Ivrr-PPG: An Illumination Variation Robust Remote-PPG Algorithm for Monitoring Heart Rate of Drivers
Ming Xu, Guang Zeng, Yongjun Song, Yue Cao, Zeyi Liu, Xiao He
Abstract
Heart rate can be accurately measured by Remote PhotoPlethysmoGraphy (rPPG) in stable illumination and noise-controlled environments such as indoors. However, heart rate measure in automotive scene is still faced with several challenges since the lighting illumination in the cab often changes dramatically. To address such issues, a novel method named Illumination Variation Robust Remote-PhotoPlethysmoGraphy (Ivrr-PPG) is proposed for monitoring the heart rate of drivers during on-road driving. The prototype is based on a commercial off-the-shelf imaging system, which includes a monochrome camera with 940 nm narrowband NIR filter and two 940 nm NIR active light sources to collect real-time video of drivers. Based on a dichromatic model, the core idea of our method is to model the non-linear relationship between illumination changing and the rPPG signal. This paper also proposed to use a background region to sense the illumination changing and consequently suppress its impact on the pulse signal. A singletone sinusoidal signal-based frequency estimation method is then used to compute the heart rate. Experimental results conducted on four different on-road driving scenarios show that the Ivrr-PPG significantly outperforms the state-of-the-art single-channel narrowband NIR algorithms.