ADVANCED SMART ATTENDENCE SYSTEM USING REAL-TIME MULTIFACE RECOGNITION WITH DEEP LEARNING
Mrs.Deepthi Nair P, Mr. Gokul E,
Mr. Hemabal S, Mr. Harish R
Assistant Professor, Department of Computer Science and Engineering, Sri Shakthi Institute of Engineering and Technology,
Coimbatore, Tamil Nadu.
Email: deepthinair@siet.ac.in
UG Students, Department of Computer Science and Engineering, Sri Shakthi Institute of Engineering and Technology
Abstract - SmartAttend is a production-ready, real-time multi-face recognition attendance system engineered specifically for the complexities of modern college environments. Traditional manual roll-calling methods are often inefficient, prone to errors, and vulnerable to "proxy" attendance. This system addresses these limitations by automating the identification process through a standard webcam interface, utilizing a robust tech stack comprising Flask, PostgreSQL, and InsightFace.
The core of the system relies on deep learning architectures, specifically MTCNN for high-precision multi-face detection and FaceNet for generating 512-dimensional facial embeddings. By achieving a recognition accuracy of over 99%, SmartAttend provides a reliable alternative to traditional biometric or manual systems. The system is designed with a Role-Based Access Control (RBAC) model, offering specialized dashboards for Admins, Teachers, and Class Advisors to manage departments, schedules, and cumulative reports.
Data is managed through a hierarchical storage structure, organizing face images and embeddings by Year, Department, and Section to ensure scalability and prevent data collisions. With features such as real-time MJPEG streaming, automatic teacher conflict detection in the timetable, and the ability to export detailed CSV reports, SmartAttend streamlines the administrative workflow of educational institutions while maintaining a modern, minimalist "Ocean Cream" aesthetic
Keywords - Face Detection, Machine Learning, Haarcascade, Computer Vision, Video Analysis
