Research, Transform Domain And Pca-based Face Recognition Methods

Eigenface method has been developed for face recognition. It holds best quality in energy compression with weak ability of antijamming. It belongs to statistic method for image without using the special information of human face. The method uses K-L transformation to extract features and recognizes in the sub-space.In respect of the character of automatic face recognition system, firstly, the corresponding preprocessing for human face image is utilized and localization method for human eyes is proposed, as well as noise removal, normalization and rotation for face image. The eigenface method has been improved in the recognition process. The corresponding process for face image has been employed based on the following two domain:1. Face recognition method based on wavelet domain and PCA. This paper discusses it by combining wavelet method and eigenface method. The face image is decomposed by wavelet transform. Because low frequency subband contains approximation information of image, and high frequency subband contains detail information in different direction, different schemes are utilized in face recognition. The recognition rate has been improved with the reduction of recognition time.2. A face recognition method based on DCT domain and PCA. Because current image compression scheme are mainly focused on DCT transformation, three level multiresolution subbands from reordered DCT coefficients for face image are acquired ahead of the using of eigenface method. Different results are achieved by applying the eigenface method on each single method. We further propose torecognize by the eigenface method on subband S₀ and subband S₁ S₂ and S₃. Ifwe get different results, more face images are input under certain threshold condition. We analyze the relationship of recognition time between the proposed method and the original method. The experimental results on ORL and YALE database shows that the...