FLOW: Feature-Level Optimal Warping for Generalized Remote Physiological Measurement

2026-09-30

AI Summary

This study addresses domain shift in remote photoplethysmography (rPPG) caused by illumination variations, motion artifacts, and sensor discrepancies. We propose FLOW, a lightweight cross-domain adaptation framework based on optimal transport. FLOW introduces a novel prototype-based cross-temporal optimal transport mechanism that integrates temporal refinement with a learnable prototype alignment module. By modeling soft cross-temporal correspondences, it effectively preserves the intrinsic rhythms of physiological signals. Furthermore, we derive a theoretical generalization bound under conditional optimal transport. Experimental results demonstrate that FLOW achieves state-of-the-art cross-domain performance across four benchmark datasets, combining a lightweight architecture with high physiological fidelity.

Abstract

Remote photoplethysmography (rPPG) enables non-contact physiological measurement but remains vulnerable to domain shifts from illumination, motion, and sensors. We propose \textbf{FLOW (Feature-Level Optimal Warping)}, an \emph{optimal transport--driven} framework for domain-generalized rPPG. FLOW integrates a \textbf{Temporal Refinement Module (TRM)} to stabilize temporal dynamics and a \textbf{Prototype-based Cross-Temporal Optimal Transport (PCOT)} module to achieve domain-invariant alignment via learnable prototypes.Beyond feature alignment, FLOW employs soft cross-temporal correspondence modeling that aligns temporal features in a flexible manner, allowing the model to respect and preserve the intrinsic rhythmic patterns of physiological signals. Moreover, the lightweight design of our modules allows seamless integration into existing end-to-end rPPG architectures without additional preprocessing. Two regularization terms further enforce source consistency and identity preservation. Theoretically, we derive a generalization bound under conditional optimal transport. Extensive experiments across four rPPG benchmarks show that FLOW achieves state-of-the-art cross-domain performance with lightweight design and strong physiological fidelity.

Problem

• Remote photoplethysmography
• Domain generalization
• Domain shift
• Physiological measurement

Innovation

• Optimal Transport
• Remote Photoplethysmography
• Domain Generalization
• Cross-Temporal Alignment
• Learnable Prototypes