In this paper, we propose an efficient illumination preprocessing algorithm for face recognition. One of the best known illumination preprocessing method, based on anisotropic smoothing, enhances the edge information, but instead deteriorates the cont...
In this paper, we propose an efficient illumination preprocessing algorithm for face recognition. One of the best known illumination preprocessing method, based on anisotropic smoothing, enhances the edge information, but instead deteriorates the contrast of the original image. Our proposed method reduces the deterioration of the contrast while enhancing the edge information, and thus the preprocessed image does not lose features like Gabor features of the original images much.. The effectiveness of the proposed illumination preprocessing method is verified through experiments of face recognition.