Pulse-PPG: An Open-Source Field-Trained PPG Foundation Model for Wearable Applications Across Lab and Field Settings

2025-02-03

AI Summary

Existing photoplethysmography (PPG) models exhibit poor generalization in real-world, unconstrained settings due to motion artifacts, low signal-to-noise ratios, and device heterogeneity. To address this, we introduce the first open-source PPG foundation model trained exclusively on 100 days of raw, in-the-wild PPG data from 120 participants—eliminating reliance on clinical datasets. Methodologically, we employ self-supervised pretraining (via contrastive learning and masked reconstruction), end-to-end temporal modeling, rigorous multi-source in-the-wild data cleaning and alignment, and establish a cross-domain transfer evaluation framework. Experiments demonstrate that our model consistently outperforms clinical-data-trained state-of-the-art models on downstream tasks—including heart rate and blood oxygen saturation estimation—while exhibiting significantly enhanced robustness across diverse environments, devices, and demographic groups. We fully open-source the code, pretrained weights, and training protocols to foster reproducible and equitable foundational research in PPG.

Technology Category

• Computer Vision: Other Foundations of Computer Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Other Foundations of Cognitive Modeling & Systems

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Photoplethysmography (PPG)-based foundation models are gaining traction due to the widespread use of PPG in biosignal monitoring and their potential to generalize across diverse health applications. In this paper, we introduce Pulse-PPG, the first open-source PPG foundation model trained exclusively on raw PPG data collected over a 100-day field study with 120 participants. Existing PPG foundation models are either open-source but trained on clinical data or closed-source, limiting their applicability in real-world settings. We evaluate Pulse-PPG across multiple datasets and downstream tasks, comparing its performance against a state-of-the-art foundation model trained on clinical data. Our results demonstrate that Pulse-PPG, trained on uncurated field data, exhibits superior generalization across clinical and mobile health applications in both lab and field settings. This suggests that exposure to real-world variability enables the model to learn fine-grained representations, making it more adaptable across tasks. Furthermore, pre-training on field data surprisingly outperforms its pre-training on clinical data in many tasks, reinforcing the importance of training on real-world, diverse datasets. To encourage further advancements in robust foundation models leveraging field data, we plan to release Pulse-PPG, providing researchers with a powerful resource for developing more generalizable PPG-based models.

Problem

• PPG Model
• Generalization Ability
• Real-world Application

Innovation

• Pulse-PPG
• Real-world Data
• Generalization Performance