Heart rate estimation using remote photoplethysmography with multi-objective optimization. (March 2019)

• Record Type: Journal Article
• Authors: Macwan, Richard; Benezeth, Yannick; Mansouri, Alamin
• Is Part Of: Biomedical signal processing and control. Volume 49(2019)
• Page Start: 24
• Page End: 33
• Publication Date: 2019-03
• Subjects: Remote photoplethysmography -- Multi-objective optimization -- Semi-blind source separation -- Independent Component Analysis -- Constrained Independent Component Analysis -- Remote heart rate measurement; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2018.10.012
• Languages: English
• ISSNs: 1746-8094

Abstract

Abstract: Remote photoplethysmography (rPPG) is being increasingly used to measure heart rate from recorded or live videos. The rhythmic flow of arterial blood, referred to as the blood volume pulse, results in periodic variations in the skin color which are then quantified into a temporal signal for analysis. Independent Component Analysis (ICA) has been used to extract the blood volume pulse which is assumed to have been mixed into the RGB channels of the skin pixels. We propose a novel semi-blind source extraction method for measuring rPPG using a multi-objective optimization approach with autocorrelation as a periodicity measure. Our method was tested on our inhouse video database UBFC-RPPG and the MMSE-HR database[33] . Our in-house database is made publicly available and is specifically aimed towards testing rPPG measurements. Our method showed improved performance over other state of the art rPPG algorithms in terms of accuracy with all the databases.