---
license: cc-by-nc-4.0
task_categories:
- image-segmentation
language:
- en
tags:
- medical
- image
pretty_name: 'autopet-iii-lite'
size_categories:
- n<1K
---


## About
This is a subset of the [autoPET-III dataset](https://autopet-iii.grand-challenge.org/dataset/).
- 1038 pairs of CT and PET images, and corresponding segmentation masks
- No change to any image or segmentation mask
- Files are rearranged into `Images-PET`, `Images-CT`, and `Masks` folders
- Images with empty segmentation masks are excluded from the official release

This dataset is released under the `CC BY-NC 4.0` license.


## Official Release
For more information, please go to these sites:
- Data License (official): [CC BY-NC 4.0](https://autopet-iii.grand-challenge.org/rules/)
- Challenge (official): https://autopet-iii.grand-challenge.org
- Data (official): https://autopet-iii.grand-challenge.org/dataset/


## Segmentation Labels
```python
labels_map = {"1": "tumor"}
```


## Download from Huggingface
```bash
#!/bin/bash
pip install huggingface-hub[cli]
huggingface-cli login --token $HF_TOKEN
```
```python
# python
from huggingface_hub import snapshot_download
snapshot_download(repo_id="YongchengYAO/autoPET-III-Lite", repo_type='dataset', local_dir="/your/local/folder")
```