Study Of Process Method For Man's Pulse Signal Based On Wavelet Analysis And Support Vector Machine

Modern medicine considers that there are abundant bodies physiological, pathological and other information in pulse signal of Chinese machine, which is one of the important window for transferring and looking into function change in human's body. All of the sick phenomenon will reflect in pulse signal. So, it is an important way for Chinese medicine,physical examination and health care to detect sickness early by analyzing pulse signal.Some healthiness and hypertensive was the object of study in this paper. The formative mechanism,influence factor,basic property,familiar method of detection and time/frequency character of signal were systemically researched and analyzed. The method of suppressing noise in pulse signal based on wavelet transform was studied. The characters of envelop spectrum of each layer of detail information based on wavelet decompose was analyzed. The time-scale character of pulse signal were extracted as the character for classify. The classified method based on SVM for identifying pulse pattern was studied. Using MATLAB, the noise in signal was suppressed, the characters were extracted, the classified model was built, and the pattern identification was carried out. Based on these studies, a miniaturized system for extracting character and identifying pattern was constructed based on VI and HC2000C pulse sensor. The system was proved in having high power of identifying pattern by test and analyzing using hypertensive signal.A miniaturized system for extracting character and identifying pattern was constructed based on wavelet and SVM in this paper. The result of study has very high practical value in studying detecting sick early based on pulse signal of Chinese machine. The method of study can be used in detecting other sick early; the result of study has important practical value and can be extend applied.