New Advances in Remote Heart Rate Estimation and Its Application to DeepFake Detection

Abstract

Estimation and monitoring of heart rate is an important indicator of a person’s physiological and psychological status. With breakthroughs in sensing, real-time signal processing and machine learning technologies, heart rate measurement techniques based on video analysis have been derived, remote photo plethysmography (rPPG), a technique that captures periodic changes in skin color caused by the heartbeat cycle through sensors such as cameras. Heart rate measurement based on rPPG has attracted widespread attention in recent years, especially with the growing maturity of deep learning-based methods and the improved accuracy of the algorithms, as well as the ability to better cope with the adverse effects of lighting changes and motion artifacts on the measurement, which has made possible the applications of rPPG in more fields, and a prominent emerging application is DeepFake videos detection based on rPPG technology. In this paper, we will introduce the principle, the recent progress of rPPG algorithms, and the applications of rPPG in the field of DeepFake detection.