Non-contact Measurement of Blood Volume Pulse Using 3 signals from CCD Camera

Abstract

Imaging-photoplethysmography (iPPG) is a non-contact technique to monitor blood volume pulse (BVP) via detecting tiny color changes in skin caused by blood circulation using a CCD camera. The most basic approach is to extract BVP from only green channel, however such method can be easily affected by ambient light noise and motion artifacts. In this study, we focus on improving the robustness of extracting the BVP from CCD camera, thereby accurately measuring beat-to-beat intervals for calculating heart rate variability (HRV) indices. We proposed a method to reconstruct BVP waveform using all RGB channels to optimize a linear function to improve the signal-to-noise ratio based on the cardiac physiological properties. The proposed method was evaluated on HRV indices with 4 healthy subjects comparing among the conventional green channel alone and the ground truth ECG. The results show that the proposed method is more effective for HRV measurement than the conventional method.