Comprehensive comparison study on different ICA/BSS methods in IPPG techniques for obtaining high-quality BVP signal

Abstract

Independent Component Analysis (ICA) has been proven to be appropriate for extracting blood volume pulse(BVP) signal in Imaging Photoplethysmography (IPPG) techniques. However, related researches commonly employed joint approximate diagonalization of eigenmatrices (JADE) algorithm, with seldom further exploration. In this article, we took comprehensive comparison studies on different ICA or blind source separation (BSS) methods in IPPG techniques from aspects of separation effect, waveform of BVP signal, and CPU time. The experimental results indicate that, Second-Order Blind Identification (SOBI) algorithm has a better performance in BVP extraction, compared to JADE and other ICA algorithms. This study might provide a promising IPPG approach for obtaining high-quality BVP signal, which is significant for non-contact physiological measurements.