Student attendance with face recognition (LBPH or CNN): Systematic literature review

Abstract

Technology growth is speedy, and more and more things can be solved easily with the existence of sophisticated technology. One of them is solving the problem of student attendance at the university. The current attendance system has developed into RFID (Radio Frequency Identification) from the previous manual. However, many things still become obstacles. For example, students who miss their cards cannot take attendance, and the problem of leaving attendance can be cheating. Now, this has developed much technology in the form of face recognition with various algorithms that can be used. The use of face recognition can overcome the previous problem because it only uses faces for attendance. However, the many algorithms for facial recognition make it difficult to determine the best algorithm to implement. The main purpose of this literature review is to compare algorithms suitable for implementation in an environment at universities, especially CNN and LBPH. Based on the literature review, it was found that CNN's accuracy is superior in terms of accuracy compared to LBPH, CNN also produces more stable accuracy if there are external factors that can affect accuracy.