This paper addresses the problem of lip segmentation in color space, which is a crucial issue to the success of a lip-reading system. We present a new segmentation approach to lip contour extraction by taking account of the color difference between lip and skin in color space. Firstly, we obtain a lip segment sample via a color transformation sequence in 1976 CIELAB and LUX color spaces. Secondly, we establish a Gaussian model and make use of the hue and saturation value of each pixel within the lip segment to estimate the model parameters. Subsequently, the memberships of lip and non-lip regions are calculated, respectively. Thirdly, we employ morphological filters to obtain the desirable lip region approximately based on the memberships. Finally, we extract the lip contour via convex hull algorithm with the prior knowledge of the human mouth shape. Experiments show the efficacy of the proposed approach in comparison with the existing lip segmentation methods.
展开▼
