Anan Nugroho, Mona Subagja, Ledi Diyanasari, Muhammad Iqbal, Aisya Fathimah, Bagus Fachriansyah
Attendance systems in higher education institutions play a strategic role in ensuring academic and operational accountability. However, conventional systems continue to face challenges such as high latency and significant risks of proxy attendance. This study aims to design and implement a face verification-based attendance system that operates efficiently within a distributed edge-cloud architecture. The system was developed using a hardwaresoftware co-design approach, utilizing a Raspberry Pi 4 as an edge computing unit responsible for real-time face acquisition, quantized RetinaFace-based detection, user interface management, and door access control via EMlock actuator. Preprocessed facial images are securely transmitted to the cloud layer over HTTPS, where identity verification is performed using FaceNet trained with triplet loss (anchor, temporalpositive, and hard negative). The embedding database is managed using a MySQL sharding scheme based on identity ranges (A-F, G-M, N-Z) and load balancing via Nginx. The system is further supported by WebSocket-based real-time communication, Telegram notifications, and a no-touch sensor fallback for redundancy. Experimental results indicate that the system achieves 100 % face detection accuracy across three distance scenarios (0.5 m to 1.5 m), a 90 % face recognition accuracy using FaceNet, and an overall functional accuracy of 9 7. 0 0% across 1, 2 0 0 real-world trials. The EMlock actuator responded with an average activation time of 1. 9 9 seconds, while the complete face verification pipeline consistently operated with inference times under one second. These findings confirm the system's reliability, responsiveness, and practical viability for academic attendance applications in edge computing environments. © 2025 IEEE.
Universitas Negeri Semarang, Department of Electrical Engineering, Semarang City, Indonesia; Universitas Negeri Semarang, Directorate of Information System, Semarang City, Indonesia; Universitas Negeri Semarang, Department of Computer Engineering, Semarang City, Indonesia; Universitas Negeri Semarang, Educational Department of Computer, Semarang City, Indonesia
