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

net/publication/327671423

Automated Student Attendance Management System Using Face Recognition

Article · September 2018

CITATIONS READS

0 4,495

5 authors, including:

Godswill Ofualagba Ise Anderson Orobor

Federal University of Petroleum Resources Federal University of Petroleum Resources
25 PUBLICATIONS   73 CITATIONS    6 PUBLICATIONS   12 CITATIONS   

SEE PROFILE SEE PROFILE

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

FACE RECOGNITION AUTOMATED ATTENDANCE MANAGEMENT SYSTEM (FACECUBE) View project

Intelligent monitoring and control using robotic systems View project

All content following this page was uploaded by Ise Anderson Orobor on 16 September 2018.

The user has requested enhancement of the downloaded file.

 International Journal of Educational Research and Information Science
2018; 5(4): 31-37
http://www.openscienceonline.com/journal/eris

Automated Student Attendance Management

System Using Face Recognition
Ofualagba Godswill1, *, Omijie Osas1, Orobor Anderson2, Ibhadode Oseikhuemen1, Odiete Etse1
1
Electrical and Electronics Engineering Department, Federal University of Petroleum Resources, Effurun, Nigeria
2
Information and Communication Technology Department, Federal University of Petroleum Resources, Effurun, Nigeria

Email address

*
Corresponding author

To cite this article

Ofualagba Godswill, Omijie Osas, Orobor Anderson, Ibhadode Oseikhuemen, Odiete Etse. Automated Student Attendance Management
System Using Face Recognition. International Journal of Educational Research and Information Science. Vol. 5, No. 4, 2018, pp. 31-37.

Received: July 17, 2018; Accepted: August 13, 2018; Published: September 13, 2018

Abstract
Attendance management system is a necessary tool for taking attendance in any environment where attendance is critical.
However, most of the existing approach are time consuming, intrusive and it require manual work from the users. This research
is aimed at developing a less intrusive, cost effective and more efficient automated student attendance management system
using face recognition that leverages on cloud computing (CC) infrastructure called FACECUBE. FACECUBE takes
attendance by using IP camera mounted in front of a classroom, to acquire images of the entire class. It detect the faces in the
image and compares it with the enrolled faces in the database. On identification of a registered face on the acquired image
collections, the attendance register is marked as present otherwise absent. The system is developed on Open Source image
processing library hence, it is not vendor hardware nor software dependent.

Keywords
Student Attendance, Face Recognition, Face Detection, FACECUBE, Cloud Computing

to such characteristics, the face recognition technique is

1. Introduction applied widely, not only to security applications but also to
image indexing, image retrievals and natural user interfaces
Maintaining attendance is very important in all educational [3].
institutions. Every institution has its own method of taking Faces are highly challenging and dynamic objects that are
student attendance. Some institutions uses paper based employed as biometric evidence, in identity verification.
approach and others have adopted automated methods such Biometrics systems have proven to be an essential security
as fingerprint biometric techniques. However, these methods tool, in which bulk matching of enrolled people and watch
subjects students to wait in a queue which consumes time and lists is performed every day [4].
it is intrusive [1]. The importance of developing a new solution to improve
Humans often use faces to recognise individuals but the performance of face identification methods has been
advancement in computing capability over the past few highlightedand CC is one of the most promising technology
decades now enable similar recognitions automatically [2]. paradigms that can be used to achieve it [5].
Face recognition technology is one of the least intrusive Face recognition technology has also been used for both
and fastest growing biometric technology. It works by verification and identification of students in a classroom.
identification of humans using the most unique This research is aimed at developing a less intrusive, cost
characteristics of their faces. effective and more efficient automated student attendance
Face recognition has characteristics that other biometrics management system using face recognition, leveraging CC
do not have. Facial images can be captured from a distance infrastructure.
and any special action is not required for authentication. Due
32 Ofualagba Godswill et al.: Automated Student Attendance Management System Using Face Recognition

2. Related Works burden of maintaining costly facilities for enrollment and

recognition, it is suggested that organisations transfers this
Several work exist in automated attendance system [6-8]. responsibility to third-party vendors who can maintain CC
However, very few have been implemented using a less infrastructures for recognition [4]. CC presents an attractive
intrusive approach such as face recognition. In this research, option for delivering educational services in a more reliable
our focus is on face recognition and a cost effective and economic way without compromising quality [11].
architecture for its implementation such as CC. CC offers a more convenient platform for image
Face recognition based attendance management system processing through large data management, distributed
with raspberry pi 2 using Eigen faces algorithm has been parallel processing, horizontal scaling etc, hence considering
proposed [9]. In the work, a camera is placed at the entrance it ideal for FACECUBE implementation.
of the class that interface with a raspberry pi 2 module for
capturing students entering the class. The images are stored
in the raspberry pi 2. The raspberry pi 2 module is used to 4. Proposed System
achieve high speed of operation. 4.1. Research Method
An attendance management system that uses CCTV
camera fixed at the entry point of a classroom has also been To achieve the set objectives, FACECUBE components
proposed. The camera automatically captures the image of a were developed as window service and web application. The
person entry the classroom and compares the image with the system is designed to shift compute intensive task like face
face database using android enhanced smart phone. The detection and recognition from client on-premise local server
Android mobile was used because of its rich face recognition to cloud server, since image processing could be compute
capability [10]. intensive especially when large data is involved. This enable
Due to computationally intensive task involved in face client local server to be any commodity hardware when
detection and identification, a system that leverages cloud implementing FACECUBE.
resources to improve facial identification performance in FACECUBE window service component is a lightweight
large video databases was proposed [5]. application that runs on the background of the on-premise
CC architecture which utilises PCA-characterised face local server. It is responsible for listening to attendance
instances has also been used to overcome the need to provide request, starting IP cameras irrespective of their location on
high computing resource for bulk image enrolment and the Campus Area Network (CAN) and continuous acquisition
matching [4]. of images from the IP cameras mounted in front of the
classrooms. The service sends all acquired images to the
3. Cloud Framework cloud web server for storage through web service mechanism
where they are processed.
Every day, sophisticated technologies emerge which The web aspect of FACECUBE is developed to provide an
becomes difficult for institutions to keep up with the trend in intuitive and simple self-service interface for users to interact
time of limited budget [11]. With the advancement of CC, with the system. It performs all the task of image storage,
many organisations are rapidly adopting IT enabled services image processing (face enrolment, face detection, face
that are hosted by cloud service providers [4]. recognition) and attendance management. It also enables
Cloud is a large pool of easily usable and accessible users perform all administrative task.
virtualized resources (such as hardware, development tools FACECUBE image processing engine is built on an Open
and/or services). These resources can be dynamically Source image processing library. Hence, the system is not
reconfigured to adjust to a variable load (scale) and can be vendor hardware or software dependent.
accessed by several users simultaneously (multi-tenant), thus
allowing optimum resource utilization [12]. 4.2. System Description
CC is a model for enabling convenient, on-demand FACECUBE is an automated student attendance
network access to a shared pool of configurable computing management system that uses face recognition technology in
resources (e.g., networks, servers, storage, applications, and its attendance process.
services) that can be rapidly provisioned and released with FACECUBE provides online access for 3 user roles:
minimal management effort or service provider interaction Student, Instructor, and Administrator role.
[13]. Student role: Student can create personalised account,
CC allow for scalable and flexible systems to be built with enroll their faces on the system in 5 orientations using its
an easy on-demand access to virtual resources in a pay-as- self-service capability and could also register their courses as
you-go utility model [5]. well as view lecture time-table. Every student that must
In face detection and recognition, rapid image processing participate in FACECUBE attendance process must be
is essential in large-scale real time system. Computational registered.
cost of image processing is quite expensive. To minimise the
 International Journal of Educational Research and Information Science 2018; 5(4): 31-37 33

Figure 1. Student profile.

Figure 2. Student face enrolment.

Instructor role: The instructor can also create a account management, course management etc.
personalised account, setup personal time-table, initiate FACECUBE sends notification to both student and
request for attendance using his/her handheld device when in instructor ahead of a class.
class and can generate attendance list after class anytime and When a request for attendance is initiated by an instructor
anywhere. from the web platform, FACECUBE windows service takes
Administrator role: The administrator is responsible for the request details, get the classroom the request was made
setting up IP cameras in different classrooms on the system from, by using it IP address to locate and start the camera in
as well as other administrative activities including user that class to begin image acquisition.
34 Ofualagba Godswill et al.: Automated Student Attendance Management System Using Face Recognition

Figure 3. FACECUBE windows service.

The system uses IP Cameras mounted in front of a registered for the course being taught when the images where
classroom irrespective of their location to continuously captured is processed. On correct identification of a student
capture image of the entire class at set interval, throughout face, the attendance register for the course is flagged as
the period of a lecture and sends images over the internet to present for the student, otherwise absent.
the cloud server for processing. The server processes the
images by identifying/detecting the human faces contained, 4.3. System Design
extract the faces and matches them with the enrolled faces of FACECUBE is designed to run part of its components in
the students stored on the database. on-premise local server as well as on a cloud server.
During the process of face matching, only the students that

Figure 4. FACECUBE architecture.

 International Journal of Educational Research and Information Science 2018; 5(4): 31-37 35

The users interact with FACECUBE through its web the data into a lower dimensionality space [11].
application interface. Suppose G is an N2x1 vector, corresponding to an NxN
The local server connects all the IP cameras in different face image I.
classroom locations using a router. It also host the windows The idea is to represent G (F=G - mean face) into a low-
service that runs on the background when the server is dimensional space:
powered. The windows service gets or post request to the
cloud server using a web service call. Fˆ - mean = w1u1 + w2u2 +...wKuK(K<<N2)
The cloud server host the web application the windows Computation of the eigenfaces [14]
service and users interact with. The web application Step 1: obtain face images I1, I2,..., IM (training faces)
components are: Step 2: represent every image Ii as a vector Gi
Face Enrolment: The enrolment process involves creating Step 3: compute the average face vector Y:
face templates from set of images taken with a webcam. The
face template are stored in the database for face recognition. 1
Image Storage: All images acquired from the IP camera Ѱ Г
video live feed as well as student enrolled face templates are
converted to binary and stored in the database.
Step 4: subtract the mean face:
Image Enhancement: Most times, acquired images have
darkness probably as a result of poor lightning in the ɸ Г Ѱ
environment. Before these images can be used for enrolment,
detection or recognition purpose, it has to be normalized. Step 5: compute the covariance matrix C:
Normalization enhances the accuracy of face detection and
∑ ɸ ɸ = AAT
recognition. This process begins by converting the RGB
image into a gray scale. Histogram normalization is then
Where A = [F1 F2... FM]
applied for contrast enhancement.
Step 6: compute the eigenvectors ui of AAT
Face Detection: This is the ability to recognise human face
Step 6.1: consider the matrix AT A
in an image. FACECUBE uses Haar Classifier for its face
Step 6.2: compute the eigenvectors vi of AT A
detection. Haar classifier is a face recognition algorithm
trained with thousands of human faces with different face ATAvi = µivi
positions and lighting condition. FACECUBE uses this
algorithm to detect multiple faces in an image and draw a Step 6.3: compute the M best eigenvectors of AAT: ui = Avi
rectangle on each detected face. The face images are Step 7: keep only K eigenvectors (corresponding to the K
extracted and resized to 100x100. largest eigenvalues)
Each face (minus the mean) Fi in the training set can be
represented as a linearcombination of the best K
eigenvectors:
ɸ ∑ Wjuj, (wj=uTjFi)u j ’s is called
eigenfaces)
Each normalized training face Fi is represented in this
basis by a vector:

%
!% , 1,2, … . ,
…%
#$

4.5. System Implementation

Figure 5. Detected faces.
FACECUBE was implemented using.NET technology on
Face Recognition: FACECUBE uses Eigenface algorithm Visual Studio and Azure SDK. Visual Studio and the Azure
for face recognition. The algorithm is use to determine if SDK provide the capabilities of creating and deploying
enrolled face template stored in the database is found in the projects to Azure cloud infrastructure directly from the IDE.
acquired classroom image. The source code was developed using C# ASP.NET and the
face detection and recognition engine was built on EmguCV,
4.4. Face Recognition and Detection a.NET wrapper to OpenCV image processing library.
FACECUBE was tested in Intel(R) Celeron(R) CPU N3050
In face recognition, problem arise when performing
@1.60Hz computers with a 4GB RAM and windows 7
recognition in a high-dimensional space. However,
operating system.
significant improvements can be achieved by first mapping
36 Ofualagba Godswill et al.: Automated Student Attendance Management System Using Face Recognition

Figure 6. FACECUBE experimental setup.

Further test was carried out by deploying FACECUBE on attendance management system called FACECUBE.
Microsoft Azure cloud environment. Using cloud platform will FACECUBE takes attendance by using IP camera mounted in
make compute intensive task faster. This will also enable front of a classroom to acquire images of the entire class,
scalability from a single machine to clusters of servers on a detect the faces in the image and compare with the enrolled
public or private cloud infrastructure and improved performance. faces in the database. On identification of a registered student
Cloud platform could also help in commercialising FACECUBE face on the acquired image collections, the attendance
as a pay per use software service. register is marked as present otherwise absent.
The system is designed to be cost effective with no
5. Conclusion specific vendor hardware and software required for
deployment.
Attendance management system is a necessary tool for
taking attendance in any environment where attendance is References
critical. Most of the existing approaches are time consuming,
intrusive and required manual work from the users. In this [1] Balcoh N. K Yousaf M. H, Ahmad W. and Baig M. I. (2012)
research, we eliminated the above challenges by Algorithm for Efficient Attendance Management: Face
Recognition based approach. IJCSI International Journal of
demonstrating the use of face recognition in student Computer Science Issues, 9 (4): 146-150.
 International Journal of Educational Research and Information Science 2018; 5(4): 31-37 37

[2] FBI (2017) Face Recognition. Available at: [9] Gaddam S. C. 251, Ramesh N. V. and Dhanekula H. (2016)
https://.fbi.gov%2Ffile-repository%2Fabout-us-cjis- Face Recognition Based Attendance Management System
fingerprints_biometrics-biometric-center-of-excellences-face- with Raspberry Pi 2 Using Eigen Faces Algorithm. ARPN
recognition.pdf. Accessed 14 January 2017. Journal of Engineering and Applied Sciences, 11 (13): 8107-
8112.
[3] Sato A., Imaoka H., Suzuki T. and Hosoi T. (2005) Advances
in Face Detection and Recognition Technologies. NEC Journal [10] Muthu K. and VeeraA. (2013) Smart Application for AMS
of Advance Technology, 2 (1): 28-34. Using Face Recognition. Computer Science & Engineering:
An International Journal (CSEIJ), 3 (5): 13-20.
[4] Kisku R. and Rana S. (2016) Multithread Face Recognition in
Cloud. Available at: [11] Orobor A. and Orobor H. (2018) Cost-Saving and Internally
https://www.hindawi.com/journals/js/2016/2575904/.Accessed Generated Revenue in Tertiary Educational Institutions: The
14 January 2017. Role of Cloud Computing. International Journal of Education
and Information Technology, 3 (2): 7-16.
[5] Caldeira R. (2013) FaceID-Cloud: Face Recognition
Leveraging Utility and Cloud Computing. Available at: [12] Vaquero L., Rodero-Merino M., Caceres J. andLindner M.
https://fenix.tecnico.ulisboa.pt/downloadFile/395145529147/d (2008) A Break in the Clouds: Towards a Cloud Definition.
issertacao.pdf. Accessed 28 May 2017. ACM SIGCOMM Computer Communication Review, 39 (1):
50-55.
[6] Dhanalakshmi N., Kumar S. and Sai Y. (2017) Aadhaar Based
Biometric Attendance System Using Wireless Fingerprint [13] Mell P. and Grance T. (2009) The NIST Definition of Cloud
Terminals. Available at: Computing. Available at:
https://ieeexplore.ieee.org/iel7/7974917/7976734/07976871.p http://www.nist.gov/itl/cloud/upload/cloud-def-v15.pdf.
df. Accessed 6 August 2018. Accessed: 24 February 2017.
[7] Mohamed M. (2017) Class Attendance Management System [14] Turk M. and Pentland A. (1991) Eigenfaces for Face
Using NFC Mobile Devices. Journal Intelligent Automation & Detection/Recognition. Journal of Cognitive Neuroscience, 3
Soft Computing, 23 (2): 251-259. (1): 71-86.

[8] Rahim A., Ismail N., Razak F., Zulkifli Z., Jamian H., Razi F.,
and Mohammad N. (2017) Automated Attendance
Management and Alert System. Journal of Fundamental and
Applied Sciences, 9 (65): 59-80.

View publication stats