by Yicheng Wu*, Zongyuan Ge, Donghao Zhang, Minfeng Xu, Lei Zhang, Yong Xia, and Jianfei Cai.
by Yicheng Wu, Minfeng Xu, Zongyuan Ge, Jianfei Cai*, and Lei Zhang.
<01.07.2022> Our paper entitled "Mutual Consistency Learning for Semi-supervised Medical Image Segmentation" has been accepted by Medical Image Analysis;
<18.04.2022> We provided our pre-trained models on the LA, Pancreas-CT and ACDC datasets, see './MC-Net/pretrained_pth/';
<16.04.2022> We released the codes;
This repository is for our paper: 'Mutual Consistency Learning for Semi-supervised Medical Image Segmentation'. Note that, the MC-Net+ model is named as mcnet3d_v2 in our repository and we also provide the mcnet2d_v1 and mcnet3d_v1 versions, which are similar to the MC-Net model in MICCAI 2021: 'Semi-supervised Left Atrium Segmentation with Mutual Consistency Training'.
This repository is based on PyTorch 1.8.0, CUDA 11.2 and Python 3.8.10; All experiments in our paper were conducted on a single NVIDIA Tesla V100 GPU.
- Clone the repo.;
git clone https://github.com/ycwu1997/MC-Net.git
-
Put the data in './MC-Net/data';
-
Train the model;
cd MC-Net
# e.g., for 20% labels on LA
python ./code/train_mcnet_3d.py --dataset_name LA --model mcnet3d_v2 --labelnum 16 --gpu 0 --temperature 0.1
- Test the model;
cd MC-Net
# e.g., for 20% labels on LA
python ./code/test_3d.py --dataset_name LA --model mcnet3d_v2 --exp MCNet --labelnum 16 --gpu 0
If our MC-Net+ model is useful for your research, please consider citing:
@inproceedings{wu2021semi,
title={Semi-supervised left atrium segmentation with mutual consistency training},
author={Wu, Yicheng and Xu, Minfeng and Ge, Zongyuan and Cai, Jianfei and Zhang, Lei},
booktitle={International Conference on Medical Image Computing and Computer-Assisted Intervention},
pages={297--306},
volume={12902},
year={2021},
doi={10.1007/978-3-030-87196-3\_28},
organization={Springer, Cham}
}
@article{wu2022mutual,
title={Mutual Consistency Learning for Semi-supervised Medical Image Segmentation},
author={Wu, Yicheng and Ge, Zongyuan and Zhang, Donghao and Xu, Minfeng and Zhang, Lei and Xia, Yong and Cai, Jianfei},
journal={Medical Image Analysis},
volume={81},
pages={102530},
year={2022},
publisher={Elsevier}
}
Our code is adapted from UAMT, SASSNet, DTC, URPC and SSL4MIS. Thanks for these authors for their valuable works and hope our model can promote the relevant research as well.
If any questions, feel free to contact me at '[email protected]'