Journal
IEEE TRANSACTIONS ON MEDICAL IMAGING
Volume 37, Issue 6, Pages 1322-1332Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TMI.2018.2799231
Keywords
Inverse problems; tomography; deep learning; primal-dual; optimization
Categories
Funding
- Swedish Foundation of Strategic Research [AM13-0049, ID14-0055]
- Elekta
- National Institute of Biomedical Imaging and Bioengineering [EB017095, EB017185]
- Swedish Foundation for Strategic Research (SSF) [ID14-0055, AM13-0049] Funding Source: Swedish Foundation for Strategic Research (SSF)
Ask authors/readers for more resources
We propose the Learned Primal-Dual algorithm for tomographic reconstruction. The algorithm accounts for a (possibly non-linear) forward operator in a deep neural network by unrolling a proximal primal-dual optimization method, but where the proximal operators have been replaced with convolutional neural networks. The algorithm is trained end-to-end, working directly from raw measured data and it does not depend on any initial reconstruction such as filtered back-projection (FBP). We compare performance of the proposed method on low dose computed tomography reconstruction against FBP, total variation (TV), and deep learning based post-processing of FBP. For the Shepp-Logan phantom we obtain >6 dB peak signal to noise ratio improvement against all compared methods. For human phantoms the corresponding improvement is 6.6 dB over TV and 2.2 dB over learned post-processing along with a substantial improvement in the structural similarity index. Finally, our algorithm involves only ten forward-back-projection computations, making the method feasible for time critical clinical applications.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available