Abstract
Abstract In recent year biometric technology has received a great attention. One of the newest area in biometric technologies is the automatic face recognition. Face recognition has developed over last decades and still a rapidly growing research area. Although, face recognition systems have reached a level of practical success but still remains a challenging problem due to large variation in face images. The aim of the proposed work is to build an efficient automatic face recognition. Data base of gray-level images for the proposed system are selected from the Face Recognition Technology FERET. Then primary processing to these images are performed through the downsampled to each face by bilinear method. Then these images were masked by a rectangle that include face region only. Wavelets transformation is based for face recognition in this experiment due to their powerful efficiency in face recognition area. Face features were extracted through the use of the 2D 2-level wavelets decomposition. The 2D Vertical and Horizontal subimages are selected. These subimages are selected due to their less sensitivity to image variations. As well as their components form the most informative subimage equipped with the highest discriminating power. Then the images are segmented blocks, the Statistical moment is used to extract features per block. The proposed work used accurate techniques to analysis the recognition which reflects significant enhancement results. These results are represented by accurate measures varied from 75% to more than 100% compared with other system on the same area.
