Estimating vital signs through non-contact video-based approaches: A survey

Abstract

The human body exhibits many vital signs, such as heart rate (HR) and respiratory rate (RR) used to assess fitness and health. Vital signs are typically measured by a trained health professional and may be difficult for individuals to accurately measure at home. Clinic visits are therefore needed with associated burdens of cost and time spent waiting in long queues. The widespread use of smart phones with video capability presents an opportunity to create non-invasive applications for assessment of vital signs. Over the past decade, several researchers have worked on assessing vital signs from video, including HR, RR and other parameters such as anemia and blood oxygen saturation (SpO2). This paper reviews the different image and video processing algorithms developed for vital signs assessment through non-contact methods, and outline the key remaining challenges in the field which can be used as potential research topics. The CHROM algorithm produces highest accuracy in detecting the signals from rPPG. There are different challenges of handling large database and motion stabilization which is not provided by any algorithm, this is main area of research in rPPG.