This week, the Philadelphia Eagles and Kansas City Chiefs are spending countless hours in film rooms in preparation for the forthcoming Super Bowl. However, BYU engineers have developed a new AI technology that could significantly reduce the time and expense of film analysis for NFL and college teams.

Professor D.J. Lee and his team of students at BYU have developed an algorithm capable of accurately locating and labelling players from game footage and determining the offensive team formation. Using deep learning and computer vision, the team has achieved over 90 percent accuracy in player detection and labelling, and 85 percent accuracy in determining formations.

Lee and Newman utilised actual game footage from the BYU football team before purchasing Madden 2020 and manually labelling 1,000 images and videos to further train their algorithm. When player locations and labels were accurate, the team was able to achieve a 99.5% success rate.

The AI system could also be utilised in other sports, including baseball, soccer, and basketball, to assist teams in refining their defensive strategies against specific batters or players, or in determining more effective formations. The researchers believe that this technology could one day eliminate the need for the inefficient and laborious manual annotation and analysis of recorded video.

“Big data can help us understand this team’s strategies and this coach’s tendencies,” said Lee. “It could help you determine whether they will go for it on 4th Down and 2 or punt. It will be worthwhile if we can give them even a 1% advantage through the use of artificial intelligence in sports.”

As players and coaches prepare for the upcoming Super Bowl, engineers from Brigham Young University are developing an artificial intelligence system that could revolutionise sports analytics. The algorithm is capable of accurately locating and labelling players from game footage, as well as determining offensive formations with an accuracy of over 90%. The technology could eliminate the need for manual annotation and analysis of recorded video, providing teams with a strategic advantage that could help them win games.

About The Author

Leave a Comment

Your email address will not be published. Required fields are marked *