Efficient Approach for Measuring Heart Pulse from Real Time Facial RGB Color Videos

Abstract

The heart, the vital organ of the human body, needs to function with proper care, and therefore it is important to monitor. Detecting Heart Rate (HR) often requires costly equipment which is complex especially with sensors. Various methods are defined over the years to measure heart rate. One of them being a non-contact method of measuring heart rate using remote photoplethysmography (rPPG). This paper presents a working solution on the live video i.e., in real-time. A set of rectangles are established around the face for detection along with object detection, which is later used in the pipeline of color variations. The mean of the colors (RGB-red, green, blue) in the Region of Interest (ROI) selected on the face (forehead region). Thus, the signal is obtained from the heart rate frequency by the signal processing method. This paper discusses the reimplementation of one such approach that uses Fast Fourier Transform (FFT). Researchers can now refer to the reviewed and classified algorithms as a starting point to improve and increase the robustness of the experiment.