DDR: Deep-Discrete spherical Registration
This repository contains the code for performing spherical (cortical) surface registration using deep learning. This is the official PyTorch implementation of the paper A Deep-Discrete Learning Framework for Spherical Surface Registration, accepted at the MICCAI 2022 conference.
The data used for the image registration task comes from the HCP dataset.
To train the affine registration network, run:
DDR_affine_train.py
To train the first stage of the non-linear registration network, run:
DDR_coarse_train.py
Note that you have to manually set the directories where you want to save your corresponding models and results at the beginning of these codes.
More Coming soon
Coming soon
Please cite this work if you found it useful:
A Deep-Discrete Learning Framework for Spherical Surface Registration
@article{suliman2022deep,
title={A Deep-Discrete Learning Framework for Spherical Surface Registration},
author={Suliman, Mohamed A and Williams, Logan ZJ and Fawaz, Abdulah and Robinson, Emma C},
journal={arXiv preprint arXiv:2203.12999},
year={2022}
}