A Two-level authentication for Attendance Management System using deep learning techniques

Monitoring attendance is an essential administrative function in all educational institutions and organizations. A well-structured framework will facilitate the expansion of institutions. It reduces the instructors’ time and effort by assisting both students and teachers in improving attendance. The existing conventional physical classroom system is insecure, disruptive to teaching, and time-consuming to gather and store student attendance, which hampers the educational activities. The proposed system is a hybridized framework of face detection and recognition, and ID card detection and card text verification that adds to the two level authentication system. At the first level, the proposed system recognizes the individual, authenticates it with database data, and detects the ID card using deep Hog based ResNet feature extraction syttem. At the second level, YoloV7 based Easy OCR reads the details and marks the concerned individual as present. This hybridized framework is accurate in identifying the persons irrespective of the illumination conditions and an efficient attendance system.