zoonono / D2MC-Net-main

【MICCAI 2023 Early Accept】Dual Domain Motion Artifacts Correction for MR Imaging Under Guidance of K-space Uncertainty

D2MC-Net

---

Pytorch codes for model in "Dual Domain Motion Artifacts Correction for MR Imaging Under Guidance of K-space Uncertainty" (MICCAI 2023)

If you use these codes, please cite our paper:

[1] Jiazhen Wang, Yizhe Yang, Yan Yang, Jian Sun. Dual Domain Motion Artifacts Correction for MR Imaging Under Guidance of K-space Uncertainty (MICCAI 2023).

http://gr.xjtu.edu.cn/web/jiansun/publications

All rights are reserved by the authors.

Jiazhen Wang and Yizhe Yang -2023/07/04. For more detail or traning data, feel free to contact: [e-mail]

---

Installation

This installation guide shows you how to set up the environment for running our code using conda.

First clone the D2MC-Net repository

git clone https://github.com/Jiazhen-Wang/D2MC-Net.git
cd D2MC-Net

Then start a virtual environment with new environment variables

conda create --name D2MC-Net python=3.8
conda activate D2MC-Net 

Install PyTorch

pip install torch torchvision

Install all requirements

pip install -r requirements.txt

Usage:

The net of training was implemented by end-to-end in D2MC-Net of pytorch version. For retraining you should simulate motion data and change the data path in train.py and saved in the logs_D2MCNet directory.

Training:

python train.py

---