A Dataset and Toolkit for Multiparameter Cardiovascular Physiology Sensing on Rings Volume: 1 1 1 5 CCS: Human-centered computing Ubiquitous and mobile computing

Jiankai Tang, Kegang Wang, Yingke Ding, Jiatong Ji, Zeyu Wang, Xiyuxing Zhang, Ping Chen, Yuanchun Shi, Yuntao Wang · 2025

Read A Dataset and Toolkit for Multiparameter Cardiovascular Physiology Sensing on Rings Volume: 1 1 1 5 CCS: Human-centered computing Ubiquitous and mobile computing by Jiankai Tang, Kegang Wang, Yingke Ding, Jiatong Ji, Zeyu Wang, Xiyuxing Zhang, Ping Chen, Yuanchun Shi, Yuntao Wang (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-04172.txt.

Introduction

Jiankai Tang and co-authors share a ring sensor data set for heart health in a 2025 paper. Smart rings can track body signals all day with ease, but open data and shared tools for ring based heart gauges lag behind the gear. Their set, tau-Ring, is the first open one of its kind.

Two ring forms, one bounce based and one pass through, logged light and motion data from 34 people across seven tasks, in still and moving states plus stirred abnormal states. True labels cover heart rate, breath rate, oxygen level, and blood pressure. Their RingTool kit then graded physics and deep learning methods, beating shop rings.

Catalog record

CREATORJiankai Tang, Kegang Wang, Yingke Ding, Jiatong Ji, Zeyu Wang, Xiyuxing Zhang, Ping Chen, Yuanchun Shi, Yuntao Wang
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE80 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-04172

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A Dataset and Toolkit for Multiparameter Cardiovascular Physiology Sensing on Rings Volume: 1 1 1 5 CCS: Human-centered computing Ubiquitous and mobile computing Jiankai Tang, Kegang Wang, Yingke Ding, Jiatong Ji, Zeyu Wang, Xiyuxing Zhang, Ping Chen, Yuanchun Shi, Yuntao Wang 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 τ \tau -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. τ \tau -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. 1. INTRODUCTION Smart rings have quickly grown in popularity due to their compact size, attractive design, and convenience, driving a significant increase in market demand ( Liang et al. 2021 ; Yu et al. 2024 ) . The global smart ring market was valued at $340.9 million dollars in 2024 and is projected to expand substantially, reaching $2,525.5 million by 2032 with a compound annual growth rate of 29.3% 1 1 1 https://www.fortunebusinessinsights.com/smart-ring-market-111418 . Commercial products like the Oura Ring 2 2 2 https://ouraring.com/ and Samsung Galaxy Ring 3 3 3 https://www.samsung.com/us/rings/galaxy-ring/ use photoplethysmography (PPG) sensors for continuous, non-invasive cardiovascular monitoring, such as heart rate (HR) measurement and sleep stage tracking. The anatomy of the finger uniquely enables effective sensing through its thin skin, although finger-based sensing faces distinct challenges, such as motion artifacts due to frequent hand movements. Recent studies have shown smart rings can measure other health parameters, including blood oxygen saturation (SpO2) ( Magno et al. 2019 ) , blood pressure (BP) ( Osman et al. 2022 ) , and bioimpedance ( Usman et al. 2018 ; Usman et al. 2019 ) , highlighting their potential for personal health management and early detection of cardiovascular diseases ( Wang et al. 2025 ; Fiore et al. 2024 ; Kim et al. 2019 ) . As smart ring devices gain popularity, users increasingly expect reliable monitoring of various physiological parameters across diverse scenarios, including daily activities ( Haveman et al. 2022 ) and varying health conditions ( Breteler et al. 2025 ) . However, comprehensive validation of ring reliability remains limited. Most studies rely on processed metrics like heart rate and sleep stage provided by commercial rings ( Koskimäki et al. 2018 ; Malakhatka et al. 2021 ; Poongodi et al. 2022 ; Rajput et al. 2023 ) , with little access to raw sensor data. This limitation restricts algorithm validation and innovation in ring-based cardiovascular sensing. Furthermore, no publicly available datasets currently provide extensive multi-parameter physiological data from wearable rings, presenting a gap in the field. The absence of standardized benchmarks further complicates direct comparisons between algorithms, restricting the robustness and general applicability of existing models. Addressing these gaps is essential for improving ring-based health monitoring technologies. To address these limitations, we implemented two custom ring devices featuring reflective and transmissive PPG optical paths, designed to collect high-quality raw physiological signals alongside ground-truth labels. Each device integrates infrared and red-channel PPG sensors and a 3-axis accelerometer (ACC), capturing synchronized data at sampling rates up to 100 Hz. Our comprehensive data collection involved 34 subjects, with a total of 28.21 hours of recordings across various daily activities, including sitting, standing, talking, and walking. We also induced physiological variability with low oxygen saturation and elevated blood pressure through low-oxygen stimulus and deep squat exercise. The resulting \tau -ring dataset includes extensive raw PPG and ACC data, along with labels for blood volume pulse waveforms (BVP), respiratory waveforms (Resp), HR, respiratory rate (RR), SpO2, BP, and specific activity annotations. This dataset is the first available benchmark dedicated to ring-based cardiovascular sensing, facilitating future algorithm development and validation. Using the \tau -ring dataset, we evaluate four state-of-the-art methods — ResNet ( He et al. 2016 ) , InceptionTime ( Ismail Fawaz et al. 2020 ) , Transformer ( Vaswani et al. 2017 ) , and Mamba ( Gu and Dao 2023 ) —across four cardiovascular monitoring tasks: HR, RR, SpO2, and BP. Our benchmarking results achieved mean absolute errors (MAE) of 5.18 BPM for HR, 2.98 breaths/min for RR, 3.22% for SpO2, 13.48 mmHg for systolic blood pressure (SBP), and 7.47 mmHg for diastolic blood pressure (DBP). These results demonstrate performance comparable to commercial devices such as the Oura and Samsung rings. To facilitate reproducibility and future research, we provide an open-source toolkit, RingTool, which integrates both traditional physics-based signal processing methods (peak detection and spectral analysis) and supervised learning approaches. RingTool enables researchers to flexibly configure methods, select sensor channels, and apply various data filtering techniques, while also supporting the integration of additional datasets and custom algorithms. The core contributions of this work are: (1) \tau -ring Dataset 4 4 4 Available at Kaggle: https://github.com/thuhci/RingTool : We designed and implemented two rings with reflective and transmissive optical paths, establishing the first dataset dedicated to ring-based cardiovascular sensing. The dataset includes 28.21 hours of raw PPG and ACC data from 34 subjects engaged in various activities and physiological states, with synchronized ground-truth labels. (2) Algorithm Performance Benchmarking: We evaluated both physics-based and supervised algorithms for predicting HR, RR, SpO2, and BP in multiple scenarios. The results are comparable with commercial rings, bringing researchers a starting point for better algorithms. (3) RingTool Evaluation Toolkit 5 5 5 Available at GitHub: https://github.com/thuhci/RingTool : We open-sourced RingTool, a toolkit that integrates traditional signal processing with deep learning approaches, enhancing the reproducibility and accessibility of ring-based cardiovascular sensing research. 2. RELATED WORK Table 1 . Summary of Open Wearable Cardiovascular Datasets Dataset Year Subjects Wearables Labels Tasks TROIKA ( Zhang et al. 2015 ) 2015 12 Wrist HR Walk, Run WPPG-Exercise ( Jarchi and Casson 2017 ) 2017 23 Wrist HR Walk, Run, Bike Fitbit2 ( Benedetto et al. [n. d.] ) 2018 15 Wrist HR Bike BAMI ( Lee et al. 2019 ) 2018 24 Wrist HR Stand,Walk, Run WESAD ( Schmidt et al. 2018 ) 2018 17 Wrist HR, EDA, RR, TEMP Sit, Stand FatigueSet ( Kalanadhabhatta et al. 2022 ) 2018 12 Ear, Wrist HR,RR,TEMP Bike DaLiA ( Reiss et al. 2019 ) 2019 15 Wrist HR, EDA, RR, TEMP Sit,Stand,Walk,Exercise EarSet ( Montanari et al. 2023 ) 2023 30 Ear HR, RR Sit, Walk, Run WildPPG ( Meier et al. [n. d.] ) 2024 16 Wrist, Head, Ankle HR Sit, Stand, Walk, Exercise WF-PPG

Read the complete text