Skip to main content

Soccer Summarization

Soccer game captions and summary in English for game summarization.

PublicationGitHub repoLast updated Jun 2026

This page contains information for the datasets mentioned in the paper Soccer Game Summarization using Audio Commentary, Metadata, and Captions .

📚 Datasets

This dataset contains the following:

  • SoccerNet-BBC (Games form SoccerNet, extanded with news, commentaries and lineups from BBC)
  • K-SportsSum-EN (Translated, CN to EN, captions and summary)
  • SportsSum-EN (Translated, CN to EN, captions and summary)

You can access the datasets right here.

💻 Audio Intensity Visualization Dashboard

The codes for Audio Intensity Visualization Dashboard as mentioned in the paper can be access right here.

📎 Cite

If you use contents from this in your research, Please cite the following paper:

    @incollection{Gautam2022Oct,
        author = {Gautam, Sushant and Midoglu, Cise and Shafiee Sabet, Saeed and Kshatri, Dinesh Baniya and Halvorsen, P{\aa}l},
        title = {{Soccer Game Summarization using Audio Commentary, Metadata, and Captions}},
        booktitle = {{NarSUM '22: Proceedings of the 1st Workshop on User-centric Narrative Summarization of Long Videos}},
        pages = {13--22},
        year = {2022},
        month = oct,
        date = {2022-10-10},
        urldate = {2022-10-10},
        isbn = {978-1-45039493-2},
        publisher = {Association for Computing Machinery},
        address = {New York, NY, USA},
        doi = {10.1145/3552463.3557019}
    }
    

⚖ Terms of Use ️

The data is released fully open for research and educational purposes. The use of the dataset for purposes such as competitions and commercial purposes needs prior written permission. In all documents and papers that use or refer to the dataset or report experimental results, a reference to the related article needs to be added: https://dl.acm.org/doi/10.1145/3552463.3557019.

👋 Contact

Please contact sushant@simula.no, cise@simula.no, or paalh@simula.no for any questions regarding the dataset. We always welcome collaboration and joint research!