Author: Zaccharie Ramzi, Francois Lanusse
Year: 2020
This tutorial was intended for the Covid-19 crisis learning-slots in March-April 2020. It presents the use of TensorFlow in 2 different settings, for the same problem of MRI reconstruction:
- as an optimisation library to perform classical iterative reconstruction
- as a deep learning library to train a CNN to reconstruct fourier undersampled images
In order to run the tutorial you will need Python 3.5+.
All the Python requirements are listed in requirements.txt except cython which is to be installed beforehand.
You can then install all the requirements using pip install -r requirements.txt.