Research On Key Points Recognition For Pulse Waveform

The pulse wave is a propagating wave. Its source is the cardiac ejection. The blood strikepercutaneous vascular wall. Pulse signal can reflect some of the physiological andpathological information of human body. So the pulse signal is always a concern and studythe important medical signals. In recent years, with the development of sensors, measuringand preservation of pulse signal has become extremely convenient. Noninvasive derivation ofarterial blood pressure and calculation of various physiology becomes a very important aspectof clinical research. In this process, the feature points of pulse analysis and recognition is thederivation of the premise. In accordance with blood flow in blood vessels of the process, thepulse wave is divided into seven feature points. The feature points recognition is a hotresearch, but also a difficult.The pulse wave shape reflect all aspects of human conditions. Pulse wave shape is notthe same from different people or different physical conditions. It is due to the uncertainty ofpulse signal. The pulse wave feature recognition faces a lot of difficulties.According to the characteristics of pulse wave point identification issues, the traditionalrecognition method using differential method for dividing cycle is not accurate. And featurepoints are inflection points. So the differential method cannot accurately identify thesefeature points location. In view of the above problems, this paper proposes a hybrid algorithmfor identifying feature points. It combining two differential and wavelet transform method.The wavelet transform used to divide wave cycle. Then we can find the reference point c.Using two points of c as a pulse cycle, identification of feature points can be done throughquadratic differential method. The experimental results show that, compared with thetraditional method, the algorithm of recognition accuracy has been improved obviously.In the study there is another kind of phenomenon. Because of the diversity of pulsewave, part of pulse wave tidal wave is not obvious. The use of traditional methods it isdifficult to determine its location. Based on the pulse wave in the frequency domaincomponent, the frequency domain of tidal wave is found that its frequency is about5Hz. Anduse of wavelet transform multi characteristics, the pulse wave is decomposed into a pluralityof frequency components. Then using of different order of different frequency range ofcompositions, we calculate the different order of the wavelet coefficients. Then we identifythe original pulse wave of the feature points using the subtraction. Through of study, we findthe maximum value after second maxima corresponding to pulse wave tidal wave peak insingle cycle. Experimental results show that, this method can identify the obvious point e andnot obvious efficiently.At the same time, this paper research the application. Its application mainly focus ontwo typical parts. The pulse wave velocity (PWV) can be accurately calculate using pulsewave feature point. In clinical, judgment of human atherosclerosis is an important index for diagnosis, making it more accurately. Another important application is noninvasive derivationof the central artery pressure. The central arterial pressure as a clinically important index, ismore value than the traditional brachial artery blood pressure. In this paper, using the pulsewave characteristic point position, we can determine the central arterial pressure and thehighest point of the rising point. From this, we create feasible data preparation fornoninvasive derivation of central arterial pressure.