SoccerNet 2025 Challenges Results

2025-08-26

AI Summary

SoccerNet 2025 Challenges address four core visual problems in football video understanding: (1) team-ball action detection, (2) monocular depth estimation, (3) multi-view foul recognition, and (4) 2D top-down match state reconstruction. Methodologically, the challenges integrate temporal modeling, multi-view geometry, semantic segmentation, and pose estimation, augmented by fine-grained action localization and multi-view collaborative analysis. The primary contributions include establishing a unified evaluation protocol and a large-scale, densely annotated benchmark—enhancing openness and reproducibility in sports vision research. State-of-the-art solutions across all tasks substantially surpass baselines, achieving notable advances in action attribution under occlusion and motion ambiguity, geometric 3D recovery from monocular inputs, and severity-aware foul classification. These improvements collectively accelerate the transition of football video understanding from algorithmic exploration to real-world deployment.

Technology Category

• Computer Vision: Video Understanding & Activity Analysis
• Intelligent Robots: State Estimation
• Machine Learning: Multi-instance/Multi-view Learning

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

The SoccerNet 2025 Challenges mark the fifth annual edition of the SoccerNet open benchmarking effort, dedicated to advancing computer vision research in football video understanding. This year's challenges span four vision-based tasks: (1) Team Ball Action Spotting, focused on detecting ball-related actions in football broadcasts and assigning actions to teams; (2) Monocular Depth Estimation, targeting the recovery of scene geometry from single-camera broadcast clips through relative depth estimation for each pixel; (3) Multi-View Foul Recognition, requiring the analysis of multiple synchronized camera views to classify fouls and their severity; and (4) Game State Reconstruction, aimed at localizing and identifying all players from a broadcast video to reconstruct the game state on a 2D top-view of the field. Across all tasks, participants were provided with large-scale annotated datasets, unified evaluation protocols, and strong baselines as starting points. This report presents the results of each challenge, highlights the top-performing solutions, and provides insights into the progress made by the community. The SoccerNet Challenges continue to serve as a driving force for reproducible, open research at the intersection of computer vision, artificial intelligence, and sports. Detailed information about the tasks, challenges, and leaderboards can be found at https://www.soccer-net.org, with baselines and development kits available at https://github.com/SoccerNet.

Problem

• Detecting ball-related actions in football broadcasts and assigning team actions
• Recovering scene geometry from single-camera clips via pixel depth estimation
• Analyzing multiple camera views to classify fouls and their severity levels

Innovation

• Team Ball Action Spotting detection
• Monocular Depth Estimation geometry recovery
• Multi-View Foul Recognition classification