This study evaluates a novel camera-based system designed to support remote rehabilitation by measuring hand and upper-limb biomechanics in real time. Many patients recovering from musculoskeletal or neurological conditions require frequent monitoring during rehabilitation, but regular clinic visits may be difficult due to distance, cost, or limited access to specialized care. Current telehealth approaches typically rely on qualitative assessments or self-reported feedback rather than objective biomechanical measurements.

The purpose of this study is to determine whether a computer vision-based system can accurately estimate biomechanical parameters such as joint angles, range of motion, muscle force, and joint torque using only a standard camera. The system analyzes hand movement using artificial intelligence and biomechanical modeling to provide real-time measurements during rehabilitation exercises.

Participants will perform guided hand-movement tasks while the system records video and extracts anatomical landmarks. These data will be used to compute biomechanical parameters and assess whether the system can reliably monitor rehabilitation progress remotely. The results will help determine whether this technology can provide clinicians with objective, continuous data to support personalized rehabilitation and improve patient outcomes.

This study aims to develop and validate a camera-based tele-rehabilitation platform capable of estimating biomechanical parameters of the human hand and upper limb in real time. Musculoskeletal and neurological conditions often require continuous monitoring during rehabilitation, yet many patients-particularly those in rural or underserved regions-have limited access to frequent in-person therapy sessions. Existing telehealth systems primarily rely on subjective reporting or periodic video consultations and often lack quantitative biomechanical measurements necessary for precise monitoring of recovery.

The objective of this research is to evaluate whether computer vision and biomechanical modeling can provide accurate, quantitative measurements of joint motion and force using a single camera. The central hypothesis is that artificial intelligence algorithms can detect anatomical landmarks of the hand from video data and combine them with mechanical modeling techniques to estimate joint angles, torques, and muscle forces in real time. Continuous biomechanical tracking may allow clinicians to better monitor rehabilitation progress and make timely adjustments to therapy protocols.

Participants will perform standardized hand-movement exercises while video data are captured using a consumer-grade camera such as a smartphone or laptop camera. Computer vision algorithms will identify hand landmarks and calculate joint kinematics. These measurements will then be integrated with inverse dynamics modeling to estimate biomechanical parameters including joint torque, range of motion, and force generation.

The study will evaluate the reliability and validity of the proposed system by comparing the computed biomechanical measurements with established biomechanical models and reference datasets. Key outcomes include the accuracy of landmark detection, reliability of biomechanical parameter estimation, and feasibility of remote monitoring during rehabilitation exercises.

Successful completion of this study will demonstrate the feasibility of a low-cost, accessible tele-rehabilitation platform capable of delivering objective biomechanical feedback to clinicians and patients. This approach has the potential to improve access to rehabilitation services, enhance patient engagement, and support data-driven clinical decision-making in remote healthcare settings.

Inclusion Criteria:

• Adults aged 18 years or older.
• Individuals undergoing or recovering from upper-limb or hand rehabilitation following musculoskeletal or neurological injury or surgery.
• Ability to perform basic hand or upper-limb movement tasks required for the rehabilitation exercises.
• Ability to understand study instructions and provide informed consent.

Exclusion Criteria:

• Severe cognitive impairment preventing understanding of study procedures.
• Medical conditions that prevent safe participation in hand or upper-limb rehabilitation exercises.
• Severe visual impairment preventing interaction with the camera-based monitoring system.
• Participation in another interventional study that could affect rehabilitation outcomes.

De-identified individual participant data (IPD) from this study may be shared upon reasonable request following study completion. Data to be shared include derived biomechanical parameters (e.g., joint torque, range of motion), summary performance metrics, and associated metadata necessary for interpretation.

Raw video data will not be shared due to potential identifiability concerns. All shared data will be de-identified in accordance with institutional policies and IRB requirements.

Access will be provided to qualified researchers upon approval of a data use agreement and for purposes consistent with the original study objectives.