Smart Attendance Monitoring System Using Face Registering

Abstract

Face recognition is one of the main branches of biometric verification as the face is the identity of an individual and it is utilized by many organizations to mark attendance of employees. Currently student attendance is recorded physically in the classroom by calling their unique ids, which utilizes more time and it is tedious to verify and identify each student if the number of students increases beyond normal range and it is tough to cross verify whether the authenticated students are actually responding. This project demonstrates a technique for attendance monitoring with facial recognition method by using two different algorithms one will be the existing algorithm such as Principal Component Analysis (PCA) algorithm and the other one is proposed by us which is Unconstrained FACE REGISTERING ALGORITHM. This method will automatically record the attendance of the scholars who are present in the classroom and it will also maintain a login and logout time of students and faculties and administration can easily access all the data of the students. However, it is difficult to estimate the outcome of facial recognition as most of the systems currently present have low detection rate and takes 20-100 images of a person for better identification. In this project attendance is marked by continuous observation which helps the system to improve and it also eliminates few features which affects the performance of the system that are different poses, light effects, partial occlusion etc. which helps the system to achieve better accuracy. This method will save approx 15-20 min of valuable class time and can be used to interact or clear doubts with teachers.