Self-supervised learning for human activity recognition using 700,000 person-days of wearable data
Year: 2024
Type: article
Abstract: Accurate physical activity monitoring is essential to understand the impact of physical activity on one's physical health and overall well-being. However, advances in human activity recognition algorithms have been constrained by the limited availability of large labelled datasets. This study aims to leverage recent advances in self-supervised learning to exploit the large-scale UK Biobank accelerometer dataset-a 700,000 person-days unlabelled dataset-in order to build models with vastly improved generalisability and accuracy. Our resulting models consistently outperform strong baselines across eight benchmark datasets, with an F1 relative improvement of 2.5-130.9% (median 24.4%). More importantly, in contrast to previous reports, our results generalise across external datasets, cohorts, living environments, and sensor devices. Our open-sourced pre-trained models will be valuable in domains with limited labelled data or where good sampling coverage (across devices, populations, and activities) is hard to achieve.
Source: npj Digital Medicine
Authors: Hang Yuan, Shing Chan, Andrew P. Creagh, Catherine Tong, Aidan Acquah, David A. Clifton, Aiden Doherty
Institution: University of Oxford
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $3,790
Funders: GlaxoSmithKline, Wellcome Trust, UK Research and Innovation, National Institute for Health and Care Research, British Heart Foundation, Department of Health and Social Care, Royal Academy of Engineering, University of Exeter, University of Oxford, Novo Nordisk, Samsung, Engineering and Physical Sciences Research Council, H2020 European Research Council, Health Data Research UK
Awards: Using wearable sensors to transform the understanding of cardiovascular disease: Prediction, mechanisms, and patient relevant endpoints., MOA: High Efficiency Deep Learning for Embedded and Mobile Platforms (Full EPSRC Fellowship Submission), Research Excellence (round 3), IRC Next Steps Plus : OPERA - Opportunistic Passive Radar for Non-Cooperative Contextual Sensing, Re-thinking Efficiency in Deep Learning under Accelerators and commodity and processors
OpenAlex ID: W4394757986
