Intel AI helps coaches analyse basketball games
March 28, 2019
College basketball teams are using artificial intelligence (AI) technology from New York-based Keemotion and Intel to capture and analyse live game footage.
Team coaches can make adjustments during games and practices in a way not previously possible. This technology helped catapult Yale to the Ivy League title and its fourth NCAA Tournament appearance.
"As a coaching staff, we want to put our players in a position to succeed,” said James Jones, Yale head basketball coach. “The ability to analyse game and practice footage in real time with Keemotion's solution allows me to teach in the moment, and is a real competitive advantage."
Keemotion uses three cameras to capture video of the court, scoreboard and surroundings. The cameras track player movement, following the action as though controlled by a human operator. An Intel Xeon scalable processor-powered server stitches the footage into a single panoramic view of the court, then the Intel AI algorithm identifies and labels every scoring play and possession, making it easy for coaches to locate and review specific actions in real time.
Keemotion's AI algorithm was trained on the Intel AI DevCloud, which allowed for multiple models to be verified and fine-tuned in parallel.
"By working with Intel Xeon processors, we are able to deliver analysed, searchable, panoramic video in less than three seconds, allowing coaches to have immediate access to a specific play or action," said Milton Lee, CEO of Keemotion.
Keemotion's offering, powered by Intel Xeon scalable processors optimised with Intel’s maths kernel library software, allows tablet-armed coaches to re-watch a play immediately or focus on a specific point on the court. This provides coaches with the information needed both to make adjustments in real time and understand where their players are and which plays may need more practice.