Computed Tomography Reconstruction Using Generative Energy-Based Priors

Martin Zach, Erich Kobler, Thomas Pock

Publikation: Beitrag in Buch/Bericht/KonferenzbandBeitrag in einem KonferenzbandBegutachtung


In the past decades, Computed Tomography (CT) has established itself as one of the most important imaging techniques in medicine. Today, the applicability of CT is only limited by the deposited radiation dose, reduction of which manifests in noisy or incomplete measurements. Thus, the need for robust reconstruction algorithms arises. In this work, we learn a parametric regularizer with a global receptive field by maximizing it’s likelihood on reference CT data. Due to this
unsupervised learning strategy, our trained regularizer truly represents higher-level domain statistics, which we empirically demonstrate by synthesizing CT images. Moreover, this regularizer can easily be applied to different CT reconstruction problems by embedding it in a variational framework, which
increases flexibility and interpretability compared to feedforward learning-based approaches. In addition, the accompanying probabilistic perspective enables experts to explore the full posterior distribution and may quantify uncertainty
of the reconstruction approach. We apply the regularizer to limited-angle and few-view CT reconstruction problems, where it outperforms traditional reconstruction algorithms by a large margin.
TitelProceedings of the OAGM Workshop 2021
UntertitelComputer Vision and Pattern Analysis Across Domains
Redakteure/-innenMarkus Seidl, Matthias Zeppelzauer, Peter M. Roth
Herausgeber (Verlag)Verlag der Technischen Universität Graz
ISBN (elektronisch)978-3-85125-869-1
PublikationsstatusVeröffentlicht - Dez. 2021
VeranstaltungOAGM Workshop 2021: Computer Vision and Pattern Analysis Across Domains - University of Applied Sciences St. Pölten, abgesagt, Österreich
Dauer: 24 Nov. 202125 Nov. 2021


KonferenzOAGM Workshop 2021


Untersuchen Sie die Forschungsthemen von „Computed Tomography Reconstruction Using Generative Energy-Based Priors“. Zusammen bilden sie einen einzigartigen Fingerprint.

Dieses zitieren