A Group Photo-Based Attendance System using Deep Learning

Abstract

Traditional methods of marking attendance in educational institutions, such as manual roll calls or even single-person biometric scans, are often time-consuming, prone to proxy attendance, and inefficient for large groups. This paper presents the design and implementation of an automated attendance system that leverages the power of deep learning for face recognition within a single group photograph. The system is built upon a robust face recognition library to detect and identify multiple faces simultaneously. The core process involves capturing a group image, detecting all faces present, generating unique facial embeddings for each detected face, and comparing these against a pre-enrolled database of students. The results are seamlessly populated and managed through a web-based interface powered by the Flask micro-framework, with attendance records stored in an SQLite database. This approach offers a non-intrusive, scalable, and efficient solution to automate the attendance process, significantly reducing manual effort and improving accuracy.