Author(s)

Xiao Xu¹, HaoNan Qin², LanShu Zhou^2*

Affiliation(s)

¹School of Health Science and Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China.

²School of Nursing, Naval Medical University, Shanghai 200093, China.

Corresponding Author

LanShu Zhou

ABSTRACT

Stroke rehabilitation is going through some pretty major changes right now. We're seeing a gradual move away from those subjective clinical scales toward more objective, automated assessments—though it's not happening as fast as some people hoped. Traditional methods like the Fugl-Meyer Assessment (FMA) are still considered the gold standard in clinics, but they've got some real problems. They're time-consuming, different raters often give different scores, and they tend to hit ceiling effects that miss subtle improvements patients are making.This review looks at important studies from 2016 to 2025. We trace how automated detection systems evolved from those lab-based depth sensors to RGB cameras you can use at home and wearable IMUs. We compare different sensing technologies and look at public datasets. There's also this "sim-to-real" gap issue - basically, models trained on healthy actors often don't work well with real patients, which is a big problem. The algorithms have changed a lot too, going from traditional machine learning to foundation models. Clinical metrics have also expanded beyond simple detection to things like severity grading.At the end, we talk about what this all means for tele-rehabilitation and why getting clinicians to actually adopt these technologies is still really challenging. We suggest a roadmap that focuses on privacy-preserving collaboration and long-term validation studies, though honestly implementing all of this won't be easy.

KEYWORDS

Stroke rehabilitation; Compensatory movement; Deep learning; Multimodal fusion

CITE THIS PAPER

Xiao Xu, HaoNan Qin, LanShu Zhou. Deep learning in stroke rehabilitation compensatory movement detection: from laboratory benchmarks to clinical reality. Eurasia Journal of Science and Technology. 2026, 8(1): 1-6. DOI: https://doi.org/10.61784/ejst3127.

REFERENCES

[1] Langhorne P, Bernhardt J, Kwakkel G. Stroke rehabilitation. The Lancet, 2011, 377(9778): 1693-1702. DOI: 10.1016/S0140-6736(11)60325-5.

[2] Cirstea M C, Levin M F. Compensatory strategies for reaching in stroke. Brain, 2000, 123(5): 940-953. DOI: 10.1093/brain/123.5.940.

[3] Gladstone D J, Danells C J, Black S E. The Fugl-Meyer assessment of motor recovery after stroke: a critical review of its measurement properties. Neurorehabilitation and Neural Repair, 2002, 16(3): 232-240. DOI: 10.1177/154596802401105171.

[4] Capecci M, Ceravolo M G, Ferracuti F, et al. The Kimore dataset: KInematic assessment of MOvement and clinical scores for REmote monitoring of physical rehabilitation. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2019, 27(7): 1436-1448. DOI: 10.1109/TNSRE.2019.2923060.

[5] Vakanski A, Jun H, Paul D, et al. A data set of human body movements for physical rehabilitation exercises. Data, 2018, 3(1): 2. DOI: 10.3390/data3010002.

[6] Zhi Y X, Lukasik M, Li M H, et al. Automatic Detection of Compensation During Robotic Stroke Rehabilitation Therapy. Journal of Translational Engineering in Health and Medicine, 2018, 6: 1-7. DOI: 10.1109/JTEHM.2017.2780836.

[7] Sohn W J, Sipahi R, Sanger T D, et al. Portable Motion-Analysis Device for Upper-Limb Research, Assessment, and Rehabilitation in Non-Laboratory Settings. Journal of Translational Engineering in Health and Medicine, 2019, 7: 1-14. DOI: 10.1109/JTEHM.2019.2953257.

[8] Lin H P, Chen Y, Wu C. Exploring the feasibility of computer vision for detecting post-stroke compensatory movements. International Conference on Rehabilitation Robotics (ICORR), 2023, 1-6. DOI: 10.1109/ICORR58425.2023.10304697.

[9] Yamamoto M, Tanaka S, Sato K. Single camera-based gait analysis using pose estimation for ankle-foot orthosis stiffness adjustment on individuals with stroke. IEEE Journal of Translational Engineering in Health and Medicine, 2025, 13: 333-340. DOI: 10.1109/JTEHM.2025.3585442.

[10] Pilla-Barroso M, Garcia J, Lopez A. Assessing the impact of depth data on human kinematics: A hybrid RGB-D and MediaPipe method. IEEE Medical Measurements & Applications (MeMeA), 2025, 1-6. DOI: 10.1109/MeMeA65319.2025.11068070.

[11] Zheng S, Liang G, Chen J, et al. Severity Assessment of Cervical Spondylotic Myelopathy Based on Intelligent Video Analysis. Journal of Biomedical and Health Informatics, 2022, 26(9): 4486-4496. DOI: 10.1109/JBHI.2022.3184870.

[12] Dousty M, Zariffa J. Tenodesis Grasp Detection in Egocentric Video. Journal of Biomedical and Health Informatics, 2021, 25(5): 1463-1470. DOI: 10.1109/JBHI.2020.3003643.

[13] Okita S, Shiro Y, Takeda S. Movement diversity and complexity increase as arm impairment decreases after stroke: Quality of movement experience as a possible target for wearable feedback. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2024, 32: 2961-2970. DOI: 10.1109/TNSRE.2024.3439669.

[14] Lu T Y, Liu K C, Hsieh C Y, et al. Instrumented shoulder functional assessment using inertial measurement units for frozen shoulder. Biomedical and Health Informatics (BHI), 2021, 1-4. DOI: 10.1109/BHI50953.2021.9508588.

[15] Chua M X, Peng S, Do T N, et al. A Wearable Strain-Sensor-Based Shoulder Patch for Fatigue Detection in Bicep Curls. Transactions on Instrumentation and Measurement, 2025, 74: 1-12. DOI: 10.1109/TIM.2025.3600715.

[16] Gao S, Li X, Zhang Y. Temporal dynamics and physical priori multimodal network for rehabilitation physical training evaluation. Journal of Biomedical and Health Informatics, 2024, 28(9): 5613-5623. DOI: 10.1109/JBHI.2024.3414291.

[17] Ding K, Wang Z, Liu H, et al. System for detection and quantitative evaluation of compensatory movement in post-stroke patients based on wearable sensor and machine learning algorithm. IEEE Sensors Journal, 2024, 24(14): 22830-22842. DOI: 10.1109/JSEN.2024.3408323.

[18] Wu Q, Wang Z, Chen Y. sEMG-Based Adaptive Cooperative Multi-Mode Control of a Soft Elbow Exoskeleton Using Neural Network Compensation. Transactions on Neural Systems and Rehabilitation Engineering, 2023, 31, 3384-3396. DOI: 10.1109/TNSRE.2023.3306201.

[19] Zhu K, Wu Y, Zhang H. Real-Time IMU-Based Kinematics in the Presence of Wireless Data Drop. Journal of Biomedical and Health Informatics, 2024, 28(11): 6512-6524. DOI: 10.1109/JBHI.2024.3432668.

[20] Mesquita G, Silva D, Santos A. Frame-Level Real-Time Assessment of Stroke Rehabilitation Exercises from Video-Level Labeled Data: Task-Specific vs. Foundation Models. Engineering in Medicine and Biology Society (EMBC), 2025: 1-7. DOI: 10.1109/EMBC58623.2025.11251613.

[21] Cóias A R, Lee M H, Bernardino A, et al. Learning Frame-Level Classifiers for Video-Based Real-Time Assessment of Stroke Rehabilitation Exercises From Weakly Annotated Datasets. Transactions on Neural Systems and Rehabilitation Engineering, 2025, 33: 3334-3345. DOI: 10.1109/TNSRE.2025.3602548.

[22] Hsu T C, Chang D M. Retrainable hand gesture recognition model driven by edge AI for occluded hand tracking. International Conference on Dependable Systems and Their Applications (DSA), 2024: 43-48. DOI: 10.1109/DSA63982.2024.00017.