This article provides an in-depth review of action recognition, spotting, and spatio-temporal localization in soccer. It discusses the complexities of understanding actions in soccer due to the dynamic nature of the game and player interactions. The review focuses on multimodal methods that integrate information from multiple sources like video and audio data. It also examines the potential of these methods to improve the accuracy and robustness of models. The article also highlights open research questions and future research directions in the field of soccer action recognition.
Publication date: 22 Sep 2023
Project Page: ?
Paper: https://arxiv.org/pdf/2309.12067
