Non-contact Pulse Rate Measurement of Hand and Wrist Using RGB Camera

This study selected the Region of Interest in the video to record the positions of the hand and wrist with a non-contact method, then filtered out the video noise and amplified the invisible changes through Eulerian Video Magnification and restored the pulse wave signal with the independent component analysis after the separation of the RGB channels. In the 10 sets of measurement data of 5 people, it has been found that when the spatial decomposition of Gaussian Pyramid was applied to the palm, the data of the estimated pulse rate in spectrum was closest to the data obtained from Pulse Oximeter with the correlation analysis of R-squared 0.886 and the agreement analysis of mean deviation 3.57 BPM. In additional to capturing pulse waves at different locations around the palm and wrist, the method applied in this study can estimate heart rate or pulse rate and improve the feasibility of monitoring non-contact pulse waves with RGB camera.