Combination of polar edge detection and active contour model for automated tongue segmentation

Abstract

Tongue diagnosis is an important diagnosis method in Traditional Chinese Medicine (TCM) and recently the development of automated tongue image analysis technology has been carried out. Automated tongue segmentation is difficult due to the complexity of pathological tongue, variance of tongue shape and interference of the lips. In this paper we present a novel method for automated tongue segmentation by combining polar edge detector and active contour model. First a novel polar edge detector is proposed to effectively extract the edge of the tongue body. We then introduce a method to filter out the edge that is useless for tongue segmentation. A local adaptive edge bi-thresholding technique is also proposed. Finally an initialization and active contour model are proposed to segment the tongue body from the image. Experimental results demonstrate that the novel tongue segmentation can segment the tongue accurately. A quantitative evaluation on 50 images indicates that the mean DCP (the distance to the closest point) of the proposed method is 5.86 pixels, and the average true positive (TP) percent is 97.2%.