This project focuses on extracting meaningful play-by-play information from segments of basketball video by identifying salient actions such as passes, shots, and rebounds. By combining statistical methods with recent advances in vision-language models, the system enables automated understanding of basketball gameplay for analytics, broadcasting, and sports intelligence applications.
Scientific coordinators: Gabriele Giudici, Andrea Maurino, Paola Zuccolotto
Giudici G., Maurino A., Zuccolotto P.
In this work, we present the BARD dataset. It is designed to advance Basketball Action Recognition task. Our contributions include significantly improved annotation accuracy and the integration of additional variables, such as player jersey numbers and team color information with respect to existing datasets. We also introduce a novel output format that supports multi-class classification tasks. Finally, we conducted a human validation study on a subsample of the annotations, with expert reviewers assessing the labeling quality and reporting the evaluation results.
Author: Gabriele Giudici
Supervisor: Andrea Maurino
Second Supervisor: Paola Zuccolotto
