DSB17

Curated release of Data Science Bowl 2017–related public thoracic CT data hosted for research use.

Dataset viewer (Hugging Face)

The Dataset viewer loads only data/train-00000-of-00001.parquet: a tabular index (subject_id, ct_path, original_name, volume_shape). NIfTI volumes are not rendered in the browser. Download ct.nii.gz from the paths in ct_path on the Files tab.

Row-level metadata from the previous root metadata.csv is preserved as case_index.csv (same columns: original_name, renamed_id, shape).

OpenVAE / SMILE project

This dataset is published alongside the OpenVAE open-source stack for medical CT/MRI latent modeling and diffusion-style pipelines. The same public cohorts are used in our benchmarking and demo workflows.

• OpenVAE (code): github.com/KumaKuma2002/OpenVAE
• OpenVAE (pretrained VAEs on Hugging Face): huggingface.co/SMILE-project/OpenVAE

Main paper (cite for OpenVAE / SUMI): arXiv:2604.07329 — Distilling Photon-Counting CT into Routine Chest CT through Clinically Validated Degradation Modeling · PDF

Citation

If you use this release with our OpenVAE / SUMI stack or benchmarks, cite the paper above:

@misc{liu2026distillingphotoncountingctroutine,
  title={Distilling Photon-Counting CT into Routine Chest CT through Clinically Validated Degradation Modeling},
  author={Junqi Liu and Xinze Zhou and Wenxuan Li and Scott Ye and Arkadiusz Sitek and Xiaofeng Yang and Yucheng Tang and Daguang Xu and Kai Ding and Kang Wang and Yang Yang and Alan L. Yuille and Zongwei Zhou},
  year={2026},
  eprint={2604.07329},
  archivePrefix={arXiv},
  primaryClass={cs.CV},
  url={https://arxiv.org/abs/2604.07329},
}

Related (earlier anatomy-aware contrast diffusion): arXiv:2512.07251

Please also follow the original Data Science Bowl 2017 / data provider citation and license requirements when using the underlying challenge data.