A Face Recognition Based Automatic Attendance Management System by Combining LBP and HOG Features

Abstract

In the growing technology era, educational institutes are particular about the regularity of the students. Because the academic performance and evaluation depends on the attendance of the student. However, the method of taking attendance of the students still remains the orthodox way i.e., calling the roll number or taking the signature of the students in a sheet of paper. The shortcomings of these methodologies are that they sluggish and are quite often influenced by duplicate data entries by the students. So in this paper we present a novel methodology of taking student’s attendance through face recognition technique. The facial features of the students are extracted via Local Binary Pattern (LBP) and Histogram of Oriented Gradients (HOG). Both LBP and HOG features are combined to create a new feature vector. A classification model is implemented using Support Vector Machine (SVM) classifier which predicts student based on comparison made between the features of the query image and the features of the images stored in the student database.