# WSSS4LUAD: lung adenocarcinoma H&E patches for weakly supervised tissue segmentation (tumor, stroma, normal)

## Facts

- Provider: Chu Han, Xipeng Pan et al. (WSSS4LUAD organizers); repackaged by MedOtter
- License: CC-BY-4.0 (https://wsss4luad.grand-challenge.org/License/)
- Size: 6 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/MedOtter/WSSS4LUAD-v2
- Access: public, readable without an account
- Category: Health & Biomedical
- Source: https://arxiv.org/abs/2204.06455
- Page: https://scigantic.com/archive/a6878e21-5392-5275-8e60-36f105fd1b60
- Keywords: Health & Biomedical, histopathology, image segmentation, Parquet, Hugging Face

## About

WSSS4LUAD is the dataset of the Grand Challenge on weakly supervised tissue semantic segmentation for lung adenocarcinoma (LUAD): H&E-stained histopathology patches where training images carry only image-level labels and the model must still produce pixel-level tissue maps. The 10,091 training patches are labeled with which of three tissue classes they contain (tumor epithelial tissue, tumor-associated stroma, normal tissue); 40 validation and 80 test patches have pixel-level masks plus a background mask for unstained areas. Each row has the PNG patch, filename, width and height, the three binary labels (training) or the masks (validation and test). Three Parquet files (train, validation, test), about 1.6 GB (measured 2026-11-21); read with Hugging Face datasets or pandas plus Pillow. Use for weakly supervised semantic segmentation, class activation map methods and tumor microenvironment analysis in digital pathology. The challenge site states the dataset is licensed CC BY 4.0. Cite Han et al., WSSS4LUAD: Grand Challenge on Weakly-supervised Tissue Semantic Segmentation for Lung Adenocarcinoma, arXiv:2204.06455 (2022).

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("MedOtter/WSSS4LUAD-v2")
print(ds)
```

## Data structure

File types in sampled files: Parquet

Example file: `test.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/a6878e21-5392-5275-8e60-36f105fd1b60
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/a6878e21-5392-5275-8e60-36f105fd1b60
- Over MCP: `get_schema_card` with id `a6878e21-5392-5275-8e60-36f105fd1b60` at https://api.scigantic.com/mcp

## Related topics

- [Parquet datasets](https://scigantic.com/topics/parquet)
- [Image segmentation datasets](https://scigantic.com/topics/image-segmentation)
- [Health & Biomedical datasets](https://scigantic.com/topics/health-and-biomedical)
- [Histopathology datasets](https://scigantic.com/topics/histopathology)
