| Face recognition has developed over 30 years and is still a rapidly growing research area due to increasing demands for security in commercial and law enforcement applications. Although, face recognition systems have reached a significant level of maturity with some practical success, face recognition still remains a challenging problem due to a large variation in face images. In this thesis, we develop a face recognition system with applications to security and individual face identification and verification. For this purpose, we study and compare two methods for face recognition---one using the Binary Phase-Only Filter (BPOF) and the other using an Extended Template Matching Correlation method. Influence of factors like illumination, pose of the subject and rotation on the recognition results is studied. An optimal face recognition system is developed. Experimental results are presented to show the potential of the proposed face recognition method. We show that the extended correlation method outperforms the BPOF by an order of magnitude. However, the BPOF showed less sensitivity to the effect of illumination than the extended correlation. The proposed optimal face recognition system was implemented using the BPOF and demonstrated an accuracy of 91%. |
