A Browser-Based Remote Photoplethysmography Method for Live Heart Rate Estimation
Year: 2026
Type: preprint
Abstract: Remote photoplethysmography (rPPG) enables contact-free heart-rate estimation from facial video, but most web-based implementations process recordings offline, limiting their use in real-time experimental applications. We developed and, to our knowledge, provide the first ECG validation of a live rPPG implementation embedded directly within Gorilla Experiment Builder, a widely used platform for online behavioural research. We validated the method in 27 participants against simultaneously recorded electrocardiography (ECG) across nine trial durations ranging from 10 to 50 s. rPPG-derived heart rate showed excellent agreement with ECG-derived heart rate, with a mean absolute error of 1.48 bpm, a root mean square error of 2.19 bpm, and an ICC(A,1) of .960. Mean bias was small (0.40 bpm), with repeated-measures 95% limits of agreement from -3.85 to 4.66 bpm, and heart-rate error did not differ significantly by trial durations. These findings support the feasibility of live rPPG within a browser-based experimental platform for physiological feedback and adaptive online experiments without requiring webcam videos to be stored or processed offline.
Source: PsyArXiv (OSF Preprints)
Language: English
Subfield: Biomedical Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7214427211