# image-registration-cnn

**Repository Path**: songxiao1234/image-registration-cnn

## Basic Information

- **Project Name**: image-registration-cnn
- **Description**: [Incomplete] A PyTorch implementation of CNN based MRI image registration based on MICCAI 2018 paper "Linear and Deformable Image Registration with 3D Convolutional Neural Networks"
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 1
- **Forks**: 0
- **Created**: 2021-06-23
- **Last Updated**: 2021-12-01

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Linear and Deformable Image Registration with 3D CNN
A PyTorch implementation of CNN based MRI image registration based on MICCAI 2018 paper [Linear and Deformable Image Registration with 3D Convolutional Neural Networks](https://arxiv.org/abs/1809.06226).

**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](./presentation/Presentation.pdf).**

## Network Architecture
![alt text](./presentation/network_architecture.png "Network Architecture")

## Requirements
- Python 3
- PyTorch
- Torchvision
- Numpy
- Skimage
- Tqdm

## Usage
### Train
```
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]
```