Ice Hockey Analytics System
Computer vision systems for real-time ice hockey analysis including player tracking, action recognition, and game event detection
Computer vision analysis of ice hockey broadcast videos for real-time game analytics.
Overview
This project focuses on developing advanced computer vision systems for ice hockey broadcast video analysis. The system provides real-time insights into player movements, game events, and performance metrics.
Key Features
- Homography Estimation: Rink-agnostic analysis for different hockey arenas
- Puck Detection: Real-time tracking of puck movement and position
- Player Action Recognition: Identification of player actions and behaviors
- Game Event Detection: Automatic detection of goals, penalties, and other game events
Technical Approach
The system utilizes state-of-the-art deep learning techniques including:
- Convolutional Neural Networks (CNNs) for feature extraction
- Temporal modeling for action recognition
- Multi-object tracking for player and puck detection
- Geometric transformations for homography estimation
Collaborations
- Stathletes Inc: Industry partner for broadcast video analysis
- University of Waterloo: Academic research collaboration
- Sport Analytics Research Group: Multi-disciplinary research team
Publications
- Buzko K., Nazemi A., Clausi DA., Chen Y. “Ice Hockey Action Recognition via Contextual Priors.” Linköping Hockey Analytics Conference, 2025. (Best Paper Award)
- Iaboni E., Negulescu S., Pitassi M., Nazemi A., Bright J., Chomko V., Clausi DA., Dickinson S., Brecht T. “New Views of Shots-Towards Measures of Net Visibility and Reachability.” Linköping Hockey Analytics Conference, 2025.
Impact
This research contributes to the growing field of sports analytics, providing coaches, analysts, and broadcasters with valuable insights for performance analysis and game strategy development.