README.md

metadata

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

About

This is a subset of the CrossMoDA dataset.

• 105 contrast-enhanced T1 MR images and corresponding segmentation masks
• No change to any image or segmentation mask
• Files are rearrange into Images and Masks folders
• This is labeled data in the training set of CrossMoDA dataset

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

Segmentation Labels

labels_map = {
    "1": "vestibular schwannoma (acoustic neuroma)",
    "2": "cochlea",
}

Official Release

For more information, please refer to the offical sites:

• Challenge (official): https://crossmoda-challenge.ml
• Data (official): https://zenodo.org/records/4662239
• Paper: https://doi.org/10.1016/j.media.2022.102628

Download from Huggingface

#!/bin/bash
pip install huggingface-hub[cli]
huggingface-cli login --token $HF_TOKEN

# python
from huggingface_hub import snapshot_download
snapshot_download(repo_id="YongchengYAO/CrossMoDA-Lite", repo_type='dataset', local_dir="/your/local/folder")