NEATX: Non-Expert Annotations of Tubes in X-rays

The Non-Expert Annotations of Tubes in X-rays (NEATX) dataset was created at PURRlab as part of a MSc thesis of Trine Naja Eriksen and Cathrine Damgaard.

This dataset contains 3.5k chest drain annotations for the NIH-CXR14 dataset, and 1k annotations for four different tube types (chest drain, tracheostomy, nasogastric, and endotracheal) in the PadChest dataset by two data science students.

Please read more about how the dataset can be used in https://arxiv.org/abs/2309.02244.
Bibtex:

@inproceedings{cheplygina2025augmenting,
  title={Augmenting Chest X-ray Datasets with Non-Expert Annotations},
  author={Cheplygina, Veronika and Damgaard, Cathrine and Eriksen, Trine Naja and Juodelyte, Dovile and Jim{\'e}nez-S{\'a}nchez, Amelia},
  booktitle={Annual Conference on Medical Image Understanding and Analysis},
  pages={133--144},
  year={2025},
  organization={Springer}
}

Data description:

This dataset contains the annotations provided by two data science students (not medical experts) for:

• A csv file with 3.5k chest drain annotations for NIH-CXR14 dataset
• A csv file with 1k annotations for four different tube types (chest drain, tracheostomy, nasogastric, and endotracheal) for PadChest dataset

We provide the raw individual annotations as well as the aggregated annotations. The annotation protocol is described in the Healthsheet.