CourtMotion: Learning Event-Driven Motion Representations from Skeletal Data for Basketball

2025-12-01

AI Summary

Traditional basketball event analysis relies solely on positional tracking data, failing to capture critical motion semantics such as body orientation, defensive posture, and shooting readiness. To address this limitation, we propose an event-driven skeletal motion representation learning framework. Our method introduces an event projection head that explicitly aligns skeletal action patterns with basketball event semantics; integrates graph neural networks to model individual player skeletal dynamics over time; and employs a specialized multi-player attention mechanism within a Transformer architecture to capture inter-player interactions. Furthermore, we adopt multi-task learning to jointly optimize event classification, trajectory prediction, and pose reconstruction. Evaluated on official NBA optical tracking data, our approach reduces trajectory prediction error by 35% and achieves significant improvements in event recognition and intent inference over state-of-the-art methods. It effectively supports downstream analytical tasks including pass decision-making and defensive response analysis.

Technology Category

• Computer Vision: Motion & Tracking
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding

Abstract

This paper presents CourtMotion, a spatiotemporal modeling framework for analyzing and predicting game events and plays as they develop in professional basketball. Anticipating basketball events requires understanding both physical motion patterns and their semantic significance in the context of the game. Traditional approaches that use only player positions fail to capture crucial indicators such as body orientation, defensive stance, or shooting preparation motions. Our two-stage approach first processes skeletal tracking data through Graph Neural Networks to capture nuanced motion patterns, then employs a Transformer architecture with specialized attention mechanisms to model player interactions. We introduce event projection heads that explicitly connect player movements to basketball events like passes, shots, and steals, training the model to associate physical motion patterns with their tactical purposes. Experiments on NBA tracking data demonstrate significant improvements over position-only baselines: 35% reduction in trajectory prediction error compared to state-of-the-art position-based models and consistent performance gains across key basketball analytics tasks. The resulting pretrained model serves as a powerful foundation for multiple downstream tasks, with pick detection, shot taker identification, assist prediction, shot location classification, and shot type recognition demonstrating substantial improvements over existing methods.

Problem

• Learning event-driven motion representations from skeletal data
• Predicting basketball events by understanding physical and semantic motion patterns
• Improving trajectory prediction and analytics tasks over position-only models

Innovation

• Graph Neural Networks process skeletal tracking data
• Transformer models player interactions with attention mechanisms
• Event projection heads link movements to basketball events