Multi Domain U-Net Model for MRI Reconstruction

Welcome to the repository of my Bachelor's thesis at the Department of Electrical and Computer Engineering at TUM. My thesis was about "Data Standardization, Multi-Domain Learning and GRAPPA preprocessing for Improved MRI". To read a more detailed discussion see my thesis and final presentation.

This directory contains implementations of U-Net and multi-domain-U-Net for MRI reconstruction in PyTorch. It also contains implementations of the different methods I used in my work, so that the reported results are easily reproducible.

To visit the main page of the fastMRI challenge, please go here.

Dependencies and Installation

I have tested this code using:

• Ubuntu 16.04.6
• Python 3.6.9
• CUDA 10.2

You can find the full list of Python packages needed to run the code in the requirements.txt file. To install them, run

pip install -r requirements.txt

Repository Structure

This repository is structured as follows:

• The directory tutorials contains three Jupyter notebooks that illustrate how to deal with the data and how to use the implemented methods. This is the best place to start.

• The other directories data, common and models contain the actual implementations of the model training pipeline. See the respective folders for a more detailed explanation.