Feature extraction and recognition of traditional Chinese medicine pulse based on hemodynamic principles

In this paper, factors contributing to the formation of pulse wave were analyzed based on hemodynamic principles. It is considered that formation of pulse wave was related to its propagation and reflection characteristics. Propagation of the pulse wave was characterized by pulse wave velocity, and reflection of the pulse wave was characterized by reflection coefficient. Pulse wave velocity and reflection coefficient were proposed as the eigenvectors of pulse wave in pulse diagnosis of traditional Chinese medicine, and support vector machine (SVM) was used to recognize slippery pulse, stringy pulse and plain pulse. Pulse wave velocity and reflection coefficient of the slippery, stringy and plain pulses in healthy people were calculated in this study, and SVM with Gaussian radial basis function was used for classifying. Results showed that pulse wave velocity and reflection coefficient with physiological and pathological significance had advantages in distinguishing slippery pulse, stringy pulse and plain pulse, which offered a new idea for recognizing pulse condition.