Hi, I'm not sure but one thing is that after reconstruction, you should mask out everything which is out of the field of view. This can be done with the FieldOfViewImageFilter <http://www.openrtk.org/Doxygen/classrtk_1_1FieldOfViewImageFilter.html> to which you should pass the reconstructed image as first input, the ProjectionsStack and the Geometry. If this is not the issue, I would check that you have reconstructed the whole field of view, that your projections do not have any truncation issue. Simon
On Wed, Aug 11, 2021 at 12:55 PM Thibault Pelletier via Rtk-users < rtk-users@public.kitware.com> wrote: > Hi everyone, > > I'm currently implementing the NMAR algorithm using RTK and I have a > problem regarding the forward projection step. > > To make sure my building blocks are correct, I decided to first read my > projection stack using the rtk::ProjectionsReader class, do a FDK Cone Beam > reconstruction and forward project this reconstructed image using the > rtk::JosephForwardProjectionImageFilter class. > > I then compare the images that I have after reading my projection stack > with the images that I have after my JosephForwardProjectionImageFilter > pass. > > My problem is that the values I obtain are quite different (sinogram of > the images attached). > The forward projected images are overall 1.7 times brighter than the > reader images. > > Does anyone have an insight on what I could be doing wrong? > > Thanks in advance, > Thibault > _______________________________________________ > Rtk-users mailing list > Rtk-users@public.kitware.com > https://public.kitware.com/mailman/listinfo/rtk-users >
_______________________________________________ Rtk-users mailing list Rtk-users@public.kitware.com https://public.kitware.com/mailman/listinfo/rtk-users
