A Non-contact Heart Rate Estimation Framework Based on Photoplethysmography Amplitude Variation Elimination and Data Fusion

Non-contact heart rate (HR) measurement is gaining attention due to its advantages over contact-based approaches. This paper presents a robust framework for estimating HR from facial videos by exploiting multiple color space representations and photoplethysmography (PPG) extraction techniques. A novel PPG Amplitude Variation Elimination Technique (AVET) based on analytic signal representation is introduced to equalize amplitude fluctuations. Multiple HR estimates are fused to reject outliers and improve robustness. Experimental evaluation on the UBFC-Phys dataset demonstrates superior performance compared to baseline ICA and Green-channel methods.

2021 28th National and 6th International Iranian Conference on Biomedical Engineering (ICBME)