Figure 3
From: Virtual multi-alignment theory of parallel-beam CT image reconstruction for elastic objects

Spread of the high-MAC area. The pixel number of the high-MAC area is reduced by spreading the MAC whenever the filtered inverse Radon transform is applied under ideal conditions such that the thickness of the object and the pixel value in the projection image are directly proportional for the Radon transform, and a well-aligned projection image set is given for the filtered inverse Radon transform. Because we expect a reduced MAC for areas after reconstruction of the reconstructed slice already having a high-MAC area, we started to observe this for an image sample without a high-MAC area at the beginning and then added the high-MAC area after reconstruction. (a) The test object has two high-MAC areas of 109 pixels for each. (b) The number of pixels maintaining a high-MAC area of 80% or more of the initial MAC in a reconstructed slice according to the projection image. (c) Reduction of the high-MAC pixels when using the Radon transform and the filtered inverse Radon transform.