A Dataset and Toolkit for Multiparameter Cardiovascular Physiology Sensing on Rings

2025-05-07

AI Summary

Smart rings offer unobtrusive, continuous physiological monitoring but suffer from limited accuracy in cardiovascular parameter estimation due to the absence of public benchmark datasets and robust algorithms. To address this, we introduce τ-Ring—the first open-source smart ring physiological signal dataset—comprising synchronized infrared/red photoplethysmography (PPG) and triaxial accelerometer data across resting, dynamic, and pathological conditions, with ground-truth labels for heart rate (HR), respiratory rate (RR), peripheral capillary oxygen saturation (SpO₂), and systolic/diastolic blood pressure (SBP/DBP). We also release RingTool, a standardized analysis framework for multi-scenario evaluation. Methodologically, we propose a dual-loop optical acquisition architecture integrating reflective and transmissive PPG pathways, and jointly leverage biophysical models (e.g., pulse transit time) with advanced deep learning architectures (1D-CNN, Transformer). Experiments demonstrate state-of-the-art performance: mean absolute errors of 5.18 BPM (HR), 2.98 BPM (RR), 3.22% (SpO₂), and 13.33/7.56 mmHg (SBP/DBP)—significantly outperforming leading commercial wearables.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Humans and AI: Brain-Sensing and Analysis
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

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

Abstract

Smart rings offer a convenient way to continuously and unobtrusively monitor cardiovascular physiological signals. However, a gap remains between the ring hardware and reliable methods for estimating cardiovascular parameters, partly due to the lack of publicly available datasets and standardized analysis tools. In this work, we present $ au$-Ring, the first open-source ring-based dataset designed for cardiovascular physiological sensing. The dataset comprises photoplethysmography signals (infrared and red channels) and 3-axis accelerometer data collected from two rings (reflective and transmissive optical paths), with 28.21 hours of raw data from 34 subjects across seven activities. $ au$-Ring encompasses both stationary and motion scenarios, as well as stimulus-evoked abnormal physiological states, annotated with four ground-truth labels: heart rate, respiratory rate, oxygen saturation, and blood pressure. Using our proposed RingTool toolkit, we evaluated three widely-used physics-based methods and four cutting-edge deep learning approaches. Our results show superior performance compared to commercial rings, achieving best MAE values of 5.18 BPM for heart rate, 2.98 BPM for respiratory rate, 3.22% for oxygen saturation, and 13.33/7.56 mmHg for systolic/diastolic blood pressure estimation. The open-sourced dataset and toolkit aim to foster further research and community-driven advances in ring-based cardiovascular health sensing.

Problem

• Lack of public datasets for ring-based cardiovascular sensing
• Absence of standardized tools for cardiovascular parameter estimation
• Need for reliable methods in motion and abnormal physiological states

Innovation

• Open-source dataset for ring-based cardiovascular sensing
• Toolkit evaluates physics-based and deep learning methods
• Superior performance in multi-parameter physiological estimation