A PyTorch implementation of CNN based MRI image registration based on MICCAI 2018 paper Linear and Deformable Image Registration with 3D Convolutional Neural Networks.
I am still waiting for Data Access Rights so the code has not been tested. I have had some bug reports so I suggest you NOT to use this code unless you are willing to use it as a headstart to adapt it to your needs.
For a quick introduction to image registration and summary of the paper check out this presentation.
- Python 3
- PyTorch
- Torchvision
- Numpy
- Skimage
- Tqdm
python ./bin/train.py
optional arguments:
--linear whether to include the affine transform branch or not [default value is True]
--epochs number of epochs to train [default value is 300]