Hi all,I have been attempting to apply a transformation that corresponds to a particular deformationField.mhd file, to a 3D-CT image.
The reason I am doing this is because I have developed a motion model in MATLAB that predicts the deformation field (inversedeformationFieldapprox.mhd) . Hence, I would like to find and apply the transformation that corresponds to this approximated dvf to a reference image (reference.mhd).
Current method: The 'deformationFieldapprox.mhd' file that I have created in MATLAB takes points from reference.mhd, and outputs points that approximate the image, 00.mhd. Then I invert the approx dvf so that 'inversedeformationFieldapprox.mhd' takes points in the moving image, 00.mhd and outputs the reference image. So now I think that it is correct to say that this inversedvf corresponds to a transformation, in which my reference image, reference.mhd, is set as the moving image. So I use the following command to try and transform my reference image according to the approximated dvf:
elastix -f _____.mhd -m reference.mhd -t0 ___.txt -p ___.txt -out ___
where -t0 is using a parameter file with (Transform "DeformationFieldTransform") and with 'inversedeformationFieldapprox.mhd' as the dvf, and -p is a dummy parameter file.
Problem: I expected that it should not matter what fixed image I use or what dummy parameter file i use, since the initial parameters (TransformParameters.0.R0.txt) would only depend on the t0 file with my deformationFieldapprox.mhd information. Through testing I have noticed that the image (result.0.R0.mhd) does depend on what fixed image I use.
Why is this? Any ideas as to how I can transform my reference image with the approx dvf that I have created in MATLAB?
Thanks for any help with this.
Regards,
-Orlando