JSEG based color separation of tongue image in traditional chinese medicine

Abstract

The process of color separation of tongue image in traditional Chinese medicine (TCM) is decomposed into two steps: region partition and color classification. In the first step, a partition method based on JSEG is proposed to obtain homogenous regions in the tongue. In the second step, a priori template and some standard color patches are designed to assist the classification according to a priori knowledge from the TCM experts, where the nearest neighbour classifier is applied for color classification. The experimental results show that the proposed approach is greatly promising to computerized tongue diagnosis. © Springer-Verlag 2004.