AUTOMATIC ATTENDANCE SYSTEM USING DEEP LEARNING

Abstract

The attendance management system in some institutes is a tedious process. Taking attendance using attendance sheets may take longer and there may be chances to lose attendance sheets and manipulations can be done. Also, students may forget to tell their attendance in class. To avoid these problems, we propose a system that identifies an individual by comparing digital image data or CCTV cameras with the stored record of that person. Using CCTV cameras, live capture of students is done, then identify and recognize the individual for marking attendance. This process is used to capture the entire classroom to mark the attendance for that session. In this system Multi Task Cascaded Neural Network (MTCNN) algorithm is used for detection of faces and FaceNet algorithm is used for recognition of an individual. After the facial recognition process it updates and generates an attendance sheet. The proposed attendance management system displays a excel sheet which contains of student’s information whether absent or present. It is flexible and reliable system than that of traditional attendance system.