Skip to content

gugababa/Transformers_Unsupervised_Anomaly_Segmentation

 
 

Repository files navigation

Transformer-based Models for Unsupervised Anomaly Segmentation in Brain MR Images

Official implementation of Transformer-based Models for Unsupervised Anomaly Segmentation in Brain MR Images.

Paper accepted in the International MICCAI Brainlesion 2022 Workshop

@article{ghorbel2022transformer,
  title={Transformer based Models for Unsupervised Anomaly Segmentation in Brain MR Images},
  author={Ghorbel, Ahmed and Aldahdooh, Ahmed and Albarqouni, Shadi and Hamidouche, Wassim},
  journal={arXiv preprint arXiv:2207.02059},
  year={2022}
}

Tags

MICCAI BrainLes 2022 Workshop, Transformer, Autoencoder, TensorFlow, Keras, Anomaly Segmentation, Unsupervised, Neuroimaging, Deeplearning

Requirements

  • Python >= 3.6

All packages used in this repository are listed in requirements.txt. To install those, run:

pip3 install -r requirements.txt

Folder Structure

Transformers_Unsupervised_Anomaly_Segmentation/
│
├── Transformers_Unsupervised_Anomaly_Segmentation.ipynb - Jupyter notebook to work on Google Colab
│
├── data/
│   └── data.txt  - datasets descriptions and download link
│
├── models/ - Models defining, training and evaluating
│   ├── Autoencoders/
│       ├── DCAE.py
│       └── ...
│   ├── Latent Variable models/
│       ├── VAE.py
│       └── ...
│   └── Transformer based models/
│       ├── B_TAE.py
│       └── ...
│
├── saved/  - saving folder
│
└── scripts/ - small utility scripts
    ├── utils.py
    └── ...    

Usage

CLI Usage

Every model can be trained and tested individually using the scripts which are provided in the models/* folders.

Google Colab Usage

Training can be started by importing Transformers_Unsupervised_Anomaly_Segmentation.ipynb in Google Colab. This github repository is linked and can directly loaded into the notebook. However, the datasets have to be stored so that Google Colab can access them. Either uploading by a zip-file or uploading it to Google Drive and mounting the drive.

Disclaimer

Please do not hesitate to open an issue to inform of any problem you may find within this repository.

Reference

This project is inspired by the comparative study paper on Autoencoders for Unsupervised Anomaly Segmentation in Brain MR Images: A Comparative Study.

@article{baur2021autoencoders,
  title={Autoencoders for unsupervised anomaly segmentation in brain mr images: A comparative study},
  author={Baur, Christoph and Denner, Stefan and Wiestler, Benedikt and Navab, Nassir and Albarqouni, Shadi},
  journal={Medical Image Analysis},
  pages={101952},
  year={2021},
  publisher={Elsevier}
}

License

This project is licensed under the GNU General Public License v3.0. See LICENSE for more details

About

Transformer-based Models for Unsupervised Anomaly Segmentation in Brain MR Images

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Languages

  • Jupyter Notebook 56.9%
  • Python 43.1%