Embedded Vital Signs Monitoring Device Based on Convolutional Neural Network

Heart rate and respiratory value are important indicators that doctors care about. In this paper, we propose an end-to-end measurement method of human respiration and heart rate based on convolutional neural network (CNN). Also, we transplant this AI model into STM32 chips to extract the feature of optical fiber signal. We make a data set of human signals collected in fiber optics to train the AI model. Experiments show that our AI model and embedded devices can predict people's respiration and heart rate effectively.