KIT | KIT-Bibliothek | Impressum | Datenschutz

Tomographic reconstruction with a generative adversarial network

Yang, Xiaogang; Kahnt, Maik; Brückner, Dennis; Schropp, Andrea; Fam, Yakub 1; Becher, Johannes 1; Grunwaldt, Jan-Dierk ORCID iD icon 1,2; Sheppard, Thomas L. 1,2; Schroer, Christian G.
1 Institut für Technische Chemie und Polymerchemie (ITCP), Karlsruher Institut für Technologie (KIT)
2 Institut für Katalyseforschung und -technologie (IKFT), Karlsruher Institut für Technologie (KIT)


This paper presents a deep learning algorithm for tomographic reconstruction (GANrec). The algorithm uses a generative adversarial network (GAN) to solve the inverse of the Radon transform directly. It works for independent sinograms without additional training steps. The GAN has been developed to fit the input sinogram with the model sinogram generated from the predicted reconstruction. Good quality reconstructions can be obtained during the minimization of the fitting errors. The reconstruction is a self-training procedure based on the physics model, instead of on training data. The algorithm showed significant improvements in the reconstruction accuracy, especially for missing-wedge tomography acquired at less than 180° rotational range. It was also validated by reconstructing a missing-wedge X-ray ptychographic tomography (PXCT) data set of a macroporous zeolite particle, for which only 51 projections over 70° could be collected. The GANrec recovered the 3D pore structure with reasonable quality for further analysis. This reconstruction concept can work universally for most of the ill-posed inverse problems if the forward model is well defined, such as phase retrieval of in-line phase-contrast imaging.

Verlagsausgabe §
DOI: 10.5445/IR/1000115487
Veröffentlicht am 03.04.2020
DOI: 10.1107/S1600577520000831
Zitationen: 17
Web of Science
Zitationen: 16
Zitationen: 26
Cover der Publikation
Zugehörige Institution(en) am KIT Institut für Katalyseforschung und -technologie (IKFT)
Institut für Technische Chemie und Polymerchemie (ITCP)
Publikationstyp Zeitschriftenaufsatz
Publikationsmonat/-jahr 03.2020
Sprache Englisch
Identifikator ISSN: 1600-5775
KITopen-ID: 1000115487
HGF-Programm 37.03.01 (POF III, LK 01) Catalysts and Mechanisms
Erschienen in Journal of synchrotron radiation
Verlag International Union of Crystallography
Band 27
Heft 2
Seiten 486-493
Vorab online veröffentlicht am 18.02.2020
Schlagwörter missing-wedge tomography; reconstruction algorithms; generative adversarial network (GAN); ptychography
Nachgewiesen in Dimensions
Web of Science
KIT – Die Forschungsuniversität in der Helmholtz-Gemeinschaft
KITopen Landing Page