PPG blood pressure prediction technology based on convolutional neural network during wartime

This article proposes a lightweight residual network model PPG light by obtaining and preprocessing PPG signals through equipment. This network is based on the UNet network model, using specialized 1D segmentation and combining MobileNet's lightweight and moderate depth with the UNet network to transform the two-dimensional model into a one-dimensional model. The input value is the processed PPG signal, and the output is an approximate blood pressure waveform. Experimental results show that this network model has a significant lightweight effect.