Related Experiment Video
Updated: May 21, 2025
06:28
Biomechanical Changes Related to Low Back Pain: An Innovative Tool for Movement Pattern Assessment and Treatment Evaluation in Rehabilitation
Published on: December 13, 2024
393
Machine Learning-Based Computer Vision for Depth Camera-Based Physiotherapy Movement Assessment: A Systematic Review
Yafeng Zhou1,2, Fadilla 'Atyka Nor Rashid1, Marizuana Mat Daud3
1Center for Artificial Intelligence Technology (CAIT), Faculty of Information Science and Technology, Universiti Kebangsaan Malaysia, Bangi 43600, Selangor, Malaysia.
Sensors (Basel, Switzerland)
|March 17, 2025
Summary
Machine learning computer vision with depth cameras shows promise for physiotherapy movement assessment. Further research is needed to address real-world validation and algorithm generalization for clinical use.
Area of Science:
- Rehabilitation Technology
- Medical Imaging
- Artificial Intelligence in Healthcare
Background:
- Physiotherapy movement assessment traditionally relies on manual observation.
- Machine learning (ML) and computer vision (CV) offer objective, quantitative movement analysis.
- Depth cameras provide rich 3D data for enhanced movement tracking.
Purpose of the Study:
- To systematically review recent advancements (2020-2024) in ML-based CV for physiotherapy movement assessment.
- To identify implementation scenarios, data collection/processing methods, and algorithms used.
- To highlight key challenges and future research directions.
Main Methods:
- Systematic literature review following PRISMA guidelines.
- Searches conducted across Web of Science, Scopus, PubMed, and ADS.
- Analysis of 18 selected studies focusing on ML/CV in physiotherapy movement assessment.
Main Results:
- Local (50%), clinical (33.4%), and remote (22.3%) implementation scenarios identified.
- Kinect series depth cameras (65.4%) were prevalent; RGB-D (55.6%) and skeletal data (27.8%) were common processing approaches.
- Algorithms included traditional ML (44.4%) and deep learning (41.7%).
Conclusions:
- ML-based CV systems demonstrate effectiveness for physiotherapy movement assessment.
- Challenges include limited real-world validation, dataset diversity, and algorithm generalization.
- Future research should focus on clinical validation and improving algorithm generalizability for practical application.
