CycleMorph

This is an official repository of CycleMorph.

Paper

• CycleMorph: Cycle consistent unsupervised deformable image registration (Medical Image Analysis, Boah Kim et al.) [paper]
• Unsupervised Deformable Image Registration Using Cycle-Consistent CNN (MICCAI 2019, Boah Kim et al.)[paper]

Implementation

A PyTorch implementation of deep-learning-based registration. We implemented this code based on voxelMorph and original cycleGAN code. (*Thanks for voxelMorph.) (*Thanks for Jun-Yan Zhu and Taesung Park, and Tongzhou Wang.)

• Requirements
  • OS : Ubuntu
  • Python 3.6
  • PyTorch 1.4.0

Data

To download the atlas brain and a sample data, visit the Data. The data should be in folder ./data.

Training

• train.py which is handled by scripts/Brain_train.sh
• You can run the code by running ./scripts/Brain_train.sh
• A code for CycleMorph is in models/cycleMorph_model.py.

Testing

• test.py which is handled by scripts/Brain_test.sh
• You can run the code by running ./scripts/Brain_test.sh

Citations

@article{kim2021cyclemorph,
  title={CycleMorph: cycle consistent unsupervised deformable image registration},
  author={Kim, Boah and Kim, Dong Hwan and Park, Seong Ho and Kim, Jieun and Lee, June-Goo and Ye, Jong Chul},
  journal={Medical Image Analysis},
  volume={71},
  pages={102036},
  year={2021},
  publisher={Elsevier}
}

@inproceedings{kim2019unsupervised,
  title={Unsupervised deformable image registration using cycle-consistent cnn},
  author={Kim, Boah and Kim, Jieun and Lee, June-Goo and Kim, Dong Hwan and Park, Seong Ho and Ye, Jong Chul},
  booktitle={International Conference on Medical Image Computing and Computer-Assisted Intervention},
  pages={166--174},
  year={2019},
  organization={Springer}
}