PhysMamba: State Space Duality Model for Remote Physiological Measurement

2024-08-02

arXiv.org

AI Summary

To address physiological signal degradation in remote photoplethysmography (rPPG) caused by facial motion, illumination variations, and noise, this paper proposes the Synergistic State-Space Dual (SSSD) paradigm. It introduces, for the first time, a dual-path architecture that tightly couples state-space models with attention mechanisms, augmented by a Multi-scale Query (MQ) mechanism to enhance time-frequency feature interaction. The method jointly leverages spatiotemporal modeling capacity and dynamic perception capability. Evaluated on PURE, UBFC-rPPG, and MMPD benchmarks, it achieves significant improvements over state-of-the-art methods—reducing average mean absolute error (MAE) by 12.7%—while demonstrating markedly improved cross-scenario generalization and real-time inference capability. This work establishes a robust, efficient paradigm for contactless remote health monitoring.

Technology Category

• Computer Vision: Multi-modal Vision
• Intelligent Robots: State Estimation
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Personalized, context-aware and across-device search
• Security and Privacy: Large-scale security measurements

Abstract

Remote Photoplethysmography (rPPG) enables non-contact physiological signal extraction from facial videos, offering applications in psychological state analysis, medical assistance, and anti-face spoofing. However, challenges such as motion artifacts, lighting variations, and noise limit its real-world applicability. To address these issues, we propose PhysMamba, a novel dual-pathway time-frequency interaction model based on Synergistic State Space Duality (SSSD), which for the first time integrates state space models with attention mechanisms in a dual-branch framework. Combined with a Multi-Scale Query (MQ) mechanism, PhysMamba achieves efficient information exchange and enhanced feature representation, ensuring robustness under noisy and dynamic conditions. Experiments on PURE, UBFC-rPPG, and MMPD datasets demonstrate that PhysMamba outperforms state-of-the-art methods, offering superior accuracy and generalization. This work lays a strong foundation for practical applications in non-contact health monitoring, including real-time remote patient care.

Problem

• rPPG
• motion artifacts
• illumination changes

Innovation

• State Space Model
• Attention Mechanism
• Multi-scale Query Technique