{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,26]],"date-time":"2026-10-26T22:17:55Z","timestamp":1793053075485,"version":"4.5.0"},"reference-count":36,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2025,11,14]],"date-time":"2025-11-14T00:00:00Z","timestamp":1763078400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Med."],"abstract":"<jats:p>As societies age, the number of individuals experiencing stroke increases, necessitating more effective rehabilitation strategies. Over half of stroke survivors suffer from upper limb impairments, making assessments of sensory-motor function crucial for both improving interventions and tracking progress. Ideally, such assessments could also be performed at home without requiring a therapist's presence. Advances in computer vision and human pose estimation allow for human movement analysis using consumer-grade cameras. This study investigates whether a single webcam, combined with human pose estimation and deep learning algorithms, can automatically detect compensatory movements in persons with stroke performing a drinking task. Twenty participants with stroke with mild to moderate upper limb impairment were recruited. Each participant performed multiple repetitions of the drinking task while being recorded by multiple cameras and an optical motion capture system (OMC) for kinematic ground truth. The videos were labeled by therapists to indicate the presence or absence of compensatory movements. Human poses were extracted from the videos using MediaPipe, and deep learning models were trained to predict these compensatory movements based on MediaPipe keypoints. Several factors affecting compensation detection accuracy were evaluated. Models trained on raw MediaPipe keypoints for inter-person compensation detection failed to generalize, achieving accuracy around 50%. Using custom features instead of raw keypoints improved the accuracy to 70%. In contrast, intraperson classification achieved high accuracy, typically exceeding 90%. Using OMC data significantly improved classification accuracy compared to using MediaPipe keypoints. Camera angle had an effect on accuracy, and convolutional neural networks outperformed long short-term memory networks. Generalizing models remain limited by (1) the measurement uncertainty of human pose estimation and (2) insufficient data representing the full spectrum of compensatory strategies (3) accurate compensation labels. The results demonstrate that deep learning approaches can differentiate between compensatory and non-compensatory movements when movement representations are sufficiently accurate. Future work should improve pose estimation and expand labeled datasets to better reflect the stroke population. While general models are limited in accuracy, personalized models using consumer cameras can support home-based rehabilitation. This digitalized assessment approach has the potential to quantify recovery progress throughout the continuum of care.<\/jats:p>","DOI":"10.3389\/fmed.2025.1645369","type":"journal-article","created":{"date-parts":[[2025,11,14]],"date-time":"2025-11-14T06:22:25Z","timestamp":1763101345000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":7,"title":["Using deep learning to detect upper limb compensation in individuals post-stroke using consumer-grade webcams\u2014A feasibility study"],"prefix":"10.3389","volume":"12","author":[{"given":"Tim","family":"Unger","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Benjamin","family":"K\u00fchnis","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Lena","family":"Sauerzopf","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Martina R.","family":"Spiess","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Alexander","family":"de Spindler","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Andreas R.","family":"Luft","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Chris","family":"Easthope Awai","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Josef G.","family":"Sch\u00f6nhammer","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Elena","family":"Gavagnin","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2025,11,14]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"E67","DOI":"10.1161\/CIR.0000000000000573","article-title":"Heart disease and stroke statistics - 2018 update: a report from the American Heart Association","volume":"137","author":"Benjamin","year":"2018","journal-title":"Circulation"},{"key":"B2","doi-asserted-by":"publisher","first-page":"2418","DOI":"10.1161\/STROKEAHA.120.029606","article-title":"Burden of stroke in europe: thirty-year projections of incidence, prevalence, deaths, and disability-adjusted life years","volume":"51","author":"Wafa","year":"2020","journal-title":"Stroke"},{"key":"B3","doi-asserted-by":"publisher","DOI":"10.1177\/17474930211065917","author":"Feigin","year":"2022","journal-title":"World Stroke Organization (WSO): Global Stroke Fact Sheet 2022"},{"key":"B4","doi-asserted-by":"publisher","first-page":"817","DOI":"10.1177\/1747493020959186","article-title":"Tracking the global burden of stoke and dementia: World Stroke Day 2020","volume":"15","author":"Markus","year":"2020","journal-title":"Int J Stroke"},{"key":"B5","doi-asserted-by":"publisher","first-page":"599","DOI":"10.1016\/j.pmr.2015.06.008","article-title":"Upper limb motor impairment post stroke","volume":"26","author":"Raghavan","year":"2015","journal-title":"Phys Med Rehabil Clin N Am"},{"key":"B6","doi-asserted-by":"publisher","first-page":"612415","DOI":"10.3389\/frobt.2021.612415","article-title":"Neurorehabilitation from a distance: can intelligent technology support decentralized access to quality therapy?","volume":"8","author":"Lambercy","year":"2021","journal-title":"Front Robot AI"},{"key":"B7","doi-asserted-by":"publisher","first-page":"2819","DOI":"10.1161\/STROKEAHA.112.654228","article-title":"Modulation of neural plasticity as a basis for stroke rehabilitation","volume":"43","author":"Pekna","year":"2012","journal-title":"Stroke"},{"key":"B8","doi-asserted-by":"publisher","first-page":"31","DOI":"10.5535\/arm.23086","article-title":"The effectiveness of goal-oriented dual task proprioceptive training in subacute stroke: a retrospective observational study","volume":"48","author":"Chiaramonte","year":"2024","journal-title":"Ann Rehabil Med"},{"key":"B9","doi-asserted-by":"publisher","first-page":"783","DOI":"10.1177\/1747493019873519","article-title":"Standardized measurement of quality of upper limb movement after stroke: Consensus-based core recommendations from the Second Stroke Recovery and Rehabilitation Roundtable","volume":"14","author":"Kwakkel","year":"2019","journal-title":"Int J Stroke"},{"key":"B10","doi-asserted-by":"publisher","DOI":"10.1177\/1545968317714576","author":"Demers","year":"2017","journal-title":"Do Activity Level Outcome Measures Commonly Used in Neurological Practice Assess Upper-Limb Movement Quality"},{"key":"B11","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1177\/23969873231191304","article-title":"Motor rehabilitation after stroke: European Stroke Organisation (ESO) consensus-based definition and guiding framework","volume":"8","author":"Kwakkel","year":"2023","journal-title":"Eur Stroke J"},{"key":"B12","doi-asserted-by":"publisher","first-page":"313","DOI":"10.1177\/1545968308328727","article-title":"What do motor \"recovery\" and \"compensationg\" mean in patients following stroke?","volume":"23","author":"Levin","year":"2009","journal-title":"Neurorehabil Neural Repair"},{"key":"B13","doi-asserted-by":"publisher","first-page":"844","DOI":"10.1177\/1545968313491008","article-title":"Responsiveness of upper extremity kinematic measures and clinical improvement during the first three months after stroke","volume":"27","author":"Murphy","year":"2013","journal-title":"Neurorehabil Neural Repair"},{"key":"B14","doi-asserted-by":"publisher","first-page":"952757","DOI":"10.3389\/fphys.2022.952757","article-title":"Classification of functional and non-functional arm use by inertial measurement units in individuals with upper limb impairment after stroke","volume":"13","author":"Pohl","year":"2022","journal-title":"Front Physiol"},{"key":"B15","doi-asserted-by":"publisher","first-page":"101","DOI":"10.1186\/s12984-021-00895-3","article-title":"How many trials are needed in kinematic analysis of reach-to-grasp? A study of the drinking task in persons with stroke and non-disabled controls","volume":"18","author":"Frykberg","year":"2021","journal-title":"J NeuroEng Rehabilit"},{"key":"B16","doi-asserted-by":"publisher","first-page":"1903","DOI":"10.3390\/s20071903","article-title":"Deep learning-based upper limb functional assessment using a single kinect v2 sensor","volume":"20","author":"Ma","year":"2020","journal-title":"Sensors"},{"key":"B17","doi-asserted-by":"publisher","first-page":"1","DOI":"10.3390\/s20236940","article-title":"Comparison of a deep learning-based pose estimation system to marker-based and kinect systems in exergaming for balance training","volume":"20","author":"Vonstad","year":"2020","journal-title":"Sensors (Switzerland)"},{"key":"B18","doi-asserted-by":"publisher","first-page":"918617","DOI":"10.3389\/fbioe.2022.918617","article-title":"Accuracy and validity of a single inertial measurement unit-based system to determine upper limb kinematics for medically underserved populations","volume":"10","author":"Hughes","year":"2022","journal-title":"Front Bioeng Biotechnol"},{"key":"B19","doi-asserted-by":"publisher","first-page":"554","DOI":"10.3390\/s24020554","article-title":"Wearable sensor to monitor quality of upper limb task practice for stroke survivors at home","volume":"24","author":"Seo","year":"2024","journal-title":"Sensors"},{"key":"B20","doi-asserted-by":"publisher","first-page":"877563","DOI":"10.3389\/fphys.2022.877563","article-title":"Using wearable inertial sensors to estimate clinical scores of upper limb movement quality in stroke","volume":"13","author":"Werner","year":"2022","journal-title":"Front Physiol"},{"key":"B21","doi-asserted-by":"publisher","first-page":"22880","DOI":"10.1038\/s41598-023-49360-2","article-title":"Comparison of markerless and marker-based motion capture systems using 95% functional limits of agreement in a linear mixed-effects modelling framework","volume":"13","author":"Das","year":"2023","journal-title":"Sci Rep"},{"key":"B22","first-page":"57228","article-title":"Kinematic analysis using 3D motion capture of drinking task in people with and without upper-extremity impairments","volume":"133","author":"Murphy","year":"2018","journal-title":"J Visual Exper"},{"key":"B23","doi-asserted-by":"publisher","first-page":"178","DOI":"10.1186\/s12984-020-00773-4","article-title":"Relationships between accelerometry and general compensatory movements of the upper limb after stroke","volume":"17","author":"Barth","year":"2020","journal-title":"J NeuroEng Rehabilit"},{"key":"B24","doi-asserted-by":"publisher","first-page":"8","DOI":"10.1093\/ptj\/84.1.8","article-title":"Development and validation of a scale for rating motor compensations used for reaching in patients with hemiparesis: the reaching performance scale","volume":"84","author":"Levin","year":"2004","journal-title":"Phys Ther"},{"key":"B25","doi-asserted-by":"publisher","first-page":"e59781","DOI":"10.2196\/59781","article-title":"Technology use for home-based stroke rehabilitation in Switzerland from the perspectives of persons living with stroke, informal caregivers, and therapists: qualitative interview and focus group study","volume":"11","author":"Sauerzopf","year":"2024","journal-title":"JMIR Rehabilit Assist Technol"},{"key":"B26","article-title":"BlazePose: on-device real-time body pose tracking","author":"Bazarevsky","year":"2020","journal-title":"arxiv.org\/abs\/2006.10204"},{"key":"B27","doi-asserted-by":"publisher","first-page":"e34307","DOI":"10.2196\/34307","article-title":"Technology-based compensation assessment and detection of upper extremity activities of stroke survivors: systematic review","volume":"24","author":"Wang","year":"2022","journal-title":"J Med Internet Res"},{"key":"B28","doi-asserted-by":"publisher","first-page":"1359776","DOI":"10.3389\/fdgth.2024.1359776","article-title":"Upper limb movement quality measures: comparing IMUs and optical motion capture in stroke patients performing a drinking task","volume":"6","author":"Unger","year":"2024","journal-title":"Front Digital Health"},{"key":"B29","article-title":"Differentiable biomechanics for markerless motion capture in upper limb stroke rehabilitation: a comparison with optical motion capture","author":"Unger","year":"2024","journal-title":"arXiv:2411.14992"},{"key":"B30","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1109\/ICORR58425.2023.10304697","article-title":"Exploring the feasibility of computer vision for detecting post-stroke compensatory movements","volume-title":"2023 International Conference on Rehabilitation Robotics (ICORR)","author":"Lin","year":"2023"},{"key":"B31","doi-asserted-by":"publisher","first-page":"40362","DOI":"10.2340\/jrm.v56.40362","article-title":"reliable and valid assessment of upper limb movement quality after stroke: the observational drinking task assessment","volume":"56","author":"Jose","year":"2024","journal-title":"J Rehabilit Med"},{"key":"B32","doi-asserted-by":"publisher","first-page":"217","DOI":"10.1186\/s12984-024-01506-7","article-title":"Evaluating inter- and intra-rater reliability in assessing upper limb compensatory movements post-stroke: creating a ground truth through video analysis?","volume":"21","author":"Sauerzopf","year":"2024","journal-title":"J Neuroeng Rehabil"},{"key":"B33","doi-asserted-by":"publisher","first-page":"6427","DOI":"10.1109\/CVPR52688.2022.00632","article-title":"MotionAug: augmentation with physical correction for human motion prediction","author":"Maeda","year":"2022","journal-title":"Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition."},{"key":"B34","doi-asserted-by":"publisher","first-page":"202","DOI":"10.1177\/09544119251315877","article-title":"Synthetic data generation in motion analysis: a generative deep learning framework","volume":"239","author":"Perrone","year":"2024","journal-title":"Proc Inst Mechan Eng Part H J Eng Med"},{"key":"B35","doi-asserted-by":"publisher","first-page":"6861","DOI":"10.3390\/s24216861","article-title":"Enhancing activity recognition after stroke: generative adversarial networks for kinematic data augmentation","volume":"24","author":"Hadley","year":"2024","journal-title":"Sensors"},{"key":"B36","doi-asserted-by":"publisher","first-page":"3613","DOI":"10.3390\/s24113613","article-title":"Optimizing rare disease gait classification through data balancing and generative ai: insights from hereditary cerebellar ataxia","volume":"24","author":"Trabassi","year":"2024","journal-title":"Sensors"}],"container-title":["Frontiers in Medicine"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2025.1645369\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,11,14]],"date-time":"2025-11-14T06:22:27Z","timestamp":1763101347000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2025.1645369\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,11,14]]},"references-count":36,"alternative-id":["10.3389\/fmed.2025.1645369"],"URL":"https:\/\/doi.org\/10.3389\/fmed.2025.1645369","relation":{},"ISSN":["2296-858X"],"issn-type":[{"value":"2296-858X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,11,14]]},"article-number":"1645369"}}