Andreas Kofler (PTB) published together with Christian Wald (Charité–Universitätsmedizin Berlin), Markus Haltmeier (University of Innsbruck), Tobias Schäffter (PTB) and Christoph Kolbitsch (PTB) an article about a method for image reconstruction based on convolutional dictionary learning.
The method can be seen as a physics-informed and supervised method fo learnind a convolutional dictionary. By unrolling an iterative scheme which is derived from a functional which involves a convolutional dictionary as regularization method, the filters as well all regularization parameters can then be learned in a supervised manner by end-to-end training of the network. The method was evaluated on a cardiac cine MR reconstruction problem and compared to other machine learning-based methods.
A pre-print of the article is available at https://arxiv.org/abs/2206.04447. An implementation of the method as well as its analysis counterpart (see https://arxiv.org/abs/2203.02166) can be found at https://github.com/koflera/ConvSparsityNNs.
Contact:
Christoph Kolbitsch, E-Mail: Christoph.Kolbitsch(at)ptb.de