Statistical Multiframe Methodology with Agnostic Thresholding for Attendance Marking System

Abstract

This paper proposes a multiframe methodology for improving the accuracy of attendance marking systems. The proposed method utilizes agnostic thresholding to segment the foreground (students) from the background, even in challenging lighting conditions and varying student appearances. By combining information from multiple frames, the system can robustly detect and identify individual students, thereby addressing common issues such as occlusions, pose variations, and illumination changes. Experimental results demonstrate that the multiframe approach significantly outperforms single-frame methods in terms of precision and recall, leading to a more reliable and efficient attendance marking system. The methodology is designed to be adaptable to different classroom environments and camera setups, making it a versatile solution for automated attendance tracking.