We’re really looking forward to CVPR this year! We hope that you’ll come by our booth (#300) to chat with our research team and learn more about how SPORTLOGiQ is applying AI to sports and shaking up the industry.
SPORTLOGiQ is an AI-powered sports analytics company transforming the way teams and fans experience the game. We use computer vision and machine learning to track and analyze the movements of every player and generate deep data insights. We track over 158 million data points per game and help professional sports teams win more games and broadcasters tell better stories.
Using standard, single-camera game footage, we flag specific game events—shots, passes and possessions—timestamp them, and record their XY coordinates. The player-location data is then transformed into meaningful and actionable insights used by experts—from coaches and players to managers and scouts to analysts and broadcasters.
Our end goal is to understand a player’s behaviour. Computer vision, or more precisely, tracking, activity recognition, and group activity analysis tell us where everyone is and what they are doing at any given time. For example, we can identify the outcome a team is seeking when applying a specific tactic or a player’s objective when making a pass. This information can be used to infer who is going to score the next goal, win the game, or win the Stanley Cup.
About the Team
Starting in hockey, SPORTLOGiQ has now expanded into soccer/football and beyond. In particular, our Kitchener-Waterloo AI lab is focused on soccer/football research. Our computer vision and machine learning team is based out of two offices and comprised of some of the best and brightest researchers in the field. Additionally, we fund and guide early stage research at 7 of the leading academic labs across Canada, which has led to 6 patents in 3 years. We were also named one of the top AI companies in the world by CB Insights.
SPORTLOGiQ is solving problems that are both interesting and groundbreaking. We’re collaborating with some of the top researchers in academia like Greg Mori and Jim Little, our in-house technology adviser. Together, we’re tackling a variety of complex problems ranging from traditional computer vision to reinforcement learning to natural language processing. This has allowed us to really move the needle when it comes to computer vision research.
More specifically, we use camera localization and mapping, real-time tracking, activity recognition and semantic segmentation for advertising, augmented reality and visualizing games from a player’s standpoint. Using hierarchical/multi-level multi-agent reinforcement learning approaches, we are now able to imagine hypothetical situations, do player assessments, player profiling, and predict games and outcomes.
At SPORTLOGiQ, we track and analyze the interactions between players. They have teammates, opponents, a strategy, and defined tactics. If we can observe, measure, quantify, and understand human behaviour in sports, then we can start thinking about generalizing everything to understand human behaviour in the real world. We’re building complex models that don’t just measure but truly understand human interactions. This technology will enable solutions to some of the biggest challenges the world will face in the future.