NVIDIA Expands AI Tools for Athlete Tracking and Sports Broadcasting

NVIDIA is expanding its presence in sports technology with artificial intelligence tools designed to improve athlete tracking, video analysis and live broadcast production.

The technology includes 3D Body Pose, which estimates human joint positions and movement in three dimensions using footage from a single camera. The system can support performance analysis, biomechanics, officiating applications and augmented broadcast graphics without requiring traditional marker-based motion capture.

Another development, Video Frame Generation, uses generative AI to create additional frames between existing video images. This can increase frame rates by two or four times, enabling smoother slow-motion replays and other high-motion sports content.

NVIDIA is also developing its Sports Intelligence Playbooks, which allow leagues, broadcasters and technology companies to customise AI models using proprietary sports footage, performance data and annotations.

The technology is designed to help organisations build systems that understand the rules, players, scoring and tactical context of individual sports, rather than relying solely on general-purpose AI models.

Industry partners including Hawk-Eye, Stats Perform, Vizrt and Wowza are among the companies working within NVIDIA's wider sports and media technology ecosystem.

NVIDIA is pursuing a technology-platform approach, providing the computing infrastructure, software and development tools that other companies can incorporate into their products.

The developments create further opportunities for sports organisations to use existing video libraries and live footage for performance intelligence, broadcast enhancement and digital content production.