IPPG信号的噪声估计与分析

Estimation and Analysis of Noise in Imaging Photoplethysmography(IPPG)Signal

目的验证成像式光电容积描记法(IPPG)信号模型中的加性噪声近似相互统计独立,提出对IP-PG信号中的噪声估计方法,探究使用空间平均方法后的残余噪声功率和信噪比变化规律。方法对13组受试者数据帧图像进行空间平均并去除基线漂移后得到IPPG信号,计算所有帧图像中所有像素点噪声方差,作为对噪声功率的估计。在此基础上进一步计算残余噪声及脉搏信号功率、信噪比,利用回归分析验证IPPG噪声与感兴趣区域(ROI)的关系。结果信噪比与感兴趣区域边长呈平方关系,残余噪声与像素点个数成反比例下降。结论实验验证了IPPG信号加性噪声近似相互统计独立的假设,证明了模型与噪声估计方法的有效性,提出的噪声估计算法对于后续滤波提高信号质量具有一定的意义。

Objective To verify that the additive noise in the imaging photoplethysmography(IPPG)signal model is approximately statistically independent of each other,a noise estimation method for the IPPG signal was proposed,and the changes of residual noise power and signal-to-noise ratio were explored using the spatial averaging method.Methods IPPG signal was obtained after averaging 13 groups of frame images of subjects and removing the baseline drift.The variance of all pixel noises in all frames was calculated as an estimation of the noise power.Based on this,the residual noise,the pulse signal power and the signal-to-noise ratio were further calculated.Regression analysis was used to verify the relationship between IPPG noise and Region of Interest(ROI).Results The analysis showed that the signal-to-noise ratio was in square relationship with the side length of the region of interest,and the residual noise decreased in inverse proportion to the number of pixels.Conclusion The experiment verified that the assumption that the additive noise of the IPPG signal was approximately statistically independent of each other,and the effectiveness of the model and the noise estimation method were proved.The proposed noise estimation algorithm is of certain significance for the subsequent filtering to improve the signal quality.