AllClip Wins 3rd Place Globally in "Spiideo SoccerNet SynLoc 2026" Competition at CVPR 2026

AllClip Inc. (Headquarters: Chiyoda-ku, Tokyo; Representative Director: Ikuma Uchida; hereinafter referred to as "AllClip") has won 3rd place globally in the "Spiideo SoccerNet SynLoc 2026" competition at CVPR 2026, one of the largest international conferences in the field of computer vision.

CVPR is one of the world's foremost international conferences in computer vision and the largest academic society gathering top-level researchers in image recognition technology. The "SoccerNet Challenge," a competition to measure the accuracy of AI processing for soccer footage, is held annually alongside CVPR. This year, 85 teams from around the world participated in one of its divisions, "Spiideo SoccerNet SynLoc 2026," with our team achieving a 3rd place global ranking.

Competition Overview

"Spiideo SoccerNet SynLoc 2026" is a task that detects players from a single frame of soccer footage and determines their position in the world coordinate system. This is a crucial element technology for building video recognition and understanding AI technologies to automate sports analysis. The evaluation design emphasizes not only the accuracy of player detection but also the precision of their position on the field.

In our approach, we significantly surpassed the baseline methods by developing a module that effectively detects small objects like "players" within 4K images, utilizing player pose information, and tuning player positions considering camera parameters.

Future Prospects

We aim to leverage the techniques used in this competition to further improve the accuracy and add new features to our AI camera system "AllClip," applying them directly to solve challenges in the sports field.

Furthermore, beyond the scope of this competition, we expect to contribute to the advancement of video AI technologies that recognize and understand sports scenes, leading to a future where the value traditionally enjoyed only in top-level environments, such as extracting individual player scenes or measuring physical data, can be delivered to all categories of sports.

We will continue to strive for further improvements in sports video analysis technology and camera systems to transform all sports scenes into video and data experiences, promoting the provision of user-friendly and practical AI sports systems for competition venues.

About AllClip Inc.

One camera becomes an AI sports camera. Easily automate game filming, editing, and sharing.

AllClip combines action cameras with AI technology to automate the filming, editing, and sharing of games and practices. We are working to lower the hurdles for filming, editing, and utilizing video footage in sports, creating an environment where more teams can leverage game footage for their growth.

If you are interested in AllClip, please feel free to contact us.

【Contact Information】

AllClip Inc.

Contact: Public Relations

Email: [email protected]

HP: https://allclip.com