DTW Based Hierarchical Filter for Pulse Waveform Classification

• Main Authors: Po-Jui Huang, 黃柏叡
• Other Authors: Duan-Yu Chen
• Format: Others
• Language: zh-TW
• Online Access: http://ndltd.ncl.edu.tw/handle/76091979376806344406

碩士 === 元智大學 === 電機工程學系 === 101 === With sharp rising population in the earth, medical personnel have to deal with more and more patients. It means each person can have relatively fewer medical resources. Faced with a situation of limited medical resources, home care will become increasingly important. In the absence of medical staff’s supervision, the patient's physiological information will be incorrect easily because of wrong way of using medical equipment, or other unstable factors. To overcome these problems, we must have a physiological information filtering mechanism to ensure the validity of measurement information in home-care system. Therefore, in this thesis, the pulse measurement of radial artery is selected as a case study since physiology of radial artery has rich information for Chinese clinical studies. To analyze the radial pulse, a hierarchical filtering mechanism is proposed, in which Dynamic Time Warping (DTW) is combined with Discrete Fourier Transform (DFT) and standard deviation analysis. Using this filtering mechanism, unusual waveform of radial pulse can be detected effectively with precision being above 90%.