基于人脸视频的光电容积脉搏波信号获取与处理  

Photo plethysmo graphy signal acquisition and processing based on facial video

作　　者：雷恒波 郭友达 莫镜清 徐雅洁[1] 邢晓曼[1] 孙明山[1] 

机构地区：[1]中国科学院苏州生物医学工程技术研究所,江苏苏州215163 [2]南京理工大学电子工程与光电技术学院,江苏南京210094

出　　处：《中国医学物理学杂志》2016年第11期1178-1182,共5页Chinese Journal of Medical Physics

摘　　要：光电容积脉搏波(PPG)中包含着人体重要的生理信息,它的监测可以有效地对心血管疾病进行早期预防和预后诊断。为了方便、准确地获得人体的PPG信号,本文通过手机摄像头拍摄一段人脸视频,根据每帧图像中像素灰度值的变化得到血液容积变化的时序曲线;然后,利用经验模态分解法(EMD)将它分解为一系列不同频率成分的本征模函数(IMF);最后通过频域分析,将处于脉搏波频率范围的IMF进行信号重构,就得到了消除低频基线漂移和高频噪声的PPG信号。该结果与现有医疗仪器同步测得的PPG数据相关系数在0.85以上,这种方法适用范围广,特别是对远程医疗中基于PPG的人体生理信息的获得具有重要意义。The monitoring of photo plethysmo graphy （PPG） which contains some important human physiological information is effective for the cardiovascular disease prevention and prognosis. In order to easily and accurately obtain the PPG signal of the body, the camera video of the human face is obtained by the cell phone, and the timing curve of the blood volume variation is obtained by investigating the variation of pixel gray values of each frame image. With the empirical mode decomposition method, the timing curve was broken down into series of intrinsic mode function （IMF） with different frequency components. Finally, using the frequency domain analysis, the IMF in the pulse frequency range is reconstructed, obtaining the PPG signal without the low-frequency and high-frequency noise baseline drift. The correlation coefficient between the result and the measured PPG data synchronized with the existing medical instruments is above 0.85. The PPG method based on facial video which is widely applied is important for obtaining the physiological information based on PPG in telemedicine.

关 键 词：摄像头 非接触成像 光电容积脉搏波 经验模态分解法 远程医疗 

分 类 号：R318[医药卫生—生物医学工程]