, 11(1), January - February 2022, 26 – 33

[18] Li-jun Y.U and Ke-feng L.I; “Application of Face

Recognition Technology in the Exam Identity Authentication
System”, 2017 3rd International Conference on Social
Science and Management (ICSSM 2017) ISBN:
978-1-60595-445-5

[19] Anuradha.S.G, Kavya.B, Akshatha.S, Kothapalli J and

Gudipati A; “Automated Face Detection & Recognition for
Detecting Impersonation of Candidate in Examination
System”, International Journal of Scientific & Engineering
Research, Volume 7, Issue 3, March-2016 ISSN 2229-5518.

[20] Olawale A and Shafi’i M.A, “E- Exams System for

Nigerian Universities with Emphasis on Security and Result
Integrity”, International Journal of the Computer, the Internet
and Management (IJCIM) Volume 18, Number 2: ISSN
0858-7027

[21] Issa T,Youssef N,Sherif S, Bassam S, Julibio D.A,

and Paulo M.F.Q; “Ensuring Online Exam Integrity Through
Continuous Biometric Authentication”, © Springer
International Publishing AG 2017, Information Security
Practices.

[22] Evwiekpaefea, A.E and Eyinlab, V.O; “Implementing

Fingerprint Authentication in Computer-Based Tests”,
Nigerian Journal of Technology Vol. 40, No. 2, 2021, 2021,
pp. 284–291

[23] Yousef S, Imane S and Amira K, “Synchronous

Authentication with Bimodal Biometrics for e-Assessment: A
Theoretical Model”,2012 6th International Conferennce on
Sciences of Electronics, Technologies of information and
Telecommunications(SETIT),2012,pp.139-145,
doi:10.1109/SETIT.2012.6481902.

[24] Kyungnam Kim, “Face Recognition using Principle

Component Analysis”, International Conference on Computer
Vision and Pattern Recognition, pp.586-591, 1996.

33