Smart Attendance Monitoring System Using Local Binary Pattern Histogram

Abstract

Attendance marking of the students in a classroom is one of the most important activity carried out by the teachers to maintain the physical presence of students as well as for records. The traditional process of taking attendance of students is time consuming, error prone, and student can proxy. To eliminate or reduce these disadvantages, a new scheme is proposed in this paper in which the attendance monitoring is done through face detection. This will eliminate the time required for the teacher to read out the names of students and also eliminate the chances of students making proxies. The system uses face recognition algorithm which captures the images of students present in the class using a camera and compares with the images stored in the database. The key challenges involved are the quality of images getting captured and the accuracy of the face recognition algorithm. Experiments prove that the proposed system achieves more than 90% accuracy when implemented with a database consisting of 493 images and tested against a class of 54 students. The accuracy can be improved by equipping better quality camera, better lighting conditions, and more accurate face recognition algorithm.