BasketEvent: Understanding Who Did What and When in Basketball Videos

2026-07-23

AI Summary

Existing methods struggle to simultaneously and accurately recognize event semantics, responsible players, and precise temporal boundaries in complex basketball videos. To address this challenge, this work introduces the player-centric BasketEvent dataset and proposes PlayNet, a novel framework that explicitly associates event labels with accountable players and provides fine-grained temporal annotations. PlayNet integrates multi-object tracking, player identity association, and hierarchical relational modeling—capturing interactions among players, between players and the ball, and across the global court context—and employs gated pooling to aggregate sparse temporal evidence. This unified approach enables joint spatial awareness and semantic reasoning. Experiments demonstrate that PlayNet significantly outperforms both video-level and cropped-region baselines, validating the efficacy of the player-centered paradigm for fine-grained basketball event understanding.

Abstract

Comprehensive basketball video understanding requires resolving not only what event occurs, but also who is responsible and when the key evidence appears. However, exist- ing methods typically treat spatial perception and semantic recognition as isolated tasks, failing to ground events to individual players or pinpoint their temporal boundaries within complex collective dynamics. To bridge this gap, we introduce BasketEvent, a player- centric basketball event understanding dataset curated from real NBA broadcasts. In BasketEvent, event labels are grounded to the responsible players, and a manually an- notated subset of 1,000 samples with precise event intervals is provided to evaluate tem- poral evidence localization. Based on this data, we propose PlayNet, a player-centric reasoning framework that maps basketball videos to player-level event predictions with temporal evidence. Concretely, PlayNet tracks key entities, associates player identities, and reasons about events by modeling player-player, player-ball, and global court inter- actions, while aggregating sparse temporal evidence via gated pooling. Extensive experi- ments demonstrate that PlayNet significantly outperforms representative video-level and crop-based baselines, proving the superiority of player-centric modeling for fine-grained sports video understanding. Our data, code, and models will be made publicly available.

Problem

• basketball video understanding
• player-centric event grounding
• temporal localization
• sports video analysis
• event attribution

Innovation

• player-centric modeling
• temporal evidence localization
• sports video understanding
• event grounding
• interaction modeling