Heart Rate Measurement on PC and Phone using Facial Videos

Abstract

Heart rate (HR) analysis has always piqued the curiosity of medical experts. Various apps have been designed using algorithms that assess the pulse using only one’s facial video. A recently developed technique called Eulerian Video Magnification (EVM) can detect temporal fluctuations in videos that are undetected by the naked human eye. It is feasible to visualize the flow of blood filling the face with this approach. Photoplethysmography (PPG) signals from the human face can be spotted by minute variations in skin tone that are connected to the blood vessels beneath the surface of the face. The output of the signals can then be used to determine the vitals of the person. In order to estimate the heartbeat of 40 participants at the initial, post-cardio, and after-resting stages, this study employed an implementation of the EVM computer vision algorithm, developed to remotely detect an individual’s HR in beats per minute from a static video of his or her face. The data from the desktop and smartphone were compared to the readings made simultaneously by an oximeter. The pulse oximeter, which likewise derives HR by PPG, and the PPG-derived HR utilizing EVM from the desktop and the smartphone both showed positive correlations.