RollVision: Implementation of a Real-Time Face Recognition Based Attendance System

Abstract

The advancement of computer vision and artificial intelligence has enabled the development of intelligent automated systems for real-world applications. This paper presents RollVision, a real-time attendance management system based on facial recognition technology. The primary objective of the system is to eliminate manual attendance processes and prevent proxy attendance by providing a contactless and efficient solution. The system is implemented using the Django framework for backend processing, along with OpenCV and Dlib libraries for face detection and recognition. It captures live video input, detects faces, extracts unique facial embeddings, and matches them with stored data to mark attendance automatically. Additionally, an interactive dashboard is developed to manage student records, subjects, and attendance data efficiently. Experimental evaluation demonstrates that the system achieves recognition accuracy between 90% and 95% with minimal processing delay under varying environmental conditions. The proposed system is scalable, reliable, and suitable for deployment in educational institutions and organizations.