Our core topics and methods
Player and ball tracking, pose estimation, event detection
We design robust vision models for real-world broadcast and training data to extract actionable spatiotemporal insights.
tracking pose estimation event detection broadcast analysis
Predictive and causal modeling for athlete and team performance
From forecasting and counterfactuals to interpretable models that support coaching and tactical decisions.
forecasting causal inference explainability optimization
Trustworthy systems to assist coaches, referees, and athletes
We study usability, transparency, and fairness of AI in sports, focusing on human-AI collaboration.
explainable AI fairness HCI sports officiating
