Self-adapting Lip Contour Detection Method based on Active Shape Model and Color Information

摘要

This paper proposes a self-adapting method for lip contour detection based on Active Shape Model and color information. In order to improve the accuracy of lip contour detection, this paper takes the advantages of ASM and color information to detect lip contour. And it is carried out without limited conditions such as light or background. Firstly, this method uses ASM to locate the mouth region from the image, this remove the interference of complex background. Secondly, color information and fuzzy reasoning theory are adopted to accurately detect the mouth region to extract the lip. Finally, this method uses curve fitting to constrain the lip region to generate the lip contour. The experiment results show that the proposed method can detect the lip contour accurately under different conditions and has strong self-adapting ability.