In this talk, I will present our recent activities in learning better models for inverse problems in imaging. We consider classical variational models used for inverse problems but generalized these models by introducing a large number of free model parameters. We learn the free model parameters by minimizing a loss function comparing the reconstructed images obtained from the variational models with ground truth solutions from a training data base. I will also show very recent results on learning "deeper" regularizers that are already able to capture semantic information of images. We show applications to different inverse problems in imaging where we put a particular focus on image reconstruction from undersampled MRI data.

This presentation is part of Minisymposium “MS55 - Advances of regularization techniques in iterative reconstruction (2 parts)”
organized by: Zichao (Wendy) Di (Argonne National Lab) , Marc Aurèle Gilles (Cornell University) .

Authors:

Thomas Pock (Graz University of Technology)

Kerstin Hammernik (Graz University of Technology)

Erich Kobler (Graz University of Technology)

Teresa Klatzer (Graz University of Technology)

Florian Knoll (New York University)

Keywords:

deep learning, image enhancement, image reconstruction, image segmentation, integral equations for image analysis, machine learning