关键词: 非接触式, 心率测量, Mamba模型, 远程光电容积描记术, RGB面部视频

Abstract: Non-contact heart rate measurement methods based on remote PhotoPlethysmoGraphy (rPPG) technology faced two main challenges: the accurate selection of facial regions of interest, and the precise interpretation of periodic patterns in rPPG from long video sequences. To address the inaccurate localization of facial regions of interest in existing works and the neglect of temporal lag among different signal sources, an end-to-end rPPG signal extraction model named RppgMamba was proposed. A hierarchical spatial-temporal Mamba model with cross-scan mechanism was designed to capture spatial perception relationships of weak signals from multi-scale spatial-temporal receptive fields. This module adaptively localized facial regions of interest and avoided interference from background and motion noise. A frequency-domain refinement module was further designed to learn the quasi-periodic patterns of rPPG signals in the frequency domain, which enhanced output signal quality. A phase compensator was incorporated to correct phase shifts, mitigating the impact of inherent temporal lags. Experimental results demonstrate that the proposed model achieves state-of-the-art performance on three public datasets, confirming its high accuracy and applicability in heart rate measurement.

Key words: non-contact, heart rate measurement, Mamba model, remote photoplethysmography, RGB facial video