Performance Evaluation of Face Recognition using PCA and N-PCA

Abstract

Face recognition has become a valuable and routine forensic tool used by criminal investigators. It is an important area of computer vision research and has gained significant interest in recent years. Efforts in improving security, such as automatic surveillance and the use of biometrics in identification, are partly responsible for this increased interest. However, several challenges remain in improving the accuracy of face recognition under illumination changes, variations in pose, occlusions, and image resolution. This paper presents performance comparison of face recognition using Principal Component Analysis (PCA) and Normalized Principal Component Analysis (N-PCA). The experiments are carried out on the ORL, Indian face database and Georgia Tech face database which contain variability in expression, pose, and facial details. The results obtained for the two methods have been compared by varying the number of training images and it has been found that as the number of training images increases efficiency also increases. The result also shows that N-PCA gives better results than PCA.