Researchers at the University of Waterloo have developed two groundbreaking AI systems that significantly enhance the analysis of hockey games using standard broadcast video, eliminating the need for expensive tracking equipment. The innovations, led by graduate student Liam Salass and engineering professors Dr. David Clausi and Dr. John Zelek, leverage advancements in computer vision and systems design to tackle longstanding challenges in tracking fast-paced, complex sports action. One system, called Puck Localization Using Contextual Cues (PLUCC), improves the accuracy of puck detection by analyzing players’ body positions and gaze direction. Since players typically focus on the puck during play, the AI uses these visual cues to infer its location even when it’s obscured by players or blurred by motion. In testing, PLUCC increased puck localization accuracy by 12% and reduced localization error by over 25% compared to existing methods. This breakthrough makes high-precision puck tracking accessible to smaller teams and amateur leagues that can’t afford million-dollar tracking systems like Hawk-Eye. “Our goal was to make puck tracking something that doesn’t require a million-dollar setup,” said Salass, lead author of the study. “If a coach can analyze a game using only video, that’s a big win for accessibility in sports analytics. Finding the puck in broadcast video is one of the toughest problems in sports vision, so seeing our system accurately predict its location using contextual cues was incredibly rewarding. It was like we’d given computers real game sense.” The second innovation, SportMamba, is an adaptive AI framework designed to track multiple players simultaneously during dynamic gameplay. It uses state space models to anticipate movements and handle challenges such as rapid motion, blocked camera angles, and shifting camera perspectives. Tested across hockey, soccer, and basketball footage, SportMamba outperformed current tracking technologies by up to 18% in accuracy and efficiency, enabling real-time, data-driven performance analysis without reliance on wearable sensors or fixed-camera networks. “Tracking a hockey player on a breakaway is relatively easy,” said Dr. John Zelek. “It’s much more difficult to track and differentiate players in a scrum along the boards or in front of the net. SportMamba can tackle these difficult situations and tell us, for example, who deflected the puck and scored.” The research, presented at the 2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops and published on arXiv, marks a major step forward in automated sports analytics. By combining contextual intelligence with adaptive tracking, the tools promise to empower coaches with deeper insights, enhance broadcast storytelling, and democratize access to advanced analytics across all levels of hockey and team sports.