Research on the combination of color channels in heart rate measurement based on photoplethysmography imaging

Significance: The measurement of human vital signs based on photoplethysmography imaging (PPGI) can be severely affected by the interference of various factors in the measurement process; therefore, a lot of complex signal processing techniques are used to remove the influence of the interference. Aim: We comprehensively analyze several methods for color channel combination in the color spaces currently used in PPGI and determine the combination method that can improve the quality of the pulse signal, which results in a modified plane-orthogonal-to-skin based method (POS). Approach: Based on the analysis of the previous studies, 13 methods for color channel combination in the different color spaces, which can be seen as having potential abilities in measuring vital signs, were compared by employing the average value of signal-to-noise ratio (SNR) and the box-plot in the public databases UBFC-RPPG and PURE. In addition, the pulse signal was extracted through the dual-color space transformation (sRGB -> intensity normalized RGB -> YCbCr) and fine-tuning on the CbCr plane. Results: Among the 13 methods for color channel combination, the signal extracted by the Cb +Cr combination in the YCbCr color space includes the most pulse information. Furthermore, the average SNR of the modified POS for all the used databases is improved by 69.3% compared to POS. Conclusions: The methods using prior knowledge are not only simple to calculate but can significantly increase the SNR, which will provide a great help in the practical use of vital sign measurements based on PPGI. (C) The Authors. Published by SPIE under a Creative Commons Attribution 4.0 Unported License.

Automated summary

This study compares several methods for color channel combination in different color spaces used in PPGI, and identifies that the Cb +Cr combination in the YCbCr color space can extract the most pulse information. Furthermore, the average signal-to-noise ratio (SNR) of the modified POS is improved by 69.3% compared to the original POS, indicating significant improvement in vital sign measurements based on PPGI.