MSc Disseration Inverse Problems in Medical Imaging

This project contains a variety of functions that use ADMM for CT reconstructions of images. The scripts in the learned section are inspired by the Learned Primal-Dual Reconstruction algorithm (https://arxiv.org/abs/1707.06474).

ODL (https://github.com/odlgroup/odl)

$ pip install https://github.com/odlgroup/odl/archive/master.zip

Utility library adler (https://github.com/adler-j/learned_primal_dual)

$ pip install https://github.com/adler-j/adler/archive/master.zip

Tensorflow (It is best to create a new virtual environment and use)

$ pip install --upgrade tensorflow

• matplotlib

• numpy

• scipy

The code in this project has been designed to run as scripts

$ python3 (script_name_here).py

$ py (script_name_here).py

All methods in this project that use Machine Learning have a script to learn the parameters and a corresponding script with prefix ##evaluate_## to produce results using the parameters that have been learned. These will generate results into a folder of the same name as the function.

All non-learned methods return outputs a Results folder which is created on first running.