Real Time Attendance System Using Facial Recognition
Vikas Kumar, Santosh Kumar · 2024
Taking student attendance is a daily essential activity in academic institutions. Manual attendance marking is tedious, time-consuming and prone to errors or proxy attendance. Automating the attendance process using biometric technology like face recognition can overcome these challenges and provide additional benefits like analytics. This paper introduces an automated attendance system leveraging facial recognition technology to streamline attendance tracking, minimizing manual effort typically associated with the process. Achieving a high level of accuracy, the system caters to English and Hindi lectures, with potential for expansion to encompass additional subjects. Through a user-friendly Tkinter interface, students’ facial data is seamlessly enrolled in MongoDB and CSV databases, facilitating efficient management. Model training, utilizing a Keras Sequential architecture with a Softmax output layer, incorporates multi-threading for enhanced efficiency, enabling optimal performance adjustments through adjustable hyperparameters. Real-time attendance management allows users to select lectures, check existing records, and mark attendance, with recognition thresholds triggering capture. The system’s workflow, from data generation to attendance capture, is guided by intuitive User Interface interactions, ensuring systematic enrollment and accurate record-keeping. Overall, this automated solution presents a promising avenue for educational institutions, offering heightened efficiency and accuracy, with potential for future enhancements including language support expansion, algorithm refinement, and feature integration to bolster usability and scalability.