摘要
中醫辨症以望、聞、問、切四診為依據。舌診是望診中主要的內容，中醫的舌診是累積數千年的臨床經驗，舌的各部份分別對應人體的臟腑。本研究利用影像處理的技術，建立了一套即時性舌診影像取象系統。 本研究取象裝置為網路攝影機與個人電腦，可以達到攜帶方便，並具備自動引導的功能，克服傳統型舌象採集裝置需固定頭部，需要架設大型的攝像器材，設備搬動不易，設備過於昂貴無法普及，操作困難的缺點，可以讓舌象採集更為簡便。 本系統經由網路攝影機資料串流取像，原始影像經過光線補償之後，強化舌色與周遭皮膚判別，對舌色進行分析，將輪廓初始化，並分析舌體在水平與垂直的相對應位置，完成對舌體位置定位，依據舌體定位資訊，導引使用者在最佳的位置，進行舌象採集的動作，同時並且標示舌象進行儲存的動作，最後再對儲存的影像，做清晰度的判斷，挑選出清晰的影像，以進行後續影像的處理。 最後本研究設計之成果，可用來量化中醫師於臨床舌診時資訊並分析， 如對舌象中顏色的識別，對舌診資訊的全面綜合識別、分析、處理，未來可作為中醫科學化之輔助工具。
並列摘要
In Chinese medicine, doctors diagnose by 4 methods: Inspection, Listening and Smelling, Inquiring, Palpation. Visual inspection of the tongue is one of the main inspections of the Wang diagnosis. Visual inspection of the tongue in Chinese medicine has a long history. In Chinese medicine, the tongue is treated as a minimized body layout with every part of the tongue corresponding to each organ of the body. This research develops a real-time tongue inspection system based on image processing technologies. The tongue inspection system developed in this research is composed of a webcam and a PC. Thus it is portable and automatic. This tongue inspection system has overcomes several disadvantages traditional tongue inspection systems have such as the need to fix the head, large image capturing equipment, ponderosity, high equipment cost, and difficult operations. This innovative tongue inspection system facilitates the visual inspection of the tongue. This tongue inspection system captures tongue images with a stream webcam. The raw images first go through light-compensation processing, strengthening the contrast of the tongue against surrounding skins. The tongue skin is further analyzed to roughly determine the outlines of the tongue. Meanwhile, the system determines the tongue position by analyzing the tongue’s relative horizontal and vertical positions. Based on the position of the tongue, the system guides the examinee to the optical position and captures the tongue images. At the same time, the system saves the tongue images with clear notification messages displayed on the screen. Finally, the system chooses the tongue images with good image quality for later processing. The system developed in this research quantitatively analyze in real time the tongue diagnosis information a Chinese medical doctor uses, such as tongue color differentiation, the distinguishing, analysis, and processing of integrated tongue inspection information. This tongue inspection system may be a great aide to scientific analysis of the Chinese medicine.