Can it be used in the field of industrial CT？

[zeminLi66]

In industrial CT 3D reconstruction, we usually keep the object stationary and translate the X-ray source to obtain projection images. Can the 3D reconstruction algorithm in this library work？

[AnderBiguri]

@zeminLi66 yes!
You will need to play a bit with the geometry. It's not easy to set up this specific geoemtry, but it's indeed possible. The digital breast tomosynthesis demo may be what you want, try that.

[AnderBiguri]

@zeminLi66 Unofunately the data has not been uploaded. In any case, you are doing a fairly complicated geometry, and there could be hundreds of things that have gone wrong. Sadly I don't have the time to debug all the code of TIGRE users, as I do this on the side, unpaid. You will need to study the code more and find the bug yourself, sorry.

[zeminLi66]

Rec.2.mp4

[zeminLi66]

[zeminLi66]

I am very grateful that you could reply to me amidst your busy schedule. I apologize for the file upload failure just now. During this period, I have been researching TIGRE and 3D reconstruction knowledge. Here are the results and code of my refactoring. I have been debugging for a long time, but there are some issues with the results. Could you please take some time to help me find out where I set the parameters incorrectly based on your experience.

[zeminLi66]

Sorry to trouble you again，After refactoring and debugging the code, the central region of the reconstructed image shows improved visibility but remains significantly blurred. Could you please analyze potential causes (e.g., algorithm parameter configuration or hardware acceleration settings) from an implementation perspective? If the issue stems from discrepancies in the C++ porting logic, I would appreciate guidance on key areas to review as I further verify the code integrity.

[zeminLi66]

This is a screen recording

Rec.0001.mp4