Relevant to my project, I would like to have your advise how to evaluate effectively the scatter boundoary effects on the specific transparent sampling (see the attached file) on (MSCT)/CBCT* imaging, resulted as the filleting the edges and contours.

There are some relevant informations for this question, as following (for more information, please refer to my project):

1. CBCTest cube DNL3a15d4 is made from PMMA (drawing as CBCTTest cube DNL3a15d4.png), imaged by CBCTs, in the standard/ high-definition imaging mode (80 kVp and 6 mA), with the specific F.O.V. (in cm) in the range of  6x6 to 8x8, in the static environment of 60% humidity and dust-free.

2. Imaging procedure is done with Test cube alignment paired-up with 3-axis laser beam tracking system (FSD 0.5mm), built-in as default on CBCT

3. After imaging, I did the segmentation of those set DICOM data, within the set optimum threshhold for PMMA (in my opinion), from -430HU to +470HU, +/- 8HU (on the SIMPLANT Pro, a daughter-software of MIMICS, Materialise NV).

4. Export the segmented model into stl file, do the superimposing with the CAD design file on ccloud to evaluate the geometry difference.