Skip to main content
. Author manuscript; available in PMC: 2020 Feb 9.
Published in final edited form as: Nat Mach Intell. 2019 Aug 9;1(8):373–380. doi: 10.1038/s42256-019-0077-5

Figure 5. Classical analytical rebinning vs. derived neural networks.

Figure 5

The trained rebinning algorithm can directly applied to real MR projection data. Parallel-beam MR projection data is rebinned to a stacked fan-beam geometry with the analytical (a) and the learned algorithm (b). Note that the result of the learned method is much sharper as it avoids ray-by-ray interpolation.