Research on Pulse Feature Point Identification Based on Differential Threshold

Accurate recognition of pulse signal characteristics is the key to disease diagnosis, the existing methods cannot adapt to the identification of sudden pulse cycles. The dynamic threshold method is used to locate all the starting point of the ascending branch and pulse cycle. The single-cycle segmentation of the entire pulse cycle can be achieved. The results show that the starting point and the peak point of all pulse data can be identified by this method. The segment of single pulse signal and recognition of feature points can also be realized through this method. The recognition accuracy of the tidal wave points and the dichotomous wave point reaches 92.6%; it can be basically completed. The identification of characteristic points provides the basis for the analysis and digitization of traditional Chinese medicine.