Video-Based Real-Time Heart Rate Detection for Drivers Inside the Cabin Using a Smartphone

Abstract

Developing vehicles with the Internet of Thing technology including driver health monitoring systems, driver safety systems, and accident prevention has drawn the attention of researchers in the last few years. The monitoring system should prevent any dangerous situation and be comfortable for the driver inside the cabin. In this paper, we introduce a remote video-based method for detecting the heart rate in real-time using smartphone cameras, which can be used for the analysis of the driver’s physiological parameters to enhance driving safety. We propose to use 3DDFA for automatic facial landmarks detection to extract the driver’s face and a 3d-classification-based model for detecting the heart rate. The experiments showed good results with mean absolute error (MAE) equal to 6.8 on the LGI-PPGI dataset and 18.68 on our DriverMVT dataset that was recorded in the wild.