Developing a Student Monitoring System for Online Classrooms Based on Face Recognition Approaches

• Published in: Computational Collective Intelligence Vol. 13501; pp. 555 - 567
• Main Authors: Pham, Trong-Nghia, Nguyen, Nam-Phong, Dinh, Nguyen-Minh-Quan, Le, Thanh
• Format: Book Chapter
• Language: English
• Published: Switzerland Springer International Publishing AG 2022; Springer International Publishing
• Series: Lecture Notes in Computer Science
• Subjects: ArcFace; CenterFace; Face recognition; FaceNet; Student monitoring system
• ISBN: 9783031160134; 3031160134
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-031-16014-1_44

One of the primary activities that lectures usually do is to take a roll call. This activity not only helps lecturers determine the participation of students but also detect strangers in the classroom. When the number of students increases, lectures take more time to monitor and check students’ attendance. We propose a student monitoring system based on facial recognition approaches to tackle that problem. With the recent development of deep learning techniques, many new approaches have made remarkable progress in face recognition. However, most of those approaches only focus on improving accuracy, while a practical end-to-end face recognition system demands good accuracy and reasonable runtime. We make adjustments and apply CenterFace for the face detection task and ArcFace for extracting embedding features from images to achieve high efficiency in both accuracy and speed. In addition, our proposed system is designed to be lightweight and scalable, capable of running in various environments, especially in a web browser. The results show that the system takes an average of 0.22 s to register a new face and 4.3 s for identifying a face in a database of 500 samples. Experiments also indicate that the system was less likely to misrecognize faces in most of our tests.