An efficient non-contact heart rate measurement based on CIELab color space

Remote photoplethysmography (rPPG) provides an approach of non-contact heart rate measurement, which can be used in natural light conditions to record human face video using an ordinary camera. However, this kind of experiment is currently sensitive to illumination, movement and skin color. In this paper, we introduce a simple and robust heart rate signal extraction method, by converting to CIELab color space as well as using a* channel to extract the original signal, rather than commonly used RGB space and g* channel. We utilized MAHNOB-HCI-TAGGING Database and self-built dataset to evaluate the results. Compared with g* channels, using a* channel in CIELab color space achieved higher accuracy, with the error rate reduced by 4.85%. The design has the characteristics of fast and convenient operation, with low cost and non-contact feature, that are suitable for a wider monitoring heart rate in daily life.