Research On Cardiac Pulse Measurement Using Facial Videos

Heart rate is one of the most important physiological parameters. It is an important standard to human health, and also an important parameter of cardiovascular disease. People can use electrocardiogram and pulse oximetry to measure their heart rate in medicine. However, this kind of measurement cannot provide real-time monitoring because people should take the initiative to measure first. At the same time, this contact measurement will make people uncomfortable. Photoplethysmography, which usually requires dedicate lights, is a low cost and non-invasive means of measuring heart rate. Researchers has shown that pulse measurements can be acquired using digital cameras with normal ambient light as the illumination source. It makes low cost and non-contact pulse measurement possible.This paper discusses a non-contact optical pulse measurement which is based on photoplethysmography technique by analyzing a facial video using computer vision, independent component analysis and signal processing etc.. We extracted the original pulse signal from a facial video by using photoplethysmography. After detrending method and independent component analysis, we could get heart rate from the power spectra of pulse wave.To verify the effectiveness of the cardiac pulse measurement method as discussed in this paper, we compared the cardiac pulse rate extracted from facial videos recorded by iPhone5 to Samsung Galaxy S5 heart rate sensor by using Bland-Altman. Experimental results showed that this contact-free method is feasible and effective. It has certain practical value and application prospect in daily care, human interaction and telemedicine.