Tongue Body Localization Based on Image Clustering and Gray Projection

Abstract

Tongue diagnosis is often used in Traditional Chinese Medicine (TCM). Tongue image segmentation, which extracts tongue body, is a key step when manufacturing an automated system of tongue diagnosis. Localization of tongue body depicted by a rectangle is a useful preprocessing step in tongue image segmentation, which can eliminate the adverse effect of strong edges from neighboring tissues such as face and lip when extracting tongue body contour. After exploring the existing tongue body localization method based on gray projection, we propose an upgraded method combining image clustering with gray projection. Specifically, our proposed method first conducts the clustering on the image hue component in HSI (i.e., hue, saturation, and intensity) space to determine three thresholds. Then, image thresholding and morphological operations are sequentially performed to generate a binary image, and its largest object region is taken as the initial localization result. Finally, the localization result is refined by performing gray projection on the image red component. Experiments on a variety of tongue images showed that our proposed method significantly improves the accuracy of tongue body localization in comparison with the existing gray projection method.