search

RTKConsortium / RTK

Reconstruction Toolkit
Apache License 2.0
245 stars 145 forks source link

Why it is different after using CudaIterativeFDKConeBeamReconstructionFilter and CudaFDKConeBeamReconstructionFilter #638

Closed jxhno1 closed 1 hour ago

jxhno1 commented 1 day ago

As I want to write a demo using CudaIterativeFDKConeBeamReconstructionFilter, and I wonder why the Hu is 200 as it is 1000 when using CudaFDKConeBeamReconstructionFilter.

SimonRit commented 1 day ago

By Hu you mean Hounsfiel Unit?

jxhno1 commented 1 day ago

Yes, I used the same data but the CudaIterativeFDKConeBeamReconstructionFilter case give me the Hounsfield Unit not close to the CudaFDKConeBeamReconstructionFilter one.

SimonRit commented 1 hour ago

This is not what I observe on a simple example:

rtkprojectshepploganphantom -g g -o proj.mha --spacing 2
rtksimulatedgeometry -o g -n 36
rtkprojectshepploganphantom -g g -o proj.mha --spacing 2
rtkfdk -g g -p . -r ^proj.mha$ -o fdk.mha 
rtkiterativefdk -g g -p . -r ^proj.mha$ -o ifdk.mha

That must be something wrong with your data, not an RTK issue. Please provide a concrete demonstration or discuss your data on the mailing list.