生命科学仪器2023，Vol.21Issue(4)：40-45,6.DOI:10.11967/2023210806
基于智能手机PPG信号的血压估测方法
Estimation method of blood pressure based on PPG signal of smart phone
摘要
Abstract
Due to the limited application scenarios of traditional sphygmomanometer,a blood pressure estimation method based on smart phone Photoplethysmography(PPG)signal is proposed in this paper to estimate blood pres-sure more flexibly and conveniently.The smart phone is used to collect the user's finger video and convert it into PPG signal.The one-dimensional U-net network is used to de-noise and de-baseline the PPG signal.The conv-olutional neural network with residual module is used to classify the signal quality after processing the signal,and then the feature extraction of the high-quality PPG signal obtained by classification is carried out.Finally,the sig-nal characteristics were combined with the physiological characteristics of the user(height,weight,age and gen-der),and the regression model of the characteristics and blood pressure was established by random forest algorithm to complete the estimation of blood pressure.Data were collected from 83 subjects.The mean error of blood pressure estimates from the blood pressure model was:systolic blood pressure(SBP)-0.62±3.29mmHg and diastolic blood pressure(DBP)0.34±3.67mmHg in the test set of 17 subjects.This study proves that PPG signals extracted from smart phones have great value in blood pressure estimation to a certain extent,and also provides a technical reference for the research direction of blood pressure estimation by smart phones.关键词
智能手机/PPG信号/信号处理/机器学习/血压估测Key words
smartphones/PPG signal/signal processing/machine learning/blood pressure estimation分类
医药卫生引用本文复制引用
李映辉,唐晓英..基于智能手机PPG信号的血压估测方法[J].生命科学仪器,2023,21(4):40-45,6.基金项目
国家重点研发计划资助(2020YFC2007301,2020YFC2007300) （2020YFC2007301,2020YFC2007300）
