Smart Desk in Hybrid Classroom: Automatic Attendance System based on Face Recognition using MTCNN and ARCFACE

Abstract

The attendance of students plays a vital role in improving learning outcomes to assess the level of student attendance. Teachers will check student attendance in several ways, such as by calling each student’s name and signing the certificate. In addition, there are several other automatic attendance-checking solutions, such as using RFID cards, fingerprints, or entering PIN codes. These methods will distract students if they have to take attendance many times in a lesson. In this article, we propose an automatic attendance solution on Smart Desks [Students’ desks in hybrid classrooms, operating on the platform of embedded devices equipped with touch screens, camera…] using MTCNN and ARCFACE algorithms for checking attendance automatically. Besides, we use the Face-Anti-Spoofing algorithm to eliminate the possibility of students using images or videos to cheat the system. After being processed by the Smart Desk, attendance data is sent to the API Server for storage. The list of times students do not appear in front of the Smart Desk is marked and detailed on the management system. Teachers and Admins can live-view or review attendance data on the website or Smart Podium, and parents will receive a notification when their child is not in class through the mobile application.