HCDG: A Hierarchical Consistency Framework for Domain Generalization on Medical Image Segmentation

Requirements

• python 3.6.8

conda create -n HCDG python=3.6.8 

• PyTorch 1.5.0

  conda activate HCDG 
  conda install pytorch==1.5.0 torchvision cudatoolkit=9.2 -c pytorch 
  pip install tensorboardX==2.0
  pip install opencv-python
  pip install pyyaml
  pip install MedPy
  pip install tqdm
  pip install matplotlib
  conda install -c anaconda scikit-image

Usage

1. Download the Fundus dataset into your own folder and change --data-dir correspondingly.

2. Train the model.

   python train.py -g 0 --datasetTrain 1 2 3 --datasetTest 4 --batch-size 4 --resume ./pretrained-weight/test4-40.pth.tar # You need to pretrain a vanilla model

3. Test the model.

   python test.py --model-file ./logs/test4/20210910_215812.079473/checkpoint_50.pth.tar --datasetTest 4 -g 0
   

Reference

If you find this code helpful, please cite our paper:

  @article{yang2021hcdg,
     title={Hcdg: A Hierarchical Consistency Framework for Domain Generalization on Medical Image Segmentation},
     author={Yang, Yijun and Wang, Shujun and Zhu, Lei and Yu, Lequan},
     publisher={AIIM-D-22-01172}
  }