README.md

---
license: cc-by-nc-4.0
task_categories:
- image-segmentation
language:
- en
tags:
- medical
- image
- segmentation
- MRI
- knee
- cartilage
pretty_name: oaizib-cm
size_categories:
- n<1K
---


## Data
| Source       | link                                                         |
| ------------ | ------------------------------------------------------------ |
| Huggingface  | [main](https://huggingface.co/datasets/YongchengYAO/OAIZIB-CM/tree/main) |
|              | [load_dataset-support](https://huggingface.co/datasets/YongchengYAO/OAIZIB-CM/tree/load_dataset-support) | 
| Zenodo       | [here](https://zenodo.org/records/14934086)
| Google Drive | [here](https://drive.google.com/drive/folders/13_afAKSH7ZMOI_Nk2gfoihbJKwafw1l9?usp=share_link) |
- Huggingface Dataset Branch:
  - `main`: The main branch contains the same files as those in Zenodo and Google Drive
  - `load_dataset-support`: We added HF `load_dataset()` support in this branch (ref: [intended usage 2](https://huggingface.co/datasets/YongchengYAO/OAIZIB-CM#2-load-dataset-or-iterabledataset-from-the-load_dataset-support-branch-%EF%B8%8F))


## About
This is the official release of **OAIZIB-CM** dataset 
- OAIZIB-CM is based on the OAIZIB dataset
  - OAIZIB paper: [Automated Segmentation of Knee Bone and Cartilage combining Statistical Shape Knowledge and Convolutional Neural Networks: Data from the Osteoarthritis Initiative](https://doi.org/10.1016/j.media.2018.11.009)
- In OAIZIB-CM, tibial cartilage is split into medial and lateral tibial cartilages.
- OAIZIB-CM includes [CLAIR-Knee-103R](https://github.com/YongchengYAO/CartiMorph/blob/main/Documents/TemplateAtlas.md), consisting of 
  - a template image learned from 103 MR images of subjects without radiographic OA
  - corresponding 5-ROI segmentation mask for cartilages and bones
  - corresponding 20-ROI atlas for articular cartilages
- It is compulsory to cite the paper if you use the dataset
  - [CartiMorph: A framework for automated knee articular cartilage morphometrics](https://doi.org/10.1016/j.media.2023.103035)


## Changelog 🔥
- [22 Mar, 2025] Add HF `load_dataset()` support in the `load_dataset-support` branch.
- [27 Feb, 2025] Add the template and atlas [CLAIR-Knee-103R](https://github.com/YongchengYAO/CartiMorph/blob/main/Documents/TemplateAtlas.md)
- [26 Feb, 2025] Update compulsory citation ([CartiMorph](https://doi.org/10.1016/j.media.2023.103035)) for the dataset
- [15 Feb, 2025] Update file `imagesTs/oaizib_454_0000.nii.gz`
- [14 Feb, 2025] Identify corrupted files: case 454


## Files
Images & Labels
- imagesTr: training images (#404)
- labelsTr: training segmentation masks (#404) 
- imagesTs: testing images (#103)
- labelsTs: testing segmentation masks (#103) 

Data Split & Info
- `subInfo_train`: list of training data
- `subInfo_test`: list of testing data
- `kneeSideInfo`: a file containing knee side information, used in CartiMorph Toolbox


## Intended Usage
### 1. Download Files from the `main` or `load_dataset-support` Branch
```bash
#!/bin/bash
pip install --upgrade huggingface-hub[cli]
huggingface-cli login --token $HF_TOKEN
```
```python
# python
from huggingface_hub import snapshot_download
snapshot_download(repo_id="YongchengYAO/OAIZIB-CM", repo_type='dataset', local_dir="/your/local/folder")
```
```python
# python
from huggingface_hub import snapshot_download
snapshot_download(repo_id="YongchengYAO/OAIZIB-CM", repo_type='dataset', revision="load_dataset-support", local_dir="/your/local/folder")
```

### 2. Load `Dataset` or `IterableDataset` from the `load_dataset-support` Branch ‼️
```python
>>> from datasets import load_dataset

# Load Dataset
>>> dataset_test = load_dataset("YongchengYAO/OAIZIB-CM", revision="load_dataset-support", trust_remote_code=True, split="test")
>>> type(dataset_test)
<class 'datasets.arrow_dataset.Dataset'>

# Convert Dataset to IterableDataset: use .to_iterable_dataset()
>>> iterdataset_test = dataset_test.to_iterable_dataset()
>>> type(iterdataset_test)
<class 'datasets.iterable_dataset.IterableDataset'>

# Load IteravleDataset: add streaming=True
>>> iterdataset_train = load_dataset("YongchengYAO/OAIZIB-CM", revision="load_dataset-support", trust_remote_code=True, streaming=True, split="train")
>>> type(iterdataset_train)
<class 'datasets.iterable_dataset.IterableDataset'>
```

- 🔥 [Differences between Dataset and IterableDataset](https://huggingface.co/docs/datasets/about_mapstyle_vs_iterable#downloading-and-streaming)


## Segmentation Labels
```python
labels_map = {
    "1": "Femur",
    "2": "Femoral Cartilage",
    "3": "Tibia",
    "4": "Medial Tibial Cartilage",
    "5": "Lateral Tibial Cartilage",
}
```

## Citations
The dataset originates from these projects:
- CartiMorph: https://github.com/YongchengYAO/CartiMorph
- CartiMorph Toolbox:
  - https://github.com/YongchengYAO/CartiMorph-Toolbox
  - https://github.com/YongchengYAO/CMT-AMAI24paper

```
@article{YAO2024103035,
title = {CartiMorph: A framework for automated knee articular cartilage morphometrics},
journal = {Medical Image Analysis},
author = {Yongcheng Yao and Junru Zhong and Liping Zhang and Sheheryar Khan and Weitian Chen},
volume = {91},
pages = {103035},
year = {2024},
issn = {1361-8415},
doi = {https://doi.org/10.1016/j.media.2023.103035}
}
```
```
@InProceedings{10.1007/978-3-031-82007-6_16,
author="Yao, Yongcheng
and Chen, Weitian",
editor="Wu, Shandong
and Shabestari, Behrouz
and Xing, Lei",
title="Quantifying Knee Cartilage Shape and Lesion: From Image to Metrics",
booktitle="Applications of Medical Artificial Intelligence",
year="2025",
publisher="Springer Nature Switzerland",
address="Cham",
pages="162--172"
}
```


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