可攜式脈診儀之自動化與脈象辨識之研究

劉省宏
元培科技大學
本計畫之目的在延續前兩年三軸脈診儀自動化之研究，使其能縮小體積方便攜帶，同時能減少測量時間，配何所發展的分類模型，將脈學中的28脈象分類。基本上將此計畫分為三個部份來進行，(1)三軸脈診儀的縮小化，使其能直接配帶於手腕部，降低受測者晃動的誤差，並建立相關的人機界面；(2)建構影像處理之尋脈位，和脈波的時域分割，萃取出脈脊之微調參數，與Z軸的氣體加壓系統；(3)建構「自我組織的模糊類神經網路」，將脈診儀的脈象分類出28脈。如此可設計出一具有客觀測量的脈診儀，再配合臨床疾病診斷與脈象分析，使脈診儀具有自動診斷功能，視窗畫面符合使用者操作的人機界面。本計畫擬先以螢光定位脈位，透過影像處理計算出感測器與脈位的相對位置，以模糊控制驅動感測器定位，再以脈博體積最大變化量為參數，透過模糊邏輯的推論機構作正確定位，Z軸的氣體加壓系統，是以脈波調變技術控制加壓幫浦和洩氣磁閥。最後是28脈象的分類，由於分類的最重要因素是資料量的多寡，且脈象本來就是一種意念的描述，因此自我組織的模糊類神經網路具有自我學習的能力，隨著測量的數量增加，亦以模糊推論修正類神經網路的歸屬函數，而達到最佳的分類效能。在臨床測試上，與子計畫一做正常者和特殊患者測量，以對本子計畫所設計製作之脈診儀做信度和穩度之校準。
關鍵字：可攜式；脈診儀；模糊控制；脈象辨識系統

A Protable Automatic Sphygmography and Pattern Recognition of Pulse Waveform

shing-hung liou
Yuanpei University
The goal is to continu the prestudies about ？H？±Designing an Automatic Triaxial Platform Control of Sphygmography？H？H. It will minize the size of sphygmography and be easily portable. The second goal is to reduce the measurement time. The final is to develop a classification for the 28 sphygmogram. Therefore, there are three parts in this project under three years. (1) The triaxial sphygmography will be minized and put under the wrist. With this device, the artifical noise could be decreased. (2) The image process will be used to search the radial pulse. The futures to correct mediam position are exacted form the temporal signal. The control of the Z axial is an air pumping system. (3) A self organizing fuzzy neural network is used to classify the 28 sphygmogram. With these methods, we can desige a impersonal sphygmograph that has a powerful user interface and can diagnose the diseases. In this project, the first phase is marking the radial pulse position and calculating the relative distance between the marker and sensor. The fuzzy control is used to drive the sensor to the correct position. The pulse waveform modulation is used to contol the pumping system at the Z axial.
關鍵字：Portable；Sphygmography；Fuzzy Control；Pulse Recognition