I am trying to extract all the skin pixels in the image. Images are not pre-known. They could be captured in various lighting conditions. The method that I currently use involves converting the image from RGB color space to YCbCr color space and classifying pixels as skin pixels if they happen to fall within certain range of Y, Cb and Cr.
Could someone suggest a better method ? Could converting image to HSV model prove useful ? If YCbCr is best color space to work in then what should the ranges for all the factors.