Method for Lip Extraction from Face Image Using HSV Color Space

Abstract

In this paper, we propose a method for lip extraction from face image by color image processing on the basis of the distribution of H and S parameters in HSV (hue, saturation, value) color space which shows a better fit with human color sense than RGB (red, green, blue) color space. Comparing the proposed method with the conventional lip extraction method using the signal ratio of G to R of RGB signals, on the basis of the defined lip extracting index, the effectiveness of the proposed method is quantitatively demonstrated. This quantification of face image by using HSV color space can be applied to the development of color data base of human face and virtual face image.