Design of an E-Attendance Checker through Facial Recognition using Histogram of Oriented Gradients with Support Vector Machine

Abstract

The usual way of checking the attendance in a class has its own drawbacks. To be able to resolve it, automated attendance systems were introduced. In this paper, the design and development of an e-attendance checker using a facial recognition system were implemented. It can scan the faces of multiple students in a standard classroom setup. A commonly used approach for face detection called Histogram of Oriented Gradients (HOG) with Support Vector Machine (SVM) was applied to examine the effect of luminance of the surrounding, the facial orientation of the student and so as their distance from the camera in the facial detection and recognition. The obtained attendance will then be uploaded to a database with authentication. It was found that the system has an accuracy of 95.65% and can detect and recognize up to 37 students. It is suggested that the classroom should have a luminance level of about 217.39 lux or higher to achieve a better accuracy performance of the system. As for the analysis of the effect of distance in the system, it is claimed that the distance of the student does not affect the accuracy of the system. Lastly, it is suggested that the face angles of the subject should be directly facing the camera to achieve a more accurate recognition result.