BLOOD PRESSURE PREDICTION METHOD AND APPARATUS BASED ON PHOTOPLETHYSMOGRAPHY SIGNAL

摘要

A blood pressure prediction method and apparatus based on a photoplethysmography signal. The method comprises: generating PPG signal segments (1); according to an acquired wavelet basis type, contraction-expansion factors and movement factors, performing signal decomposition on the PPG signal segments by using a continuous wavelet transform means, so as to generate a PPG wavelet coefficient matrix (3); converting the PPG wavelet coefficient matrix by means of performing modulo operation on matrix elements, and performing normalization processing on the converted matrix to generate a PPG normalized matrix (4); acquiring an RGB color disk matrix, and according to the RGB color disk matrix, performing tensor conversion on the PPG normalized matrix to generate a PPG time-frequency three-dimensional tensor (5); according to a convolutional network input width threshold, performing tensor shape reconstruction on the PPG time-frequency three-dimensional tensor by using a bicubic interpolation algorithm, so as to generate a PPG convolutional three-dimensional tensor (6); and performing, by using a convolutional neural network classification and regression model, classification and regression calculation on the PPG convolutional three-dimensional tensor, so as to generate a PPG predicted blood pressure data pair (7).