A Robust Approach to Student Attendance Using Web-Based Facial Recognition

Irfan Sahputra (Universitas Malikussaleh)
Muhammad Fikry (Universitas Malikussaleh)
Kurniawati Kurniawati (Universitas Malikussaleh)

Proceedings of International Conference on Multidisciplinary Engineering (ICOMDEN)
Vol. 2 (2024): Proceedings of International Conference on Multidisciplinary Engineering (ICOMDEN)

In this paper, we introduce an innovative student attendance recording system that utilizes computer vision and machine learning to improve attendance management in educational settings. By employing YOLOv8 for real-time face detection and MobileNetV2 for face recognition, the system achieves high accuracy and efficiency across various classroom conditions. Rigorous testing in diverse lighting environments and varying student densities demonstrated a peak recognition accuracy of 98% in well-lit conditions, with an average face detection time of under one second. This system offers a more robust, efficient, and scalable solution than traditional manual attendance methods, addressing common limitations in accuracy and reliability. Future work will target optimization under low-light conditions, enhancing its applicability in real-world scenarios.