IF3IF3 and IF2IF2

Engineering
1

I am using ITK in Python. I have a general question in dealing with 3D images.

Most of the examples in https://itk.org/ITKExamples/src/index.html are for 2D images. When I apply the code to 3D images, I got TypeError frequently.

Somehow, I realize that there is an option “IF3IF3” versus “IF2IF2” when declaring a filter, which could solve my problem.

For example, instead of using

filter = itk.GradientMagnitudeRecursiveGaussianImageFilter.New()

I should use

filter = itk.GradientMagnitudeRecursiveGaussianImageFilter.IF3IF3.New()

My first question is: whether this is the correct way to deal with 3D data?

My second question is: what is the meaning of “IF3IF3”/“IF2IF2” and “ISS3ISS3”/“ISS2ISS2”?

Thanks in advance,
Jianxu

2

Hi @Jianxu_Chen,

A good practice is to use:

filt = itk.GradientMagnitudeRecursiveGaussianImageFilter.New(image_or_filter)

And the filter type created will be the correct dimension for the input image or filter output. It will work for both 2D and 3D images.

This is the name mangling for the wrapped types. For example, IF3IF23 refers to itk::Image< float, 3 >, itk::Image< float, 3 >. For more information on the naming scheme, see the Wrapping section of the ITK Software Guide.

Hope this helps,
Matt

3

For some reason, itk.GradientMagnitudeRecursiveGaussianImageFilter.New() does not work for me.

For example, I have an image, itk_img_smooth

type(itk_img_smooth)

returns itkImagePython.itkImageF3

itk_img_smooth.GetLargestPossibleRegion().GetSize()

returns itkSize3 ([149, 139, 60])

Then,

gradientMagnitude = itk.GradientMagnitudeRecursiveGaussianImageFilter.New()
gradientMagnitude.SetSigma(sigma)
gradientMagnitude.SetInput(itk_img_smooth)
gradientMagnitude.Update()

does not work. The error message is

Expecting argument of type itkImageSS2 or itkImageSourceISS2

But the following code works

gradientMagnitude = itk.GradientMagnitudeRecursiveGaussianImageFilter.IF3IF3.New()
gradientMagnitude.SetSigma(sigma)
gradientMagnitude.SetInput(itk_img_smooth)
gradientMagnitude.Update()
4

Oops … I know what you mean now.

I should use

gradientMagnitude = itk.GradientMagnitudeRecursiveGaussianImageFilter.New(itk_img_smooth)
gradientMagnitude.SetSigma(sigma)
gradientMagnitude.Update()

Great thanks!
Jianxu

2 Likes
5

Great!

Additionally, more information can be found in the Python quick start guide:

https://itkpythonpackage.readthedocs.io/en/latest/Quick_start_guide.html