TTA_Tracking

Dataset Description

TTA_Tracking is a ball tracking dataset for table tennis, recorded under realistic conditions at a professional Paralympics level. The dataset is intended to support research in sports video understanding, object detection, and ball tracking under challenging real-world conditions.

Dataset Summary

• Task: Ball tracking / object detection
• Sport: Table Tennis (Paralympics)
• Recording conditions: Realistic professional match conditions
• License: CC BY-NC 4.0 (Research use only, no redistribution)

Usage

This dataset is intended for research purposes only. Commercial use and redistribution are strictly prohibited.

Citation

If you use this dataset in your research, please cite the following work:

• Paper: TOTNet: Occlusion-aware temporal tracking for robust ball detection in sports videos

@article{xu2026totnet,
  title={TOTNet: Occlusion-aware temporal tracking for robust ball detection in sports videos},
  author={Xu, Hao and Baniya, Arbind Agrahari and Wells, Sam and Bouadjenek, Mohamed Reda and Dazeley, Richard and Aryal, Sunil},
  journal={Computer Vision and Image Understanding},
  pages={104657},
  year={2026},
  publisher={Elsevier}
}

License

This dataset is licensed under CC BY-NC 4.0. Use of this dataset is permitted for non-commercial research purposes only. Redistribution in any form is not permitted without explicit written permission from the authors.