Security system by face recognition

Abstract

Background: Authentication is one of the major challenges of the information systems era. From Among other things, recognizing the human face is one of the known techniques that can be user authentication.

Aim: The main purpose of this paper is to study the application of the facial recognition algorithms as a security system for the examination office at Omar Al-Mukhtar University (OMU) for the first time in Libyan universities. It can detect intruders into restricted or highly secure areas, and help reduce human errors.

Methods: This system consists of two parts: the hardware part and the software part. The device part consists of a camera, while the software part consists of the face detection and facial recognition algorithms program. The face was detect using the Viola Jones method and face recognition is performed by Use of independent component analysis (ICA). When a person enters the area in question, his photos are taken by the camera and sent to the program for analysis and comparison with an existing database of trusted people.

Result: In this study, we used a database of people responsible for the examination office at the College of Science, the sample consisted of 100 images. The ICA algorithm was applied to recognize faces, and it achieved a result of 86.7%, which is considered a good result, as it was compared to applying the PCA algorithm on the same sample, where the result was 76.7%.

Conclusion: Independent Component Analysis (ICA) is better than Principal Component Analysis (PCA) in some cases, but requires at least a few seconds as is training time. The face recognition-based presence system achieves the expected goal, but the accuracy of the face recognition still needs improvement.