| Zugriffsnummer | 50389 |
| Dokumenttyp | Konferenzartikel |
| Peer Review | mit Peer Review |
| Sprache | Englisch |
| Titel | Convolutional analysis operator learning by end-to-end training of iterative neural networks |
| Autor(in); Institution |
Wald, C.; Department of Radiology Charité, Universitätsmedizin Berlin, Berlin, GERMANY
Haltmeier, M.; Department of Mathematics, University of Innsbruck, Innsbruck, AUSTRIA
|
| Quelle/Jahr | 2022 IEEE 19th International Symposium on Biomedical Imaging (ISBI):(2022), 5 S. |
| ISSN | 1945-7928 (print) ; 1945-8452 (online) |
| ISBN | 978-1-6654-2924-5 (print) ; 978-1-6654-2923-8 (online) |
| DOI | |
| URL | |
| Konferenzangaben | 19th International Symposium on Biomedical Imaging (ISBI), Kolkata, 28-31 March 2022, India |
| Freie Schlagworte | iterative neural networks ; sparsity ; analysis operator ; compressed sensing ; cardiac cine MRI |
| Zusammenfassung | The concept of sparsity has been extensively applied for regularization in image reconstruction. Typically, sparsifying transforms are either pre-trained on ground-truth images or adaptively trained during the reconstruction. Thereby, learning algorithms are designed to minimize some target function which encodes the desired properties of the transform. However, this procedure ignores the subsequently employed reconstruction algorithm as well as the physical model which is responsible for the image formation process. Iterative neural networks - which contain the physical model - can overcome these issues. In this work, we demonstrate how convolutional sparsifying filters can be efficiently learned by end-to-end training of iterative neural networks. We evaluated our approach on a non-Cartesian 2D cardiac cine MRI example and show that the obtained filters are better suitable for the corresponding reconstruction algorithm than the ones obtained by decoupled pre-training. |
| Themenbereich der Metrologie | Metrologie in der Medizin |
Zitierung
Kofler, A., Wald, C., Schäffter, T., Haltmeier, M., & Kolbitsch, C. (2022). Convolutional analysis operator learning by end-to-end training of iterative neural networks. 19th International Symposium on Biomedical Imaging (ISBI), Kolkata, 28-31 March 2022, India. https://doi.org/10.1109/ISBI52829.2022.9761621