SoccerNet-v2: A Dataset and Benchmarks for Holistic Understanding of Broadcast Soccer Videos

TL;DR

SoccerNet-v2 introduces a large-scale annotated dataset of broadcast soccer videos with new tasks and benchmarks to advance holistic understanding of soccer videos in computer vision.

Contribution

It provides a comprehensive dataset with 300k annotations, extending tasks to include action spotting, shot segmentation, and replay grounding, with benchmark results for each.

Findings

300k annotations for soccer videos

Extended tasks including replay grounding

Benchmark results provided for each task

Abstract

Understanding broadcast videos is a challenging task in computer vision, as it requires generic reasoning capabilities to appreciate the content offered by the video editing. In this work, we propose SoccerNet-v2, a novel large-scale corpus of manual annotations for the SoccerNet video dataset, along with open challenges to encourage more research in soccer understanding and broadcast production. Specifically, we release around 300k annotations within SoccerNet's 500 untrimmed broadcast soccer videos. We extend current tasks in the realm of soccer to include action spotting, camera shot segmentation with boundary detection, and we define a novel replay grounding task. For each task, we provide and discuss benchmark results, reproducible with our open-source adapted implementations of the most relevant works in the field. SoccerNet-v2 is presented to the broader research community to help push computer vision closer to automatic solutions for more general video understanding and production purposes.