The Research on Facial Segmentation Algorithm for Facial Diagnosis in Traditional Chinese Medicine

Abstract

A method for accurate segmentation of human face based on region growing and ellipse fitting is presented on facial diagnosis in Traditional Chinese Medicine (TCM). Firstly, the region growing method is used to locate the face skin region, and the possible area is segmented. Then the non-face area is removed by a morphological processing step, and further the edge and the initial outline of the face is obtained. Finally, the exact contour of the face is extracted by using the direct least square ellipse fitting method according to the acquired edge and initial contour. The experimental results have shown that the algorithm can overcome the interference problems between chin contour and neck region, and meet the requirements of face contour extraction for an image processing system of TCM facial diagnosis, which lays a foundation for further segmentation and detection of facial features.