Face Detection In Color Images Using Pixel Based Skin Color Detection Techniques

The selection of the best color space for skin detection in color images is important in many computer vision areas and this paper presents a comparative study on the pixel- based skin color detection techniques. Three main issues of the face detection are the selection of the best color space, skin color pixel classification algorithm and face detection algorithm. A large set of XM2VTS face database is used to examine whether the selection of color space can enhance the compactness of the skin class and discriminability between skin and non-skin class in thirteen color spaces, six different skin color pixel classification algorithms and one face detection algorithm. The results show that 1) the selection of the color space can improve the skin classification performance 2) the segmentation performance degrades only when chrominance information is used for classification 3) Bayesian classifier is found to perform better as compared to other classification algorithms, like Gaussian classifiers. Piece-wise linear decision boundary classifier algorithm outperforms all the other skin classification algorithms when used for images with good illumination conditions. The template matching technique is used to mark the face region in the image.