MIT and FIFA develop automated system to track out-of-bounds touches
MIT researcher Henry Wang has developed a semi-automated sports officiating system that utilizes computer vision, skeletal data, and ball tracking to identify out-of-bounds events. This research project, developed in collaboration with FIFA, aims to reduce game interruptions through automated detection while maintaining a seamless experience for viewers.
Key Takeaways
- The system utilizes computer vision and skeletal data to determine ball possession at the moment it crosses the pitch boundary.
- Henry Wang’s research reached 82.5% accuracy on 2022 World Cup datasets, including over 90% in aerial duels.
- The technology is designed to deliver real-time alerts to referees, aiming to maintain game flow without visual disruption for fans.
- Development was managed through FIFA Innovation, the governing body's division for testing on-field match technology.
Why It Matters
Automating objective boundary calls represents the next step in digitizing the field of play, moving beyond high-profile offside and goal-line decisions into routine match events. For the streaming and broadcast ecosystem, this provides a new source of real-time metadata that can power instant graphic overlays and clarify controversial calls for viewers. As these systems move from research to implementation, they reduce the 'dead time' associated with manual VAR reviews, enhancing the live product's pacing. Watch for the potential integration of this 'last-touch' logic into the existing 16-camera array used for semi-automated offside tracking in future FIFA tournaments.
Additional Context
The 2026 FIFA World Cup has served as a primary testing ground for an integrated AI officiating stack. Per reports from Forbes and Mexico Business News in June 2026, the tournament employs 1,248 digital player models created from high-resolution 3D body scans. These avatars allow the semi-automated offside technology (SAOT) to track 29 specific data points per player at 50 frames per second. This spatial mapping is fused with data from the official Adidas Trionda match ball, which contains an inertial measurement unit (IMU) sensor sampling at 500Hz to detect microscopic physical contacts, according to Fox Sports and FIFA's technical documentation from July 2026.
Hardware partnerships have been critical to this infrastructure. Lenovo, serving as the official technology partner, provides the edge compute nodes required to stabilize 'Referee View' body camera footage in real time, as noted by MeshWorld in July 2026. This system filters out motion blur from sprinting officials to provide broadcasters with a clear first-person perspective. Meanwhile, the 'Football AI Pro' platform grants all 48 competing teams access to tactical generative AI assistants to analyze match data. While these tools aim to democratize data access, outlets like Forbes have highlighted ongoing debates regarding the commercial rights and privacy of player biometric data generated by these millimeter-accurate 3D scans.
Read full article at news.mit.edu
Enjoy our coverage?
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source