舌診影像自動診斷系統之研製

Tongue Image Analysis for Automatic Diagnostic System

王節宣(Chieh-hsuan Wang)

指導教授 ： 魏清泉

朝陽科技大學/資訊學院/資訊與通訊系碩士班/碩士(2011年)

摘要

中醫學之精髓在於「辨證論治」，而辨證以望、聞、問、切四診為手段，依「四診合參」原則，綜合各類訊息，加以分析歸納，達到審查病因、辨明病態、闡述病機、 確定治療原則，以及判斷預後轉歸等目的，其中舌診是中醫望診中重要項目；舌頭是反應身體內在狀態的鏡子，臟腑病變可由舌部之變化而得知，舌診即是通過對舌部之觀察而探知體內狀態，但是傳統舌診判讀往往會受醫師主觀認定、經驗累積及當時環境因素所影響，缺乏客觀指標，無法達到研究上可重複性之要求，因此為實現舌診診斷上客觀化及定量化之目標，中醫舌診電腦化之研究為必然發展趨勢。

本篇論文主要根據中醫舌診，為了實現中醫舌自動診斷系統，有效的分割邊緣檢測方法的舌頭是非常重要的。我們主要先使用HSL的Luminance的灰階影像與Otsu自動選取閥值（Threshold）的方法配合進行邊緣切割。而我們利用切割完後的舌面進行舌苔及舌質的分離方便進行顏色的分析，而針對舌苔進行顏色的量化，利用Color Threshold 以及根據傳統中醫舌診對顏色定義，將舌苔定義出量化的值，以方便我們在舌質上的擷取。再此之後，利用得到的舌質進行顏色的分析，針對傳統中醫舌診對舌質顏色定義，我們也將顏色進行測試及定義後，將以顏色與顏色之間的距離進行分類，並以分類後的比例決定舌質之顏色，根據分析結果找出對應之病理。

關鍵字

診斷系統 ； 舌診 ； 舌診影像

並列摘要

The essence of Chinese medicine is "syndrome differentiation", the differentiation to look, smell, and asked, cutting by means of four diagnostic, according to the "combination of all" principle, integrated various types of messages, to be analyzed and summarized, to review the causes, identify sick to explain the pathogenesis, treatment principles established, and the prognosis of outcome such purposes. Tongue is one of the important items in Traditional Chinese Medicine by inspection. Internal state of the tongue is a mirror of the body reaction, the internal organs of tongue lesions may change that, tongue or through the observation of the tongue in vivo state of discovery. However, interpretation of the traditional tongue diagnosis is often subjective and determined by the physician, with experience and the prevailing environmental factors, the lack of objective indicators, the study of repeatability can not achieve the requirements, so as to achieve the objective of tongue diagnosis, and quantification of objectives, Tongue computerized research as inevitable trend of development.

This thesis is based on Traditional Chinese Medicine Tongue, Tongue in order to achieve automatic diagnosis system, an effective segmentation of the tongue edge detection is very important. We mainly use HSL’s Luminance of gray-scale image and automatically selects the Otus’s method with the cutting edge. We use cutting the tongue, and after the separation of tongue body and tongue coating color analysis to facilitate, and for the tongue to color quantization. Then we use Color Threshold and follow Traditional Chinese medicine of tongue diagnosis to color’s define that will define the quantitative value of tongue, to facilitate the acquisition on our tongue. Then after that, use the tongue to get the color analysis, and Traditional Chinese medicine tongue diagnosis on tongue for color definition. We will also test the color and definition that will be the distance between each color classification, and the ratio after the decision to classify the color of tongue, according to analysis results to identify corresponding pathology.

並列關鍵字

Tongue Diagnostic ； Tongue Image ； Diagnostic System

參考文獻

[10] N. Otsu, “A Threshold Selection Method From Gray-Level Histograms”, IEEE Trans. on System, Man and Cybernetics, 1979, vol. SMC-9, no. 1, pp.62-66.

[11] Li Qingli, Liu Zhi, “Tongue color analysis and discrimination based on hyperspectral images”, Computerized Medical Imaging and Graphics, 2009, Volume 33, Issue 3, Pages 217-221.

[12] Bo Huang, Jinsong Wu, David Zhang, Naimin Li “Tongue shape classification by geometric features”, 2010, Volume 180, Issue 2, Pages 312-324.

[14] Chieh-Hsuan Wang, Ching-Chuan Wei, Che-Hao Li, “Comparison of Edge Segmentation Methods to Tongue Diagnosis in Traditional Chinese Medicine”, 2010, Volume 6422/, Pages 232-238.

[1] 北京中醫學院中醫系中醫基礎理論教研室，中醫舌診，第一版，人民衛生出版社，1995。