Multi-Agent Approach to Face Recognition in Border Control

Abstract

In this paper, we have proposed a multiagent approach to face recognition in border control. In today’s unstable world, security is the most important aspect. Border criminals, frauds, and unauthorized immigrants are burning issue in Sri Lanka within the last few years due to the lack of proper identification system. Therefore, the efficiency and accuracy of the traditional face recognition system are not good enough to overcome this disaster. The face biometrics is an ideal solution for authentication as it has an advantage over conventional systems. The main objective of the research is to provide an automated and effective solution for face recognition using limited data and time-critical application. The face biometrics is the main ingredients of the proposed system and overall process done by using multi-agent system (MAS) to compare the latest face biometrics with a physical appearance at the border points. The face recognition algorithms, Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) are used as an agent within the multi-agent environment can be identified as the novelty of the proposed system. Less human intervention and smoothly adaption to the current system can be identified as an advantage of the system. The system has been analyzed with the traditional system and evaluated with authentic biometric samples and identified with 96% accuracy with comparing to the traditional system