A Vision-Based, Game-Integrated Framework for Elbow Rehabilitation with Perspective Error Compensation: A Proof-of-Concept Study
| Author | Ali Soltani Sartakhti | en |
| Author | Meysam Siyah Mansoory | en |
| Author | Soheila Abbasi | en |
| Orcid | Ali Soltani Sartakhti [0009-0003-0647-1727] | en |
| Orcid | Meysam Siyah Mansoory [0000-0002-8160-7618] | en |
| Orcid | Soheila Abbasi [0000-0001-9312-2485] | en |
| Issued Date | 2026-06-30 | en |
| Abstract | Background: Stroke is a leading cause of disability worldwide, and a substantial proportion of stroke survivors experience upper limb dysfunction, particularly in the elbow joint. Traditional rehabilitation faces two major challenges: limitations in clinicians’ ability to objectively monitor movement and reduced patient motivation due to the repetitive nature of therapeutic exercises. Although game-based systems may increase patient motivation, they often lack a quantitative link to biomechanical indicators. In addition, red-green-blue (RGB) vision-based systems have simple hardware requirements but are sensitive to perspective errors and the lack of depth information. Objectives: This study aimed to present a proof-of-concept framework for elbow rehabilitation that integrates interactive gaming, markerless posture estimation, and a proposed machine learning-based error correction architecture to enhance patient motivation while enabling quantitative range of motion (ROM) monitoring with a conventional RGB camera. Methods: This simulation-based proof-of-concept framework was evaluated in 2 distinct nonclinical development layers. First, a web-based geometric simulator was developed to examine angular error behavior caused by body rotation and arm elevation relative to the camera. On the basis of this analysis, a correction model using normalized vector length ratios was proposed. In the second layer, an interactive game environment was implemented using Python and the MediaPipe, OpenCV, and Pygame libraries. The elbow angle was computed from the 2D coordinates of the shoulder, elbow, and wrist and mapped directly to the game logic. An initial calibration protocol and a dashboard for adjusting treatment parameters were also developed. Results: The simulations showed that angle estimation error in 2D systems increased nonlinearly with body rotation and changes in projection. Applying the proposed correction solution demonstrated, within the conceptual simulation environment, a theoretical capacity to control error growth within a specified operational range. Practical implementation of the game also enabled a direct link between treatment goals and progress within the interactive environment. Conclusions: The proposed framework demonstrates the feasibility of developing a low-cost, noninvasive elbow rehabilitation system that mitigates perspective errors and integrates interactive gaming. This study presents only a preliminary framework, rather than evidence of clinical readiness or real-user accuracy; however, it may support future research. | en |
| DOI | https://doi.org/10.5812/jcrps-170829 | en |
| URI | https://brieflands.com/journals/jcrps/articles/170829 | en |
| Keyword | Stroke Rehabilitation | en |
| Keyword | Upper Extremity | en |
| Keyword | Artificial Intelligence | en |
| Keyword | Video Games | en |
| Publisher | Brieflands | en |
| Title | A Vision-Based, Game-Integrated Framework for Elbow Rehabilitation with Perspective Error Compensation: A Proof-of-Concept Study | en |
| Type | Research Article | en |
Files
Original bundle
1 - 1 of 1
Loading...
- Name:
- jcrps-15-1-170829-publish-pdf.pdf
- Size:
- 989.72 KB
- Format:
- Adobe Portable Document Format
- Description:
- Article/s PDF