metadata

license: cc-by-nc-nd-4.0

TACDEC: Dataset of Tackle Events in Soccer Game Videos

TACDEC is a dataset of tackle events in soccer game videos. Recognizing the gap in existing open datasets that predominantly focus on official soccer events such as goals and cards, TACDEC targets a comprehensive analysis of tackles — a critical aspect of soccer that combines technical skills, tactical decision-making, and physical engagement. By leveraging video data from the Norwegian Eliteserien league across multiple seasons, we annotated 425 videos with 4 types of tackle events, categorized into "tackle-live", "tackle-replay", "tackle-live-incomplete", and "tackle-replay-incomplete", yielding a total of 836 event annotations. The dataset offers an unprecedented resource for the development and testing of machine learning models aimed at understanding and analyzing soccer game dynamics.

Zenodo

The dataset is also available on Zenodo: https://zenodo.org/records/10611979.

Terms of Use

The dataset is fully open for research and educational purposes. Use of the dataset for competitions or commercial purposes requires prior written permission. References to the related article must be included in all documents and papers that use, refer to, or report experimental results from this dataset.

How to download dataset:

Everything that is related directly to the dataset is located within this repo. For instructions related to testing the model used or verifying results from TACDEC-paper, see TACDEC-model repo.

What you can find in this repo is:

All 425 videos
All 425 labels

In addition:

A torch, sorted_cls_tokens_feature.pt, containing all the CLS-tokens (features) from DINOv2 used concatenated in a sorted order
A numpy array, sorted_cls_tokens_labels.npy, containing all the labels used, concatenated in the same sorted order

We highly recommend downloading these if you want to implement/test the model, as running the large DINOv2 model is highly computationally expensive.

Citation:


@incollection{Kassab_MMSYS_ODS,
  author = {Jåsund Kassab, Evan and Maric Solberg, Håkon and Gautam, Sushant and Shafiee Sabet, Saeed and Torjusen, Thomas and Riegler, Michael and Halvorsen, Pål and Midoglu, Cise},
  title = {{TACDEC: Dataset of Tackle Events in Soccer Game Videos}},
  booktitle = {{MMSys'24 : The 15th ACM Multimedia Systems Conference}},
  year = {2024},
  month = apr,
  date = {2024-04-15},
  urldate = {2024-04-15},
  isbn = {979-8-4007-0412-3/24/04},
  publisher = {Association for Computing Machinery},
  address = {New York, NY, USA},
  doi = {10.1145/3625468.3652166}
}

Acknowledgement:

This research was partly funded by the Research Council of Norway, project number 346671 (AI-storyteller). We would like to thank the Norwegian Professional Football League ("Norsk Toppfotball") for making videos available for the research.

Contact

For any questions regarding the dataset, or to discuss potential collaboration and joint research opportunities, please get in touch with the following people:

Evan Jåsund Kassab: evanjk@uio.no
Cise Midoglu: cise@simula.no
Pål Halvorsen: paalh@simula.no