Heart-Rate Evaluation Using Remote Photoplethysmography – A Case Study

Abstract

Heart rate of a person is a very crucial parameter about a person’s affective state. It directly is related to a person’s well-being. Although, several devices such as smart watches, portable heart-rate monitor etc. are available in the market which can be used to monitor heart-rate on a regular basis and have the capability to send alerts if any unexpected curve is observed in the pattern but these devices usually require a contact with the skin to measure the heart-rate and are expensive. Since there is an increasing widespread use of camera equipped devices in our daily lives, there are immense prospects to employ remote photoplethysmography (rPPG). A camera can be used to perceive variations in reflection and absorption of light from skin that is otherwise not noticeable with naked eye. Camera based remote photoplethysmography assists non-contact and low cost vascular activity monitoring. It is very important to manage the aspects such as illumination, selection of proper region of interest, correct estimation of signal and escalation of signal to noise ratio to remove any unwanted disturbances in order to get accurate results. In this paper, steps involved in remote photoplethysmography and the implementation of these steps are discussed.

Keywords: OpenCV, rPPG, Machine learning, plethysmographic signal, Affine Transformation, PCA, ICA, measurements, photoplethysmography, autonomous