2D similarity transform parameter map doesn't keep "CenterOfRotationPoint" parameter

[NHPatterson]

Hi all, thanks for putting together this package, I'm really enjoying the ease of installation.

It seems the similarity transform as wrapped in ITKElastix doesn't return the CenterOfRotationPoint transform parameter (although it is written to file) when doing 2D to 2D registration.

I encountered this issue using SimpleElastix as well so it may be in elastix codebase itself.

import numpy as np
import itk

zero_image = np.zeros([1024,1024])

similarity_parameter_map = parameter_object.GetDefaultParameterMap("rigid")

# set method to similarity & 2D
similarity_parameter_map["Transform"] = ["SimilarityTransform"]
similarity_parameter_map["MovingImageDimension"] = ["2"]
similarity_parameter_map["FixedImageDimension"] = ["2"]

parameter_object = itk.ParameterObject.New()
parameter_object.AddParameterMap(similarity_parameter_map)

sim_result_image, sim_params = itk.elastix_registration_method(
    zero_image.astype(np.float32),
    zero_image.astype(np.float32),
    parameter_object=parameter_object,
    log_to_console=True,
    output_directory="./transforms_similarity",
)

transform_cent_rot = np.array(
    sim_params.GetParameter(0, "CenterOfRotationPoint")
).astype(np.float32)[::-1]

assert len(transform_cent_rot) == 2

print(sim_params)

ParameterObject (0000028751D7CE00)
  RTTI typeinfo:   class elastix::ParameterObject
  Reference Count: 1
  Modified Time: 2007229
  Debug: Off
  Object Name: 
  Observers: 
    none
ParameterMap 0: 
  (CenterOfRotationPoint)
  (CompressResultImage "false")
  (DefaultPixelValue 0)
  (Direction 1 0 0 1)
  (FinalBSplineInterpolationOrder 3)
  (FixedImageDimension 2)
  (FixedInternalImagePixelType "float")
  (HowToCombineTransforms "Compose")
  (Index 0 0)
  (InitialTransformParametersFileName "NoInitialTransform")
  (MovingImageDimension 2)
  (MovingInternalImagePixelType "float")
  (NumberOfParameters 4)
  (Origin 0 0)
  (ResampleInterpolator "FinalBSplineInterpolator")
  (Resampler "DefaultResampler")
  (ResultImageFormat "nii")
  (ResultImagePixelType "float")
  (Size 1024 1024)
  (Spacing 1 1)
  (Transform "SimilarityTransform")
  (TransformParameters 1 0 0 0)
  (UseDirectionCosines "true")

file on disk

(Transform "SimilarityTransform")
(NumberOfParameters 4)
(TransformParameters 1.000000 0.000000 0.000000 0.000000)
(InitialTransformParametersFileName "NoInitialTransform")
(UseBinaryFormatForTransformationParameters "false")
(HowToCombineTransforms "Compose")

// Image specific
(FixedImageDimension 2)
(MovingImageDimension 2)
(FixedInternalImagePixelType "float")
(MovingInternalImagePixelType "float")
(Size 1024 1024)
(Index 0 0)
(Spacing 1.0000000000 1.0000000000)
(Origin 0.0000000000 0.0000000000)
(Direction 1.0000000000 0.0000000000 0.0000000000 1.0000000000)
(UseDirectionCosines "true")

// SimilarityTransform specific
(CenterOfRotationPoint 511.5000000000 511.5000000000)

// ResampleInterpolator specific
(ResampleInterpolator "FinalBSplineInterpolator")
(FinalBSplineInterpolationOrder 3)

// Resampler specific
(Resampler "DefaultResampler")
(DefaultPixelValue 0.000000)
(ResultImageFormat "nii")
(ResultImagePixelType "float")
(CompressResultImage "false")

[N-Dekker]

@NHPatterson Thanks for reporting this issue, Heath. Stefan (@stefanklein), Viktor (@ViktorvdValk) and I just discussed the issue. It appears that elastix::SimilarityTransformElastix should override CreateTransformParametersMap, in order to fix this issue. I will try to fix it soon!

[ViktorvdValk]

The same counts for AffineDTITransform, see issue 366 on elastix repo

[N-Dekker]

The same counts for AffineDTITransform, see issue 366 on elastix repo

Thanks @ViktorvdValk I just assigned issue SuperElastix/elastix#366 to myself 😄

[N-Dekker]

@stefanklein @ViktorvdValk I think this issue should be fixed by SuperElastix/elastix#379 But then, what about the following three elastix transform types?

• AdvancedAffineTransformElastix
• AffineLogTransformElastix
• EulerStackTransform

Are these three fine, or should they get a similar fix? Is it possible to check that in Python?

[stefanklein]

Hi Niels, Potentially these could suffer from the same problem. You could see it by verifying if they have a CreateTransformParameterMap function, and compare that implementation to the WriteTransformParameterFile code. Potentially, any transform that saves some transform specific settings (not just CenterOfRotationPoint) to a TransformParameter file (in the function WriteTransformParameterFile) could suffer from this problem when the CreateTransformParameterMap function is not in sync. Best, Stefan From: Niels Dekker <notifications@github.com> Sent: Tuesday, December 22, 2020 10:06 PM To: InsightSoftwareConsortium/ITKElastix <ITKElastix@noreply.github.com> Cc: S. Klein <s.klein@erasmusmc.nl>; Mention <mention@noreply.github.com> Subject: Re: [InsightSoftwareConsortium/ITKElastix] 2D similarity transform parameter map doesn't keep "CenterOfRotationPoint" parameter (#93) @stefanklein<https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fstefanklein&data=04%7C01%7Cs.klein%40erasmusmc.nl%7C4f0fcd00e2b0431124d408d8a6bd6b9c%7C526638ba6af34b0fa532a1a511f4ac80%7C0%7C0%7C637442679778961169%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=Gl3Kxwt%2B%2B8XHyG1l85qISSFssbZPQXDeFpi%2FE1uTD5o%3D&reserved=0> @ViktorvdValk<https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FViktorvdValk&data=04%7C01%7Cs.klein%40erasmusmc.nl%7C4f0fcd00e2b0431124d408d8a6bd6b9c%7C526638ba6af34b0fa532a1a511f4ac80%7C0%7C0%7C637442679778961169%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=E6YGftrekd2FpoGfIUzXmdt8eo5x6OYeD1tQyy3fz94%3D&reserved=0> I think this issue should be fixed by SuperElastix/elastix#379<https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FSuperElastix%2Felastix%2Fpull%2F379&data=04%7C01%7Cs.klein%40erasmusmc.nl%7C4f0fcd00e2b0431124d408d8a6bd6b9c%7C526638ba6af34b0fa532a1a511f4ac80%7C0%7C0%7C637442679778971163%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=RMns42UBWwnz3yhQk%2BvBopLz5R3dHI92aO2TX0FeChY%3D&reserved=0> But then, what about the following three elastix transform types? * AdvancedAffineTransformElastix * AffineLogTransformElastix * EulerStackTransform Are these three fine, or should they get a similar fix? Is it possible to check that in Python? - You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FInsightSoftwareConsortium%2FITKElastix%2Fissues%2F93%23issuecomment-749775974&data=04%7C01%7Cs.klein%40erasmusmc.nl%7C4f0fcd00e2b0431124d408d8a6bd6b9c%7C526638ba6af34b0fa532a1a511f4ac80%7C0%7C0%7C637442679778971163%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=J6API7gN0t1FWYhU8G%2FkHPfM1qXWQV2x%2Fayb7CMIP00%3D&reserved=0>, or unsubscribe<https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FAAF2LNNLS7HJBJRPEOMQVV3SWEC4NANCNFSM4U5JUIPA&data=04%7C01%7Cs.klein%40erasmusmc.nl%7C4f0fcd00e2b0431124d408d8a6bd6b9c%7C526638ba6af34b0fa532a1a511f4ac80%7C0%7C0%7C637442679778981162%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=hVhi1a3HljF3pUPT1%2Fw4%2B99QWAVYye28iY%2FvjYORV5I%3D&reserved=0>.

[ViktorvdValk]

@N-Dekker I checked all the transforms in python and the same happens for:

• AffineLogStackTransform
• AffineLogTransformElastix
• EulerStackTransform

AdvancedAffineTransformElastix is just the AffineTransform which does return the CenterOfRotationPoint

[N-Dekker]

@ViktorvdValk Do you mean that those three transforms (AffineLogStackTransform, AffineLogTransformElastix, EulerStackTransform) also need to be fixed (the same like Similarity and AffineDTI)?

[ViktorvdValk]

Yes, to be certain, you check this with the verification Stefan mentioned, but if I print the output of one of these transforms, it doesn't contain the 'CenterOfRotationPoint' parameter.

[N-Dekker]

@NHPatterson Hi Heath! I'm closing the issue now, as I think we got it fixed 😃 By pull request SuperElastix/elastix#379 "Add missing parameters to transform parameter maps", which is just merged onto the https://github.com/SuperElastix/elastix develop branch. If you have the time, please check it out! And of course, if you think the issue is not entirely fixed, please feel free to reopen the issue, or submit another one. Thanks again!