摘要
为了提高非接触式心率检测的准确性,提出用PTZ相机检测心率。由PTZ相机巡航捕捉人脸,通过皮肤检测得到皮肤像素点区域。用DIS光流法精确化提取PPG信号,由PCA降维与傅里叶变换得到心率频域图。在交互界面显示数据。实验中被试者分别处于静止和运动状态,在不同机脸距离下检测,同时由血压仪测得同步心率。计算两状态各自的平均绝对误差、误差标准差和均方根误差均较小。且B-A(Bland-Altman)图里数据点均位于95%置信区间内,区间范围小,表明实验系统的准确性高。
In order to improve the accuracy of non-contact heart rate detection,this paper proposed a method to detect heart rate by using a PTZ camera.The PTZ camera cruised to capture the face and got the skin pixel area through skin detection.The PPG signal was accurately extracted by DIS optical flow method,and the heart rate frequency domain diagram was obtained by PCA dimension reduction and Fourier transform.The data was displayed in the interactive interface.In the experiment,the subjects were at rest and in motion respectively,and the heart rate was measured synchronously by blood pressure meter at different face distances.The mean absolute error,error standard deviation and root-mean-square error of the two states were calculated respectively,which were smaller.Moreover,the data points in the Bland-Altman(B-A)diagram were all located within the 95%confidence interval,with a small interval range,indicating that the experimental system was of high accuracy.
作者
高伟翔
李擎
杨玉竹
Gao Weixiang;Li Qing;Yang Yuzhu(School of Automation,Beijing Information Science and Technology University,Beijing 100192,China;School of Automation,University of Science and Technology Beijing,Beijing 100083,China)
出处
《计算机应用与软件》
北大核心
2023年第9期43-50,98,共9页
Computer Applications and Software
基金
国家自然科学基金项目(61971048)。
关键词
PTZ相机
非接触
心率检测
PPG信号
PCA降维
PTZ camera
Non-contact
Heart rate detection
PPG signal
PCA dimensional reduction
