Retinex method based on CMSB-plane for variable lighting face recognition

In this paper, we propose a new Retinex method based on CMSB-plane for variable lighting face recognition. Cast shadows, created by direct light source, seriously damage face images and ultimately deteriorate the performance of face recognition system. To eliminate cast shadows efficiently, they should be preserved in the illumination estimation process. The proposed method estimates illumination by iteratively convolving the input image with a 3times3 averaging mask weighted by coefficients determined by using the combined most significant bit-plane (CMSB-plane). Then, illumination normalization can be achieved by well-known Retinex theory. In this way, we can achieve a fast illumination normalization in which even face images with cast shadows are normalized efficiently. The proposed method has been evaluated based on the CMU PIE database and Yale face database B by using PCA. In result, the proposed method has higher recognition rates than other conventional illumination normalization methods such as SSR and SQL Therefore, we believe that the proposed method has a great potential to be applied to real time face recognition systems, especially under harsh illumination conditions.