A-eye: Attendance monitoring using face detection and recognition from CCTV footage

Monitoring and updating attendance records of students is an integral part of activities in schools and colleges. To mitigate the laborious work of keeping attendance records, an automated method of attendance monitoring using face as a biometrics is proposed. In this paper, face detection and recognition for maintaining student attendance using deep learning methodology is presented. Face detection in low-resolution CCTV footage is achieved using the Haar Cascade algorithm with 88.59% detection accuracy. The detection is not limited to frontal face detection. It also has side face selection along with varied illumination situations. These detected faces are then used to create a student face database. The Convolutional Neural Network (CNN) is trained on this face database for student recognition to mark attendance. A total of 60 students' face data is used in this for recognition and updating attendance for 5 subjects. The proposed CNN provided 76% face recognition accuracy for the implemented network.