AnAI-Based SmartAttendance Management System Using Real-Time Face Recognition

The growing demand for automated and accurate attendance tracking has driven educational institutions to explore intelligent digital solutions. Traditional attendance methods, including manual roll calls and RFID-based systems, are prone to errors, proxy attendance, and significant administrative overhead. This paper presents an AI-Based SmartAttendanceManagement Systemthat leveragesreal-time face recognitiontoautomate studentattendance inhigher education institutions. The system is developed using Python (Flask) for the backend, InsightFace (ONNX) as the primary face recognition engine, OpenCV for video processing, and a responsive HTML/CSS web dashboard for realtime monitoring. The system supports simultaneous detection of multiple students per frame, differentiates between Present and Late statuses based on a configurable attendance window, and exports dailyattendance records as CSVfiles. Experimental results demonstrate that the system achieves high recognition accuracy, low latency, and reliable performance under varied lighting and classroom conditions.