Mapping assessment practices in DanceSport: a Human-In-The-Loop systematic review

Area of Science:

• Sports Science
• Biomechanical Engineering
• Human-Computer Interaction

Background:

• DanceSport combines athleticism, artistry, and musicality, traditionally relying on subjective judging.
• Recent trends explore technology-assisted methods for objective and measurable performance assessment in DanceSport.
• A synthesis of research themes, technological applications, and validation methods in DanceSport assessment is lacking.

Purpose of the Study:

• To systematically review and synthesize major themes in technology-assisted DanceSport assessment.
• To identify the evolution of research themes and applied technologies in evaluating DanceSport performance.
• To analyze the validation strategies employed in existing DanceSport assessment systems.

Main Methods:

• Systematic literature review following PRISMA guidelines, analyzing 24 studies.
• Human-In-The-Loop topic modeling (Exploratory LDA) on abstracts to identify themes and transitions.
• Qualitative content analysis of full texts for assessment approaches, technologies, and validation designs.

Main Results:

• Four key themes emerged: 3D motion perception, posture/motion evaluation, dynamic performance image analysis, and assessment validation.
• Research shows a trend towards objective, data-driven assessment using indicators like posture, motion, and rhythm.
• Technologies applied include motion capture, wearables, computer vision, and computational models, though validation is limited by scope and context.

Conclusions:

• DanceSport assessment is advancing towards technology-assisted, multimodal systems.
• Full integration of technical measurement with aesthetic judgment and competition-level validation is pending.
• Future research should focus on comprehensive datasets, cross-dance comparisons, and involving stakeholders in validation frameworks.