Gabor Feature based Classification using Statistical Models for Face Recognition

Face recognition is one of the challenging applications of image processing. Robust face recognition algorithms should posses the ability to recognize identity despite many variations in pose, lighting and appearance. Principal Component Analysis (PCA) has been widely adopted as a potential face recognition algorithm. However, it has limitations like poor discriminatory power and large computational load. In view of these limitations with PCA, this paper proposes a face recognition method with PCA based on Gabor features. On applying the statistical models like Independent Component Analysis (ICA) and Linear Discriminant Analysis (LDA) on the output of reduced features from PCA, the more discriminating features were obtained. Two normalization methods, namely Unit Length normalization (UL) and zero Mean and unit Variance (MV) methods were employed for the normalization of extracted features in order to get a better classification results. The proposed Gabor feature based method has been successfully tested on ORL face data base with 400 frontal images corresponding to 40 different subjects which are acquired under variable illumination and facial expressions. It is observed from the results of PCA with Gabor filters that the ICA method gives a top recognition rate of about 95% when compared to LDA method with MV normalization method. (C) 2010 Published by Elsevier Ltd