Illumination removal in color images using Retinex method based on advanced adaptive smoothing

An important challenge of any skin color based intelligent system is to filter off varying illuminant in color facial images. In this paper, we propose a novel Retinex algorithm which based on an advanced adaptive smoothing filter to remove illuminant in color images. The proposed method estimates illumination by convolving the input image with some adaptive smoothing mask windows, each mask weight in any window is computed via combining measures of the illumination discontinuity at each pixel. In this way, we can achieve an illumination estimation image of the original image which contaminated by illuminant. Then, Retinex theory is used to remove illumination estimation image from original image, and achieve the result image. This new algorithm is compared with available methods, and the comparative results indicate that the proposed method present more efficiency and promising result when images under harsh illumination conditions. In the end of this paper, the proposed method is evaluated based on natural condition color images by applying Bayesian skin color detector. The experimental results prove that the new method can increase accurate hit of skin color pixel detection and maintain real-time performance.