Mobile-Based Intelligent Classroom Attendance System Using
Face Recognition
Maha Lakshmi M , Nandhini S, Poorana Selvi T, Praja J ,Dept of IT
Co-author : [Link] Viyanikaran AP/IT
[Link] Aditanar College Of Engineering
Abstract : Keywords :
The AI-Based Smart Classroom Face Detection, Attendance System,
Attendance System using Mobile Face Machine Learning, TensorFlow Lite, Flutter,
Detection is a mobileapplication designed Computer Vision
to automate the traditional attendance
I .Introduction :
process in classrooms. In many educational
institutions, attendance is still recorded Attendance management is an
manually, which consumes time and may important part of educational institutions,
lead to human errors. This project but traditional manual methods are time-
proposes an intelligent mobile-based consuming and prone to errors. To
solution that uses face detection overcome these issues, the AI-Based Smart
technology to identify students and mark Classroom Attendance System using
attendance automatically. Mobile Face Detection is proposed.
The system uses the camera of a This system uses a mobile camera to
mobile device to capture an image of the capture classroom images and applies face
classroom. The captured image is detection technology to identify students.
processed using face detection algorithms The detected faces are matched with a
to detect multiple student faces present in stored database, and attendance is marked
the classroom. Each detected face is automatically. It reduces manual work,
analyzed and matched with the student prevents proxy attendance, and improves
database stored in the system. accuracy.
Once the faces are recognized, The application is developed using
the application automatically records Flutter, with TensorFlow Lite for face
attendance for the identified students. The recognition and [Link] for backend
mobile application is developed using support. Overall, the system provides a
Flutter for the user interface and integrates fast, reliable, and smart solution for
machine learning models using TensorFlow modern classroom attendance
Lite for face detection and recognition. management.
[Link] Statement :
In many educational institutions,
attendance is still recorded manually,
which is time-consuming, inefficient, and
prone to human errors. It also allows issues
like proxy attendance, where students
mark attendance for others. Existing
systems such as RFID or biometric methods
require additional hardware and physical
interaction, increasing cost and complexity. The system consists of three main
There is a need for a smart, accurate, components:
and contactless attendance system that A. Mobile Application (Frontend)
can automatically identify students and
record attendance using a simple mobile Developed using Flutter
device. The system should reduce manual Provides user interface for teachers
effort, improve accuracy, prevent misuse,
and work efficiently in real-time classroom Captures classroom images
environments. B. Machine Learning Module
III. PROPOSED SYSTEM : Uses TensorFlow Lite
The proposed system is a mobile Performs face detection and recognition
application that automates attendance
Processes images locally or via API
using face detection and recognition.
C. Backend Server
Key Features:
Developed using [Link]
➢ Real-time face detection using
mobile camera Stores student database and attendance
➢ Recognition of multiple students records
simultaneously
Handles API requests and report
➢ Automatic attendance marking
generation.
➢ Attendance report generation
➢ Cloud-based data storage V. METHODOLOGY :
The system captures a classroom
image, detects faces, and compares them
with stored student data to mark
attendance instantly.
IV. SYSTEM ARCHITECTURE :
The working process of the system includes ➢ improved accuracy compared to
the following steps: manual methods
Capture classroom image using mobile The system performs efficiently
under normal lighting conditions and
camera
provides reliable results.
Detect faces using ML model
[Link] :
Extract facial features
Advantages:
Match detected faces with stored database
➢ Saves time and effort
Mark attendance for recognized students ➢ Eliminates proxy attendance
Store attendance data in backend ➢ Cost-effective (only smartphone
required)
Generate reports ➢ Easy to use
VI. TECHNOLOGIES USED : ➢ Real-time processing
Limitations:
Frontend: Flutter
➢ Performance may vary under poor
Backend: [Link]
lighting
Machine Learning: TensorFlow Lite ➢ Face occlusion (mask, angle) may
Database: MongoDB / Firebase affect accuracy
➢ Requires initial database setup
Programming Languages: Dart, JavaScript
Objectives :
VII. RESULTS AND DISCUSSION :
➢ To develop a mobile-based
The system was tested in a application for automatic attendance
classroom environment with multiple using face detection.
students. The following observations were ➢ To reduce manual effort and save
made:
time in the attendance process.
➢ To accurately identify students using
face recognition technology.
➢ To prevent proxy attendance and
improve reliability.
➢ To provide real-time attendance
marking using a smartphone camera.
➢ To store and manage attendance
data securely using a backend system.
➢ Accurate detection of multiple
faces in a single frame XI. CONCLUSION :
➢ Attendance marked within seconds
➢ Reduced manual effort significantly This paper presents an AI-based
smart attendance system that automates
classroom attendance using mobile face in Proc. Advances in Neural Information
detection. The system reduces manual Processing Systems (NIPS), 2012.
workload, improves accuracy, and provides
[5] S. Z. Li and A. K. Jain,
a scalable solution for educational
institutions. Future enhancements may "Handbook of face recognition," Springer,
include integration with cloud AI services, 2011.
improved recognition accuracy, and
[6] R. Brunelli and T. Poggio,
support for live video attendance tracking.
"Face recognition: Features versus
X. FUTURE WORK : templates," IEEE Trans. Pattern Analysis
➢ Integration with AI-based cloud and Machine Intelligence, 1993.
recognition systems [7] M. Turk and A. Pentland,
➢ Masked face recognition
➢ Real-time video attendance tracking "Eigenfaces for recognition," Journal of
➢ Integration with college ERP systems Cognitive Neuroscience, 1991.
[8] H. Kaur and S. Kaur,
[Link]
"Face recognition based attendance
[1] W. Zhao, R. Chellappa, P. J. Phillips, and
system,"
A. Rosenfeld,
International Journal of Advanced
"Face recognition: A literature survey,"
Research in Computer Science, 2018.
ACM Computing Surveys, vol. 35, no. 4, pp.
399–458, 2003.
[2] O. M. Parkhi, A. Vedaldi, and A.
Zisserman,
"Deep face recognition," in Proc. British
Machine Vision Conf. (BMVC), 2015.
[3] F. Schroff, D. Kalenichenko, and J.
Philbin,
"FaceNet: A unified embedding for face
recognition and clustering,"
in Proc. IEEE Conf. Computer Vision and
Pattern Recognition (CVPR), 2015.
[4] A. Krizhevsky, I. Sutskever, and G. E.
Hinton,
"ImageNet classification with deep
convolutional neural networks,"