TY  - JOUR
A1  - Yang, Xiaogang
A1  - Kahnt, Maik
A1  - Brückner, Dennis
A1  - Schropp, Andreas
A1  - Fam, Yakub
A1  - Becher, Johannes
A1  - Grunwaldt, Jan-Dierk
A1  - Sheppard, Thomas L.
A1  - Schroer, Christian G.
T1  - Tomographic reconstruction with a generative adversarial network
Y1  - 2020
VL  - 27
IS  - 2
SP  - 486
EP  - 493
JF  - Journal of Synchrotron Radiation
DO  - 10.1107/S1600577520000831
DO  - 10.23689/fidgeo-5047
N2  - 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.
UR  - http://resolver.sub.uni-goettingen.de/purl?gldocs-11858/9393
ER  -