Toward a Complete System for Pupil Monitoring Using Deep Learning for Digital Twins
Résumé
School dropout is a significant issue with profound economic and social implications. Research suggests that addressing key factors – such as repetitive absenteeism, parental disengagement, and a negative school climate – can effectively mitigate the problem. This chapter presents a comprehensive system designed to prevent school dropout by targeting these underlying causes. The system consists of four subsystems implemented on an embedded platform. The first is an attendance detection system with GUI facilitating the use and the update. The second is an access control system that safeguards students by preventing unauthorized access to restricted areas like laboratories. Both systems use face detection and recognition techniques. The third subsystem focuses on behavior recognition to detect and prevent bullying and assaults. Lastly, the system integrates IoT technology for data logging and real-time parent notifications. With an accuracy rate of 92.12%, further enhanced by a multi-test confirmation mechanism, the system demonstrates a promising solution for reducing school dropout rates.