A BVP Signal Detection Method of Anti-Motion Interference

The noncontact detection methods of blood volume pulse (BVP) based on facial videos have become a hot spot in recent years. However, these kinds of methods are highly sensitive to face movement. To address this problem, a novel BVP detection method based on kanade-Iucas-tomasi (KLT) and independent component analysis (ICA) was proposed, in which, KLT method was employed for tracking and locating the region of interest (ROI) in facial images, and ICA method was used to improve the signal to noise ratio (SNR) of BVP signal. Based on 120 recorded facial videos, we carried out the comparative study of the proposed method and other commonly-used methods, the experimental results show that our method has obvious advantages in eliminating motion interference. In the implementation of the blind source separation method, we also conducted four separation methods. The results show that the second order blind identification (SOBI) algorithm is the best one to separate the BVP signal.