Fast adaptive smoothing based on LBP for robust face recognition

Adaptive smoothing for face image normalisation under variation of illumination is presented. The illumination is estimated by iteratively convolving the input image with a 3-by-3 averaging kernel weighted by a simple measure of the illumination discontinuity at each pixel. In particular, weights of a kernel are encoded into a local binary pattern (LBP) to achieve fast and memory efficient processing. In the framework of the retinex based method the proposed method gives superior experimental results.