Automated Attendance System using FPGA based on Face Recognition

Abstract

The manual method of taking attendance involves the director or the school instructor doing it by hand, which takes a lot of time and effort and involves mistakes and proxy attendance. Automated systems involving the use of biometrics like fingerprint and iris recognition have been well developed in recent times, but they are economically highly expensive for deployment on a large scale. Biometric features such as face recognition can be utilized to address these problems,which involve several stages, including image gathering, face identification, feature extraction, facial recognition, and finally recording the attendance. In this design, we have attempted to develop a successful and effective ”Automated Attendance System Using FPGA Based on Facial Recognition”. The goal of this design is to detect and recognize the face in a captured image or frame using MTCNN and FaceNet for attendance purposes using FPGA, where we have explored two approaches: PS-PS(Processing System) and PS-PL (Programmable Logic)approach. Our system achieved an accuracy of 88.7 percent using the PS-PS approach, and 72.5 percent with the PS-PL approach. While the PS-PL approach demonstrated reduced accuracy, it offered improved performance and decreased latency, making it a promising alternative for real-time applications in automated attendance.