A convolutional neural network for fast upsampling of undersampled tomograms in X-ray CT time-series using a representative highly sampled tomogram

Bellos, D.; Basham, M.; Pridmore, T.; French, A.P.

doi:10.1107/S1600577519003448

Journal of Synchrotron Radiation Journal of
Synchrotron Radiation

IUCr IT WDC

home archive editors for authors for readers submit subscribe open access

view article

Figure 23
The selected area marked by the red window in Fig. 22(A)

enlarged (area [1501:1800, 701:1000] of [1:2560,1:2560]). (A, B, C) Reconstruction of the 1101th sinogram (area [550:2000, 550:2000] of [1:2560,1:2560], pixel-value [−4.501 × 10⁻⁴:1.097 × 10⁻⁴]). (D, E, F) Reconstruction of the 1601th sinogram (area [550:2000, 550:2000] of [1:2560,1:2560], pixel-value [−1/2560:1/2560]). (A, D) From the original dataset with 91 projections. (B, E) Using UDNN-128 for upscaling resulting in 181 projections. (C, F) Using UDNN-128 for upscaling resulting in 361 projections. Similar to the earlier Figs. 12

, 18

and 20

, the proposed method (panels B, C, E and F) produces more clear reconstructions with less noise.

JOURNAL OF
SYNCHROTRON
RADIATION

ISSN: 1600-5775

Volume 26| Part 3| May 2019| Pages 839-853

https://doi.org/10.1107/S1600577519003448

Open

access