First place in the CVPR 2023 SoccerNet Challenge

10 Juli 2023

In the field of sports analytics an important task is the performance measurement of individual players. From a machine learning perspective this involves the identification and tracking of players during a match based on video footage. In team sports, the identification of the jersey number (Number Recognition) can be beneficial in order to distinguish between the individual appearance of each player (Re-Identification). 

With their contrastive learning approach Konrad Habel, Fabian Deuser, and Norbert Oswald won the first place in the SoccerNet Re-Identification Challenge and the second place of the Number Recognition Challenge at the Workshop on Computer Vision in Sports at CVPR 2023. In addition, the solution for Number Recognition was awarded with the Jury Prize for the most innovative runner-up method.

SoccerNet 2023 Challenge Results
Konrad Habel, Fabian Deuser, Norbert Oswald

[arXiv]