Combination of Denoising Algorithms for Video-Based Non-contact Heart Rate Measurement

Abstract

Denoising is very important for signal processing of non-contact heart rate measurement. Most studies of video-based heart rate measurement usually use a single denoising algorithm such as principal component analysis (PCA) or independent component analysis (ICA) to reduce noise in original signal. Using only one single denoising algorithm usually can not achieve satisfactory results, because each denoising algorithm has its shortcomings. To deal with this problem, this study has proposed to combine the following three denoising algorithms of wavelet transform (WT), PCA and ICA in different ways for noise removal in the original heart rate signal. The denoising performance is evaluated by root mean square error (RMSE) and mean absolute error (MAE). The experimental results have shown that the measurement accuracy of combined algorithms is higher than that of single algorithm. The proposed combined denoising algorithm in this study can improve accuracy and enhance stability of heart rate measurement, which is of practical significance for the application of non-contact heart rate measurement technology.