Functional Profiling in Paralympic Water Polo Using Deep Learning, Stereo Vision, and Phase-Based Kinematic Analysis:...

Area of Science:

• Sports Science
• Biomechanics
• Paralympic Athletics

Background:

• Accurate classification systems are crucial for Paralympic water polo to reflect sport-specific functional performance.
• Existing methods may not fully capture the nuances of residual motor function under ecologically valid conditions.

Purpose of the Study:

• To propose and evaluate a task-specific kinematic profiling framework for objective, biomechanically interpretable descriptors of residual motor function in Paralympic water polo.
• To assess the feasibility of using pose estimation for quantitative measurement in sport-specific tasks.

Main Methods:

• A pilot study involving five male national-level water polo athletes (three with eligible impairments, two able-bodied).
• Standardized sport-specific tasks including floating, propulsion, passing, and shooting under physical opposition.
• Stereoscopic video, OpenPose-based 3D reconstruction, and phase-based analysis to extract kinematic features and composite indices.

Main Results:

• The kinematic profiling framework successfully extracted objective descriptors of postural control, propulsion, and upper-limb function.
• Task- and side-specific differences in stabilization, propulsion, and motor reorganization were identified.
• Automatic ball-release detection demonstrated high accuracy, matching manual verification.

Conclusions:

• The developed framework provides transparent and interpretable candidate descriptors for Paralympic water polo classification.
• Pose estimation serves as a viable quantitative tool, treating visibility interruptions as meaningful data.
• This approach can inform future evidence-based classification research in Paralympic water polo.