In this paper, we propose a novel non-contact pulse wave monitoring method which is robust to fluctuation of illumination by using two-band infrared videos. Since the proposed method uses the infrared light for illumination, we can detect the pulse wave on the face without visible lighting. The corresponding two-band pixel values in the videos can be separated into hemoglobin and shading components by applying separation matrix in logarithmic space for two pixel values. Since the shading component is separated, the extracted hemoglobin will be robust to the fluctuation of illumination. The pixel values of the region of interest (ROI) were spatially averaged all over the pixels for each frame. The averaged values are used to form the raw trace signal. Finally, the pulse wave and pulse rate were obtained from raw trace signal through several signal processing such as detrend, adaptive bandpass filter, and peak detection. We evaluated the absolute error rate for pulse rate between the estimated value and the ground truth obtained by the electrocardiogram. From the experiment, we found that the performance of our method was greatly improved in comparison with the conventional method by means of one-band infrared video.
Ryota Mitsuhashi, Genki Okada, Koki Kurita, Keiichiro Kagawa, Shoji Kawahito, Chawan Koopipat, Norimichi Tsumura, "Two-Band Infrared Video-Based Measurement for Non-Contact Pulse Wave Detection on Face without Visible Lighting" in Proc. IS&T 25th Color and Imaging Conf., 2017, pp 83 - 87, https://doi.org/10.2352/ISSN.2169-2629.2017.25.83