A Bayesian approach to skin detection in YCbCr color space
WU Zhong-dong, Wang Saichao, Han Zichao · 2013
Skin color has been proved to be a useful way for face detection, localization and tracking. To effectively detect the skin region in images, we implement two different skin color classification techniques. We began by first constructing the skin color models with one (Cb-Cr) lookup table, and then using two (Cb-Y and Cr-Y) lookup tables. In the model by two lookup tables (LUTs) we take the influence of luminance (Y component) on skin color into consideration. In order to use two LUTs together, Gaussian normalization and linear transformation are used to normalize the range of threshold value to [0,1]. Experimental results have demonstrated that the model by two LUTs can effectively detect skin regions, and the problems are also introduced in this paper.