Learned Primal-Dual Reconstruction

This repository contains the code for the article "Learned Primal-Dual Reconstruction".

Contents

The code contains the following

• Training using ellipse phantoms
• Evaluation on ellipse phantoms
• Training using anthropomorphic data from Mayo Clinic.
• Evaluation on example slice
• Reference reconstructions of the above using ODL.

Pre-trained networks

The pre-trained networks are currently under finalization and will be released soon.

Dependencies

The code is currently based on the latest version of ODL. It can be most easily installed by running

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

The code also requires the utility library adler which can be installed via

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

Contact

Jonas Adler, PhD student
KTH, Royal Institute of Technology
Elekta Instrument AB
jonasadl@kth.se

Ozan Öktem, Associate Professor
KTH, Royal Institute of Technology
ozan@kth.se

Funding

Development is financially supported by the Swedish Foundation for Strategic Research as part of the project "Low complexity image reconstruction in medical imaging" and "3D reconstruction with simulated forward models".

Development has also been financed by Elekta.