ENH: Lambda functions instead of functor classes in JosephForwardProjectionFilter#694
ENH: Lambda functions instead of functor classes in JosephForwardProjectionFilter#694MichaelColonel wants to merge 2 commits into
Conversation
|
Thanks a lot for doing this. Could you please modify the existing projector instead of creating a new one? That would facilitate the review and this is what we want eventually. |
I will do that, but what to do with JosephForwardAttenuatedProjectionImageFilter class? |
|
I've modified existing classes for JosephForwardProjectionFilter. |
MichaelColonel
changed the title
arobert01
left a comment
arobert01
left a comment
There was a problem hiding this comment.
Thank you for the great work! It looks really good to me. I made some comments, mostly related to style.
Regarding JosephAttenuatedForwardProjector, based on the tests I did, everything seems to work well, and I don’t see a better way to do it. Maybe @SimonRit will have additional comments before merging.
If there are no further comments, we can also implement these changes in the Joseph-based back projectors.
Thanks again!
| * | ||
| * \author Simon Rit |
There was a problem hiding this comment.
I would remove the author part of the comments.
| using InputPixelType = typename TInputImage::PixelType; | ||
| using OutputPixelType = typename TOutputImage::PixelType; | ||
| using CoordinateType = double; | ||
| using WeightCoordinateType = typename itk::PixelTraits<InputPixelType>::ValueType; |
There was a problem hiding this comment.
WeightCoordinateType doesn't seem to be used in this case
|
|
||
| std::cout << "\n\n****** Compare two resulted images ******" << std::endl; | ||
| CheckImageQuality<OutputImageType>(jfp->GetOutput(), mipfp->GetOutput(), 0.001, 0.000001, 1.E+19); | ||
| std::cout << "\n\nImages are OK! " << std::endl; |
There was a problem hiding this comment.
Maybe I would keep the same "Test PASSED! " sentence for consistency ?
| /** \brief Function to multiply the interpolation weights with the projected | ||
| * volume values and attenuation map. | ||
| * | ||
| * \author Antoine Robert |
There was a problem hiding this comment.
Remove author section in the comments
|
Thank you for your excellent contribution @MichaelColonel and your review @arobert01. This is very interesting work but I will include it in the next RTK major release. I am currently working on a minor release and will merge this one afterwards. Note that #629 will probably require additional work for this PR but I'll handle it. |
|
I rebased to main and added a commit to do the same for JosephBackProjectionFilter |
SimonRit
left a comment
SimonRit
left a comment
There was a problem hiding this comment.
@axel-grc Can you please:
- fix the wrapping (keep existing wrapping without functors)
- remove the threadId parameter in the Backprojection as this is not multithreaded
- squash the three commits of @MichaelColonel in one commit
?
@axel-grc Have you had a chance to start working on this yet? I'm currently working on a development based on this PR, so I'd be happy to step in and implement @SimonRit's requested changes if that helps. |
@arobert01 I did not start working on this yet. |
…ardProjectionFilter Lambdas were added to the classes: JosephForwardProjectionImageFilter, JosephForwardAttenuatedProjectionImageFilter, MaximumIntensityProjectionImageFilter. Python wrappers for classes above have been modified. Similar lambda functions can be added to back projection classes.
I implemented the requested changes and rebased to main. |
4 participants
Implementation of lambdas instead of functor classes in rtkJosephForwardProjectionFilter.
Test example of MIP filter using this new class.
Possible solution of issue #645.