An Improved Algorithm (KPCA) For Face Recognition

Abstract

Within computer vision face recognition has become increasingly relevant in today‟s society. The major motivating factors for this are the understanding of human perception, and a number of security and surveillance applications such as access to ATMs, airport security, tracking of individuals and law enforcement. Early work on face recognition involves methods such as principal component analysis, elastic bunch graph matching and optical flow based techniques. Facial recognition analyses the characteristics of a person‟s face images input through a digital video camera. It measures the overall facial structure, including distances between eyes, nose, mouth, and jaw edges. These measurements are retained in a data base and used as a comparison when a user stands before the camera. For faces to be a useful biometric signal, facial features used for face identification should remain invariant to factors that modify face image appearance. The proposed work is based on Kernel Principal Component Analysis which overcomes the ineffectiveness by extracting the face image features in high dimensional spaces. KPCA is an improvement of PCA which extracts feature set more suitable for categorization than classical PCA. KPCA is good at dimensional reduction and achieves better performance than PCA. KPCA is an appearance based approach that decomposes face images into small sets of characteristics feature images called eigenfaces.