Multi-ConDoS.zip

# Multi-ConDoS: Multimodality Contrastive Domain Sharing for Self-Supervised Medical Image Segmentation This repository is the official implementation of IEEE paper [Multi-ConDoS: Multimodality Contrastive Domain Sharing for Self-Supervised Medical Image Segmentation](https://ieeexplore.ieee.org/document/10167829). ## Requirements To install requirements : ```setup conda env create -r env.yml ``` To download datasets : - [MICCAI 2020: HECKTOR](https://www.aicrowd.com/challenges/hecktor) - [BraTS 2018](https://www.med.upenn.edu/sbia/brats2018.html) ## For pre-training To train the DSGAN model in the paper, run this command ```shell python pytorch-DSGAN/train.py ``` ## For fine-tuning To train the downstream network, run this command - for hecktor : ```shell python torch_unet_z/train_Hecktor.py ``` - for brats : ```shell python torch_unet_z/train_BraTS.py ``` ## Evaluation To evaluate my model - for hecktor : ```shell python torch_unet_z/predict.py ``` - for brats2018 : ```shell python torch_unet_z/predict_3c.py ``` ## Citation If Multi-ConDoS is used in your research, please cite the following article in your paper: ``` @article{Jiaojiao-TMI2023, author={Zhang, Jiaojiao and Zhang, Shuo and Shen, Xiaoqian and Lukasiewicz, Thomas and Xu, Zhenghua}, journal={IEEE Transactions on Medical Imaging}, title={Multi-ConDoS: Multimodal Contrastive Domain Sharing Generative Adversarial Networks for Self-Supervised Medical Image Segmentation}, year={2023}, volume={Early Access}, pages={1-20}, doi={10.1109/TMI.2023.3290356}} ``` Please also consider citing the following article in your paper: ``` @article{ZHANG-MIA2022, title = {Multi-modal contrastive mutual learning and pseudo-label re-learning for semi-supervised medical image segmentation}, journal = {Medical Image Analysis}, pages = {102656}, year = {2023}, volume = {83}, author = {Shuo Zhang and Jiaojiao Zhang and Biao Tian and Thomas Lukasiewicz and Zhenghua Xu} } ``` ## Acknowledgements - The code is developed based on [CycleGAN](https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix) and [SimCLR](https://github.com/Spijkervet/SimCLR). If you have some issues, please contact 2195468073@qq.com <!-- >���� Pick a licence and describe how to contribute to your code repository. -->