Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

OSL-ActionSpotting: A Unified Library for Action Spotting in Sports Videos

Yassine Benzakour, Bruno Cabado, Silvio Giancola, Anthony Cioppa, Bernard Ghanem, Marc Van Droogenbroeck

TL;DR

OSL-ActionSpotting addresses fragmentation in sports video action spotting by unifying major algorithms into a modular Python library. It provides a plug-and-play backbone–neck–head framework, GPU-accelerated data loading via DALI, and a JSON-based dataset format to standardize datasets. The library reproduces state-of-the-art results on SoccerNet-v2 while enabling faster preprocessing and easier cross-dataset experimentation. This work offers a practical, scalable tool to accelerate research and deployment of action spotting in diverse sports contexts.

Abstract

Action spotting is crucial in sports analytics as it enables the precise identification and categorization of pivotal moments in sports matches, providing insights that are essential for performance analysis and tactical decision-making. The fragmentation of existing methodologies, however, impedes the progression of sports analytics, necessitating a unified codebase to support the development and deployment of action spotting for video analysis. In this work, we introduce OSL-ActionSpotting, a Python library that unifies different action spotting algorithms to streamline research and applications in sports video analytics. OSL-ActionSpotting encapsulates various state-of-the-art techniques into a singular, user-friendly framework, offering standardized processes for action spotting and analysis across multiple datasets. We successfully integrated three cornerstone action spotting methods into OSL-ActionSpotting, achieving performance metrics that match those of the original, disparate codebases. This unification within a single library preserves the effectiveness of each method and enhances usability and accessibility for researchers and practitioners in sports analytics. By bridging the gaps between various action spotting techniques, OSL-ActionSpotting significantly contributes to the field of sports video analysis, fostering enhanced analytical capabilities and collaborative research opportunities. The scalable and modularized design of the library ensures its long-term relevance and adaptability to future technological advancements in the domain.

OSL-ActionSpotting: A Unified Library for Action Spotting in Sports Videos

TL;DR

OSL-ActionSpotting addresses fragmentation in sports video action spotting by unifying major algorithms into a modular Python library. It provides a plug-and-play backbone–neck–head framework, GPU-accelerated data loading via DALI, and a JSON-based dataset format to standardize datasets. The library reproduces state-of-the-art results on SoccerNet-v2 while enabling faster preprocessing and easier cross-dataset experimentation. This work offers a practical, scalable tool to accelerate research and deployment of action spotting in diverse sports contexts.

Abstract

Action spotting is crucial in sports analytics as it enables the precise identification and categorization of pivotal moments in sports matches, providing insights that are essential for performance analysis and tactical decision-making. The fragmentation of existing methodologies, however, impedes the progression of sports analytics, necessitating a unified codebase to support the development and deployment of action spotting for video analysis. In this work, we introduce OSL-ActionSpotting, a Python library that unifies different action spotting algorithms to streamline research and applications in sports video analytics. OSL-ActionSpotting encapsulates various state-of-the-art techniques into a singular, user-friendly framework, offering standardized processes for action spotting and analysis across multiple datasets. We successfully integrated three cornerstone action spotting methods into OSL-ActionSpotting, achieving performance metrics that match those of the original, disparate codebases. This unification within a single library preserves the effectiveness of each method and enhances usability and accessibility for researchers and practitioners in sports analytics. By bridging the gaps between various action spotting techniques, OSL-ActionSpotting significantly contributes to the field of sports video analysis, fostering enhanced analytical capabilities and collaborative research opportunities. The scalable and modularized design of the library ensures its long-term relevance and adaptability to future technological advancements in the domain.
Paper Structure (13 sections, 3 equations, 2 figures, 1 table)

This paper contains 13 sections, 3 equations, 2 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Work
  3. Sports Video Understanding
  4. Action spotting
  5. OSL-ActionSpotting
  6. Definitions
  7. Action Spotting Algorithms
  8. Data Handling and Processing
  9. Evaluation metrics
  10. User-friendly Tools and Module
  11. Experiments / Benchmark
  12. Discussions
  13. Conclusion

Figures (2)

  • Figure 1: OSL-ActionSpotting is a plug-and-play library that unifies action spotting algorithms. The design of OSL-ActionSpotting is inherently versatile, making it applicable to a broad spectrum of sports video analyses. This adaptability ensures that OSL-ActionSpotting can facilitate the development of novel action spotting techniques, and accelerate the deployment of these methods, providing a robust and comprehensive tool for researchers and analysts in various sports domains.
  • Figure 2: OSL-ActionSpotting contains methods for efficient data loading, modularized action spotting, and comprehensive evaluation. The dataloader provides videos $V$ or frame features $F$. The backbone converts the videos $V$ into frame features $F$. The neck lifts the frame features $F$ into neck features $F'$, further processed by the spotting head that predicts the actions $A$. The evaluation compares the action predictions $A$ with the ground truth $GT$.