Challenges

Discover our Challenges

The SoccerNet Challenges are a great opportunity for you to discover our dataset and tasks, while benchmarking your solutions with others. The challenges are open for anyone , regardless of your affiliation to academia or the industry. Feel free to pick one or several task(s) and start by trying the baselines that we provide. Then, you can either adapt our methods, or come up with a totally new one and try to reach the top of the leaderboard... Good luck fighting for the throne!

In 2024, we propose 4 challenges, we re-conducted the Ball Action Spotting and Dense Video Captioning tasks and proposed 2 new challenges: Multi-View Foul Recognition and Game State Reconstruction. 

In 2023, we propose 7 challenges, split into 3 main categories: Broadcast Video Understanding, Field Understanding, and Player Understanding. We re-conducted Action Spotting, Camera Calibration, Player Tracking, and Player Re-Identification tasks and proposed 3 new challenges: Ball Action Spotting, Dense Video Captioning, and Player Jersey Number Recognition. In total, 5000$ worth of prizes will be granted to the winner of each challenge, sponsored by EVS Broadcast Equipment, SportRadar, Baidu Research, and AINSports. The results of this challenge will be presented during the CVPR 2023 Workshop CVSports.

In 2022, we proposed 4 challenges around 6 tasks! Camera Calibration and Field Localization will be the guest competition of the CVsports workshop at CVPR 2022. Action Spotting and Replay Grounding are once again featured at the ActivityNet workshop of CVPR 2022. We also propose novel Player Tracking and Re-Identification challenges, which should be featured at the ActivityNet workshop of CVPR 2022. In total, 4000$ worth of prizes will be granted to the winner of each challenge, sponsored by EVS Broadcast Equipment, SportRadar, and Baidu Research.

Action Spotting and Replay Grounding were featured at the ActivityNet workshop of CVPR 2021. The challenges had a 1000$ prize sponsored by Second Spectrum. The winners of both challenges are Baidu research.