Real-time, accurate, and open source upper-limb musculoskeletal analysis using a single RGBD camera - An exploratory...

Area of Science:

• Biomechanics
• Rehabilitation Engineering
• Computer Vision

Background:

• Biomechanical biofeedback enhances rehabilitation but relies on costly motion capture systems.
• Existing computer vision methods have limitations in upper-limb joint angle accuracy and exclude scapular/clavicular motion.

Purpose of the Study:

• To introduce an open-source, low-cost solution for high-fidelity upper-limb kinematics using a depth-sensing camera.
• To enable real-time biomechanical analysis, including muscle force estimation, for upper-limb movements.

Main Methods:

• Utilized a single consumer-grade depth-sensing camera (∼$500) with semi-automatic skin marker labeling.
• Collected data from eight participants performing hand-cycling, comparing the depth-sensing camera system with an optoelectronic system.
• Recorded muscle activity (electromyography) and external load (instrumented pedals).

Main Results:

• Demonstrated significant agreement between depth-sensing and optoelectronic systems for 3D marker positions (3.3 ± 3.9 mm error).
• Showcased low error propagation in biomechanical analysis, with joint angle differences below 5°.
• Achieved real-time biofeedback at 68 Hz.

Conclusions:

• The novel depth-sensing camera approach provides a low-cost, accurate alternative for upper-limb motion capture.
• The system shows potential for comprehensive biomechanical studies and clinical applications in rehabilitation.
• Further research is needed to validate its use in pathological populations.