Face recognition using multiresolution wavelet combining discrete cosine transform and Walsh transform

Abstract

In this paper a face recognition system based on multi resolution hybrid wavelet approach has been presented. The multi resolution hybrid wavelet transform matrix is generated using Kronecker product of Walsh and DCT transform matrices. This wavelet is used to extract features from face images with different expressions of subjects' faces. A feature map is generated using energy compaction technique which is used as a template to extract features of enrolled and test images. The experiments are performed on faces94 database with different variations in facial expression, change in face position and occlusion. The recognition rates achieved are 99.24%.