BRIDGE: Borderless Reconfiguration for Inclusive and Diverse Gameplay Experience via Embodiment Transformation

Paper Title

BRIDGE: Borderless Reconfiguration for Inclusive and Diverse Gameplay Experience via Embodiment Transformation

Publication Info

• Topic area: Inclusive sports learning and embodiment transformation in parasports.
• Keywords: parasports, wheelchair basketball, embodiment transformation, tactical learning, self-efficacy, inclusive design, functional classification, sports visualization, HCI, rehabilitation.

Background and Problem

• Problem / challenge: Parasports face a scarcity of training resources, particularly video materials tailored to their specific needs. Non-disabled sports footage is often used but is difficult to interpret due to differences in embodiment, such as wheelchair-based movement constraints and functional classifications.
• Significance: Addressing this gap can improve tactical learning, self-efficacy, and equitable access to training resources for parasports athletes, fostering greater inclusion and motivation.
• Motivation and related work: Prior research in parasports and HCI has explored VR/AR systems and motion analysis but has not adequately bridged the gap between non-disabled and parasports contexts. This paper builds on embodiment theory and sports visualization to propose a system that adapts non-disabled sports footage for wheelchair basketball.

Solution

• Proposed approach: BRIDGE, a system that converts stand-up basketball footage into wheelchair basketball simulations using a multi-stage reconstruction pipeline and embodiment-aware orientation mapping.
• Novelty:
  1. A design framework for embodiment-aware orientation mapping that decomposes head, trunk, and wheelchair base orientations.
  2. A novel conversion technique for reconstructing stand-up basketball footage into wheelchair basketball contexts.
  3. Empirical validation with 20 participants, showing improved tactical understanding, functional classification, and self-efficacy.
  4. Design implications for inclusive sports learning, emphasizing equitable access and motivation.
• Procedure and key techniques:
  1. Game Streaming Reconstruction: Detects players, ball, and court geometry; maps trajectories into 3D space.
  2. Embodiment-aware Orientation Mapping: Separates head, trunk, and wheelchair base orientations, constrained by classification-based rules.
  3. Evaluation: Conducted two user studies with wheelchair basketball players to assess naturalness, tactical understanding, and self-efficacy.

Results

• Concrete findings:
  • Embodiment-aware mapping significantly improved perceived naturalness (mapped condition rated ~5/7 vs. baseline ~3/7).
  • Functional classification accuracy was ~80% for simple plays and ~74% for complex plays, with trunk mobility identified as the most influential cue.
  • Tactical action was preserved in reconstructed footage, with no significant differences in understanding between original and reconstructed videos.
  • Self-efficacy ratings were higher for reconstructed videos, especially for complex tactics (e.g., national players: 6.38/7 for reconstructed vs. 4.12/7 for original).
  • Subjective ratings (e.g., imagery vividness, ease of imagery) were consistently higher for reconstructed videos across all measures.
• Advantage over baselines:
  • Enhanced naturalness and functional classification through embodiment-aware mapping.
  • Higher self-efficacy and improved accessibility compared to untransformed stand-up basketball footage.
  • Better alignment with wheelchair basketball constraints, fostering actionable tactical understanding.
• Experiments / evaluation:
  • Two studies with 20 participants (10 national team players, 10 non-elite players).
  • Metrics included perceived naturalness, functional classification accuracy, tactical understanding, self-efficacy, and subjective evaluations.
  • Videos were reconstructed using Unity and evaluated under mapped and baseline conditions.
• Limitations and future work:
  • Small sample size and inclusion of non-disabled participants may limit generalizability.
  • Focused on wheelchair basketball; generalization to other sports remains untested.
  • Lacks multimodal feedback (e.g., haptics) and interactive training features.
  • Future work should explore larger-scale studies, multimodal systems, and applications in rehabilitation and education.

Summary

BRIDGE is a system that transforms stand-up basketball footage into wheelchair basketball simulations, addressing the lack of parasport-specific training resources. By employing embodiment-aware orientation mapping and a multi-stage reconstruction pipeline, BRIDGE enhances tactical understanding, functional classification, and self-efficacy among wheelchair basketball players. Empirical results show improved naturalness, preserved tactical actions, and higher subjective ratings for reconstructed videos. While currently focused on wheelchair basketball, the approach has potential applications in other sports, education, and rehabilitation, offering a foundation for inclusive, cross-embodiment learning technologies.