Camera Calibration and Player Localization in SoccerNet-v2 and Investigation of their Representations for Action Spotting
This work solves the problem of camera calibration and player localization for soccer video analysis, enabling improved action spotting, but it is incremental as it builds on existing datasets and architectures.
The authors addressed the lack of a large-scale calibration dataset and public calibration network for soccer broadcast videos by distilling a commercial tool into a neural network on the SoccerNet dataset, and used the resulting representations to achieve new state-of-the-art performance in action spotting.
Soccer broadcast video understanding has been drawing a lot of attention in recent years within data scientists and industrial companies. This is mainly due to the lucrative potential unlocked by effective deep learning techniques developed in the field of computer vision. In this work, we focus on the topic of camera calibration and on its current limitations for the scientific community. More precisely, we tackle the absence of a large-scale calibration dataset and of a public calibration network trained on such a dataset. Specifically, we distill a powerful commercial calibration tool in a recent neural network architecture on the large-scale SoccerNet dataset, composed of untrimmed broadcast videos of 500 soccer games. We further release our distilled network, and leverage it to provide 3 ways of representing the calibration results along with player localization. Finally, we exploit those representations within the current best architecture for the action spotting task of SoccerNet-v2, and achieve new state-of-the-art performances.