Research and Development of Attendance System for Smart Classroom Using Bluetooth Low Energy and Facial Recognition Technologies

Attendance is one of the crucial factors affecting students' academic performance and reflecting their final grades. Therefore, an effective and accurate attendance marking method is necessary. Commonly used techniques such as roll call or signing in papers can result in forgery and consume class time. Even technology-based methods like Radio Frequency Identification (RFID) cards do not ensure transparency, as anyone can use the cards. We propose a smart attendance system using facial recognition technology to address these issues. Our approach aims to track student presence throughout the class, unlike current systems that only recognize faces once to mark attendance. We use You Only Look Once (YOLO) for face detection, FaceNet for feature extraction, and SVM for classification. Bluetooth Low Energy (BLE) is also implemented as a second data source from students' mobile devices to verify presence in certain circumstances where facial recognition fails. Internet of Things (IoT) devices are deployed at the edge to collect and process data before sending it to the server. Multiple cameras and BLE stations are set up to ensure adequate information from the classroom for optimal results. We have successfully tested the system in a simulated classroom environment with seven members, recording cases of absence, entering and leaving the class, and the total participation time of each individual.