Please check out the SportsPose website to download the dataset.
You can read our papers here:
- SportsPose - A Dynamic 3D sports pose dataset
- Two Views Are Better than One: Monocular 3D Pose Estimation with Multiview Consistency
If you use our work, please cite the following
BibTeX:
@inproceedings{ingwersen2023sportspose,
title={SportsPose: A Dynamic 3D Sports Pose Dataset},
author={Ingwersen, Christian Keilstrup and Mikkelstrup, Christian and Jensen,
Janus N{\o}rtoft and Hannemose, Morten Rieger and Dahl, Anders Bjorholm},
booktitle={Proceedings of the IEEE/CVF International Workshop on Computer Vision in Sports},
year={2023}
}
@inproceedings{ingwersen2025two,
title={Two views are better than one: Monocular 3d pose estimation with multiview consistency},
author={Ingwersen, Christian Keilstrup and Tirsgaard, Rasmus and Nylander, Rasmus and Jensen, Janus Nortoft and Dahl, Anders Bjorholm and Hannemose, Morten Rieger},
booktitle={Proceedings of the Computer Vision and Pattern Recognition Conference},
pages={5915--5925},
year={2025}
}
To install the packages needed for the dataset and example, simply clone this repository and pip install it. I.e.
git clone https://github.com/ChristianIngwersen/SportsPose.git
cd SportsPose
pip install .
Examples on how to use the provided PyTorch dataset can be found in the notebook: example_notebook.ipynb