Novel Algorithm for Skin Color Based Segmentation Using Mixture of Gmms

Image segmentation techniques are considered the main artifact against which the computer can visualize the objects in that scene and for further processing, many hand segmentation techniques are adopted in this direction which consumes the color cue as the enjoinder tools for spotting the skin pixels and non-skin pixels, GMM has been implemented successfully in this area which can tone single color model using single GMM, some shortcoming may come along with this technique, our proposed technique is to combine several color models each one is modeled using GMM and the final model is mixture of GMMs thatcan get rid of the single errors that may occur with single GMM, we have achieved a very high recognition rate when compared with applying single GMM for the pertinent color models, we have adopted four color models which are commonly used for skin color which are normalized RGB which referred as rgb, HSV, YUV and L*a*b*.