In this presentation we talk about our graph-based model for action spotting in soccer games. The goal of action spotting is to identify the specific time an action occurs. Our model represents players, referees, and goalkeepers as nodes in a graph and connects according to their distance in the field. We describe in detail our methods for extracting this information from the frames and each component of our architecture. Finally, we present the experiments we carry on the SoccerNetV2 benchmark and show that our performance (57.83% mAP accuracy) surpasses similar graph-based methods and has competitive results with heavy computing methods.
Alejandro Cartas is a Postdoc at the Universitat Pompeu Fabra since 2021, working in the Intelligent Multimodal Vision Analysis (IMVA) group. Prior to that, he received a Ph.D. (cum laude) in Informatics & Mathematics from the University of Barcelona, Spain in 2020. His current research interests lie in the automatic understanding of sports video sequences. The ultimate goal of this field is to develop algorithms capable of understanding sport scenes so they can be used effectively in real world applications like video summarization, predictive analytics, and team and players performance analytics.
