Page 865
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
Haar Cascade Classifier-Based System for Student Attendance Through
Face Recognition
Theophilus Bamise Ajala, Sixtus Chimezie Ukaigwe , Chukwudi Anthony Udemba, Rashid Kehinde
Oloko, Abiodun Richard Agboola
Department of Computer Science, Caleb University, Imota, Lagos, Nigeria
DOI:
https://doi.org/10.51583/IJLTEMAS.2026.150500072
Received: 07 May 2026; Accepted: 12 May 2026; Published: 02 June 2026
ABSTRACT
Attendance systems are essential in academic and organizational settings to track individual participation and
discourage absenteeism. Traditional methods, such as paper-based attendance sheets, are tedious and susceptible
to fraud, as well as impersonation and having someone else stand in for you. To resolve these problems, this
study developed a haar cascade classifier-based system for student attendance through face recognition using
Python and a webcam to detect and identify students in real-time. The web application applies the Haar Cascade
Classifier from the OpenCV library, a machine learning-based algorithm that detects facial features. After
students are enrolled, the system captures facial images, extracts features, and stores them in a database. During
lectures, the system matches real-time images with the database, automatically recording attendance if a match
is found. Performance evaluation showed that both enrollment and attendance processes were completed in
under one minute, offering significant efficiency improvements over manual methods while usability testing
with ten participants confirmed high satisfaction ratings averaging above 4.0 on a five-point Likert scale across
navigation, clarity, form layout, and overall experience. This system effectively eliminates impersonation,
reduces lecturer workload, and encourages punctuality, thereby contributing to improved academic integrity and
institutional efficiency.
Keywords: Face Recognition, Attendance System, Haar Cascade Classifier, Web camera, Real-time Detection
How to cite: Ajala, T.B., Ukaigwe, S.C., Udemba, C.A., Oloko, R.K., & Agboola, A.R. (2026). Haar Cascade
Classifier-Based System for Student Attendance Through Face Recognition.
INTRODUCTION
Organizations, the business, and education have always required an attendance system. In order to distinguish
between participants and non-participants, attendance systems can be used to track or record people's attendance.
This evolution has led to different approaches, such as manual, timesheet, mechanical, access control, and
biometric attendance systems, due to technological developments and growing demand (Lun, 2023). The
management of people's attendance at work in order to minimize losses caused by employees not attending work
is called attendance management. Attendance management is different from attendance control as time clock
and timesheets have traditionally been used to track attendance, but attendance management strives to foster an
atmosphere where attendance is maximized (Akinduyite et al., 2013).
In order to track employee attendance whether it be automated or manual, all businesses require an attendance
management system. Student attendance on a daily basis is necessary for quality control purposes. In most cases,
most businesses utilize inefficient methods of manually calling out names or having employees sign pieces of
paper (Karunakar et al., 2020). Nevertheless, in most cases, most automatic human recognition systems use
conventional methods such as identity verification cards, passwords, and fingerprints. All these methods,
however, have a few drawbacks such as loss of a verification card or forgotten password. Therefore, using a
sophisticated face recognition system is the best way to guarantee complete confidentiality and preserve
historical information (Bhattacharya et al., 2018). It is a swiftly expanding field that is crucial to security since
it is an extremely accurate method of identifying and verifying individuals (Alhanaee et al., 2021).
Page 866
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
Additionally, attendance is a crucial criterion that is utilized for a variety of functions in numerous academic
institutions and businesses. These goals include maintaining records, evaluating students, and encouraging ideal
and regular attendance in class. Most schools in developing nations demand a minimum percentage of students
to attend class, however this policy has not been followed due to a number of issues with the current attendance
system. This conventional approach uses books or sheets of paper to record student attendance. The attendance
sheet could get lost or stolen. Impersonation can easily be done in this way. The process of taking attendance
involves much time, and it may be difficult to ascertain the number of students who have attained the necessary
percentage and are hence qualified to sit for the examination. A process that eliminates all these difficulties is
hence required (Akinduyite et al., 2013).
The biometric system is an authentication method that uses a person's distinct physiological or behavioral traits
to automatically identify them. To put it another way, it is an automatic identification of an individual based on
his behavioral or physiological characteristics. Inherited characteristics that emerge throughout the early stages
of human development are known as physiological features. Individuals' hands, faces, fingerprints, iris, and
retinas are a few unique physiological traits that can be measured. Handwriting, vocal patterns, and keystroke
dynamics are a few unique behavioral characteristics that can be measured (Omoyiola, 2018).
The study of face recognition is concerned with the manner in which biological systems recognize faces and the
potential for computers to imitate this process. Various types of optical sensing mechanisms, or eyes, exist in
biological systems. Optical sensing devices have evolved to suit their specific environment. In a similar vein,
various visual devices are used by computer systems to collect and process faces according to the specific needs
of each application. These sensors include 3D scans, infrared cameras, and video cameras (such as camcorders)
(Martinez, 2023).
Figure 1 processes in face recognition system (Omoyiola, 2018)
One of the most effective uses of image analysis is face recognition. It is used to incorporate facial recognition
data into electronic passports worldwide. In addition, it serves as a natural user interface in consumer
electronics, entertainment, security, and law enforcement. By identifying the user and his identity and offering
personalized services for consumer electronics, it improves the user experience (Omoyiola, 2018).
Related Work
This section gives a basic overview of some existing studies that was conducted relating to the current study by
other researchers:
Biometric Process
Biometric identification refers to unique traits of a person's physique or conduct that can be leveraged for
electronic authentication and granting access to devices, systems, or information. Biometrics involves analyzing
unique traits of an individual using measurement and statistics. The basic usage of the system are as follows but
not limited to restriction of access, identifying and recognizing people under monitored environment (Hamaamin
et al., 2024). Registration or enlistment, live data acquisition, extracting and comparing patterns on template
represent the techniques and operations that biometric systems depend on.
The training phase identified as enlistment or registration mode aims to collect data relating to individual
identities.To give assurance that the system can withstand changes in the data for a period of time, more than
one data might be carried out. In this stage, the sensor of the biometric system registers the traits of an individual,
Page 867
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
which are consquently illustrated as signatures and stored in the database. Given that the enlistment processing
is performed offline, hence there is indefinite duration (Guennouni et al., 2019).
The "one-to-one" verification or authentication process involves identifying the identity of the user by matching
his or her biometric data against the biometric template stored in the database. In this case, the system has to
answer the request of identity verification of the user. Presently, the methods being used include the use of smart
cards, a user ID or PIN number (Guennouni et al., 2019).
Identification mode involves “one-to-N” matching where the individual's identity can be established through the
comparison with one of the models in the database. There might be no individual in the database. As per
Guennouni et al. (2019), this approach involves the association of an identity with an individual.
The other importance of sample enrollment is to develop numerical templates that will act as benchmarks for
future comparisons, and also to collect biometric data. Should there arise the need for developing a new or
revised comparison algorithm, new replacement templates can be developed from the archives containing the
raw samples. Efficient and effective methods of sample enrollment are necessary to ensure that samples are
consistent, hence improving performance during matches of biometrics using one-to-many search (Aware,
2024).
Figure 2 Biometric Process (Aware, 2024)
Types of biometric Identification
It is anticipated that systems will become more reliable and capable of delivering good results in difficult
situations and in the face of counterfeiting as the use of biometric data increases. For any system to operate
effectively, information security must be guaranteed. Therefore, a strict procedure is required to verify each
Page 868
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
person's identification before allowing them to access the stored data (Hamaamin et al., 2024).
Figure 3 Types of Biometrics System (Hamaamin et. al., 2024)
Face recognition
The examination of facial characteristics and the relative positions of the lips, nose, and eyes is the foundation
of face recognition. Thermal and standard video There are now two types of imaging methods for capturing
facial images. Thermal imaging uses a heated wire created by the blood in the facial capillaries to create
photographs of faces, while regular video is recorded by a camera (Feng, 2018).
The topic of facial recognition systems was recommended by Umalkar et al. (2023). Authors of this research
study present the process of development and evaluation of an attendance system that uses facial recognition
and uses real-time data for effective and accurate attendance management. This system incorporates advanced
computer vision techniques and artificial intelligence tools to recognize employees/students in
workspaces/classrooms.
The Attendance Management System Using Face Recognition proposed by Soundarya et al. (2021) seems to be
a more appropriate model for handling the attendance of students not only in class but also in other places.
A paper written by Waghunde et al. (2021) shows an example of how automation can take place using the Class
Biometric Register System (OBCARS). The objective of such a device would be to address the issue of papers
that have been damaged or misplaced within various institutions of higher education.
An attendance management system using real-time face recognition technique is implemented in the project
proposed by Gayatri & Yalla (2023). Computer vision techniques have been employed in the project to
accurately identify and track individuals for attendance purposes. Face detection, feature extraction, and
matching are among the face recognition techniques utilized in the system.
Kumara et al. (2021) suggested implementing the use of a MATLAB program along with an RFID card in
managing face recognition for student attendance. Student attendance data is collected manually or through an
RFID card in the research. This problem is hoped to be addressed through the implementation of a lecturer's
laptop to collect attendance using face recognition software. This project will require MATLAB and a camera
in recognizing faces.
A number of studies have attempted to investigate the application of facial recognition technology in managing
Page 869
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
attendance; yet, significant shortcomings have been observed in terms of the use of technologies, ease of
application, portability, and usability. In this study, the development of a more interactive, portable, and user-
friendly attendance system has been attempted using the Python programming language, along with OpenCV
Haar Cascade Classifier, Streamlit, and SQLite3. There has also been no attempt made at Caleb University. The
current study will make up for this deficiency through the use of a system with a haar cascade classifier that
enables face detection to keep track of student attendance using a webcam on a lecturer’s computer and a
minimalistic SQLite3 database, making the whole process efficient without necessitating costly biometric
equipment. The current study will address this issue by opting for the Haar Cascade Classifier provided by the
OpenCV library, which detects faces in a shorter time with lower computing demands.
Statement of the Problem
Educational institutions place a high emphasis on student attendance, and Caleb University is no exception in
adhering to this principle. It is mandatory for all students to achieve a minimum attendance rate of 75% in each
class or lecture in order to qualify for examination participation. The integration of biometric technology has
demonstrated increased efficiency in tracking attendance. Making sure students are present is crucial to their
success. Each educational institution has established a number of policies about absence as well as specific
absenteeism limitations for this reason. In addition to the instructional activities, the teaching staff is inherently
burdened by the enforcement of these regulations. Attendance records are utilized in schools and other
educational settings for this purpose. The attendance sheet is either distributed around the classroom for students
to sign, or the instructor reads the names of the students from the attendance list and places a sign for the present
student (Temiz, 2022). On occasion, a different student may falsify a signature on the right student's behalf. In
fact, one of the issues teachers deal with is finding more signatures than there are kids in the class. Instructors
must verify the number of students participating with the number of signatures in order to prevent such issues.
Someone else may attend class on behalf of a student enrolled in the course, even if the number of signatures is
determined to be correct (Temiz, 2022). To ascertain whether someone unrelated to the course is participating,
identities should be verified. As is evident, the entire class must pass an identification check to prevent fraudulent
participation. Course time, which is extremely valuable for teachers and the classroom, is wasted as a result. We
need a way to automate the student attendance system, hence this project proposed facial recognition for
attendance, this system will serve as a rapid way to capture and record student attendance for the lecturer to
make use of, thereby reducing the stress of taking attendance manually. It reduces absenteeism among the
students in the class, when the students know that their friends cannot assist them in marking the attendance by
proxy or in their absence. The proposed system will aid in regulating records and provide an accurate list of
students who attend class.
Objectives of the Study
The explicit aim of the study is to develop a haar cascade classifier-based system for student attendance
through face recognition. The particular objectives are:
i. design a user interface to capture student attendance via face recognition,
ii. implement (i) using Python, OpenCV, Streamlit, and SQLite3.
iii. evaluate the implemented application on a couple of performance metrics.
MATERIALS AND METHODS
This study adopted a prototype approach, which involves designing the dashboard for the web application,
this represents the graphical user interface (GUI) to be used by the users of the system while the backend
also created using SQLite to store the students’ faces that are registered on the system. The use of Streamlit
(a python framework) enabled the programming of an interactive and cost-effective attendance management
interface that can operate directly through a lecturer’s laptop webcam, which is suitable for academic
environments. The advantage of using Streamlit is that it reduces the front-end development by using CSS
bootstrap or in-line or outline CSS with Javascript. The primary advantage of using SQLite is its serverless
and zero-configuration architecture, meaning it requires less setup and is a self-contained library that embeds
directly into an application, storing the entire database as a single file on disk.
Page 870
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
Figure 4: Architecture of the proposed system (Image Capturing & Recognition
According to Symanovich (2022), below are the basic step of facial recognition work:
1. Face detection: A camera takes a picture of the face. The person will be captured in profile view or facing
straight ahead in the final shot.
2. Face analysis: The Haar Cascade technique is used to process a taken image of a face. After that, the computer
looks at the person's facial geometry to find the salient features that set their face apart in the database.
3. Converting the image to data: The face capture procedure converts the analog data into a collection of distinct
digital data. A facial signature is created using the data gathered via facial recognition.
Page 871
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
4. Finding a match: A database with other registered faces is compared to the unique numerical code produced
from facial attributes. If a facial signature matches a picture in a facial recognition database system, the student's
attendance will be marked or their face will not be recognized.
Haar Cascade Classifier
Every stage of the Haar cascaded classifier comprises several weak classifiers. The method of boosting is
employed to train the weak classifiers so that the average of predictions made by all the weak classifiers is used
to generate a very accurate classifier. The classifier will then decide whether to continue with another window
(negative) or report the presence of the object (positive) based on the prediction made. Since most windows do
not contain anything interesting, stages are designed in such a way that they eliminate negative examples as fast
as possible (Mittal, 2020).
Sequential Diagram
UML sequence diagrams (also known as interaction diagrams) illustrate how the actions are performed and how
the objects interact within a specific collaboration. The main focus of sequence diagrams is time, and the
interaction is illustrated through the vertical axis of the diagram representing time on what messages are being
sent and at what times (Visual-paradigm, 2024). Sequence diagrams mainly depict object interactions in their
specific sequence of occurrences (Bell, 2022).
Figure 5: Sequential Diagram of the proposed system
The general structure and functioning of the suggested system are depicted in the image. All of the system's parts
and their interactions are depicted in this figure. The student will need to scan their face to take the attendance.
The lecturer will monitor the attendance as the student sits in front of the system camera to capture their faces.
PERFORMANCE METRICS
To evaluate the performance of the models, we used the following metrics:
RESPONSE TIME
By applying Haar Cascade Classifier on the face recognition-based attendance system, the user’s face
was easily captured and recognized by the system during the marking phase.
Table 1 : Model Evaluation
Haar Cascade Classifier
Time taken during enrollment phase
<=20 seconds
Time taken to mark attendance
<=30 seconds
From the table above, it takes less than one (1) minute for both enrollment and attendance to be processed and
completed. This is significant because it will eliminate impersonation carried out among the students, it reduces
the time taken to manually write attendance in the class during lecture(s).
USABILITY TEST
BY APPLYING FIVE-POINT LIKERY SCALE, WE WERE ABLE TO SCORE THE USABILITY TEST FOR
THE WEB APPLICATION.
TABLE 2: 5-POINT LIKERT SCALE QUESTIONNAIRE
ID
QUESTIONS
SCALE
Q1
THE WEB APPLICATION WILL SPEED UP MY
5
4
3
2
1
Page 872
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
ATTENDANCE PROCESS
Q2
THE MULTIPAGE OPTIONS ENABLE ME TO
NAVIGATE THE APPLICATION EASILY
5
4
3
2
1
Q3
THE WAY THE FRONTEND INTERFACE IS
DESIGNED IS EASY TO COMPREHEND
5
4
3
2
1
Performance Evaluation Showed That Both Enrollment and Attendance Processes were Completed in Under one
Minute, Offering Significant Efficiency Improvements Over Manual Methods While Usability Testing With Ten
Participants Confirmed High Satisfaction Ratings Averaging Above 4.0 On A Five-Point Likert Scale Across
Navigation, Clarity, Form Layout, and Overall Experience.
RESULT AND DISCUSSION
The user interface is easy for use. The user loads the application on the web browser, the user uses the text and
button to navigate around the interface. In the application page, the user enters input values for enrollment to
take place. Thereafter, the output is generated as attendance is marked after clicking the take attendance button.
Figure 6: Screenshot of Admin Login Page
Figure 7: Screenshot of Enrollment Page
The screenshot above shows the enrollemnt page of the system, on the right is the textbox and button which the
user will use to enroll himself/herself, the user enter his full name and ID number, then clicks the add new user
button, a python file is generated at the taskbar, the user opens the python file and the student’s image is taken
via the web camera of the machine and recorded in the system.
Page 873
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
Figure 8: Screenshot of Marking Attendance Page
Figure 9: Screenshot of Attendance Page
The screenshot above shows the attendance page of the system, on the left is the table and button which the user
will use to mark the attendance himself/herself, the user clicks the take attendance button, the student’s record
is taken and the attendance is recorded in the system.
The present study modelled the webcam-based facial recognition approach of Kumara et al. (2021) but improved
upon it by replacing MATLAB with Python-based open-source technologies and eliminating the dependence
on RFID cards. The use of Streamlit also enabled the development of an interactive and cost-effective attendance
management interface that can operate directly through a lecturer’s laptop webcam, which is suitable for
academic environments.
CONCLUSION
This project work is of benefit to eliminate the possibility of human error in marking attendance, ensuring that
the process is both precise and quick. This reduces time spent on manual attendance-taking and prevents
fraudulent activities, such as proxy attendance. The automated nature of facial recognition reduces the need for
Page 874
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
manual intervention. Students just have to be present in front of a camera, and the system automatically records
their attendance.
The contribution to knowledge is that the project shows how Haar Cascade Classifier Algorithm is used in face
recognition for student attendance system and has contributed to the body of research on haar cascade classifier-
based system for student attendance through face recognition. The current study utilized Python programming
language together with OpenCV Haar Cascade Classifier, Streamlit, and SQLite3 to create a more interactive,
lightweight, and user-friendly web-based attendance platform.
REFERENCES
1. Akinduyite, C.O., Adetunmbi, A.O., Olabode, O.O., and Ibidunmoye, E.O. (2013). Fingerprint-Based
Attendance Management System. DOI:10.12691/jcsa-1-5-4.
2. Alhanaee, K., Alhammadi, M., Almenhali, N., and Shatnawi, M. (2021). Face Recognition Smart
Attendance System using Deep Transfer Learning. Available at: https://creativecommons.org/licenses/by-
nc-nd/4.0.
3. Aware, I. (2024). What are Biometrics? Biometric Processes. Available at: https://www.aware.com/what-
are-biometrics-biometric-processes/.
4. Bell, D. (2022). Explore the UML sequence diagram. Available at: https://developer.ibm.com/articles/the-
sequence-diagram/.
5. Bhattacharya, S., Nainala, G. S., Das, P., and Routray, A. (2018). Smart Attendance Monitoring System
(SAMS): A Face Recognition Based Attendance System for Classroom Environmen, in 2018 IEEE 18th
International Conference on Advanced Learning Technologies (ICALT), 2018.
6. Feng., Z. (2018). Biometric Identification Technology and Development Trend of Physiological
Characteristics. doi :10.1088/1742-6596/1060/1/012047.
7. Gayatri, P., and Yalla, V. (2023). Real Time Facial Reconigzation Based Attendance System.
8. Guennouni, S., Mansouri, A., and Ahaitouf, A. (2019). Biometric Systems and Their Applications. DOI:
http://dx.doi.org/10.5772/intechopen.84845.
9. Hamaamin, R.A., Ali, O.M.A., amd Kareem, S.W. (2024). Biometric Systems: A Comprehensive Review.
10. Karunakar, M., Sai, C. A., Chandra, K., and Kumar, K. A. (2020). Smart Attendance Monitoring System
(SAMS): A Face Recognition Based Attendance System for Classroom Environment, International Journal
for Recent Developments in Science and Technology, vol. 4, no. 5, pp. 194-201, 2020.
11. Kumara, P.M., Tahmasebi, M., and Sethu, D. (2021). Automatic Attendance Recording System Using
Facial Recognition.
12. Lun, H.W. (2023). Smart Event Attendance System Using Face Recognition. Faculty of Information and
Communication Technology (Kampar Campus), UTAR.
13. Martinez, A.M. (2023). Face Recognition, Overview. Available at:
https://link.springer.com/referenceworkentry/10.1007/978-0-387-73003-5_84.
14. Mittal, A., (2020). Haar Cascades, Explained. Available at: https://medium.com/analytics-vidhya/haar-
cascades-explained-38210e57970d.
15. Omoyiola, B.O. (2018). Overview of Biometric and Facial Recognition Techniques. IOSR Journal of
Computer Engineering (IOSR-JCE). e-ISSN: 2278-0661,p-ISSN: 2278-8727, Volume 20, Issue 4, Ver. I
(Jul - Aug 2018), PP 01-05.
16. Soundarya, S., Ashwini, P., Rucha, W., and Gaurav, K. (2021). A Review Paper on Attendance
ManagementSystem Using Face Recognition. Volume 9, Issue 11 November 2021 | ISSN: 2320-2882.
17. Umalkar, A., Manhas, S.S., Chandiwala, I., and Bhagat, N. (2023). Face Recognition Based Attendance
System Using Real Time Data. Volume 11, Issue 8 August 2023 | ISSN: 2320-2882.
18. Visual-paradigm (2024). What is Sequence Diagram.
19. Waghunde, P.P., Tayde, M.P., Sabalae, A.A., and Sawant, Y.B. (2021). Biometric Attendance System
Using Facial Recognition.
