Installing itk-vtkglue module using Pip

Hi all,

I am trying to install the itk-vtkglue module using pip, as I would like to use the transform inverse functions from vtk on some .nrrd data. However, when I use “python -m pip install itk-vtkglue” I get the following error message:

Could not find a version that satisfies the requirement itk-vtkglue (from versions: )

No matching distribution found for itk-vtkglue

I would much prefer it to use pip to install the module instead of going through CMake etc. Is this possible?

Thanks in advance,

RK

Hi RK,

We are currently expanding the currently supported distributions. What is your Python version and operating system?

Thanks,
Matt

Hi Matt,

Thank you for your quick reply. I am working on Windows 7 and using python 3.6.3.

I encountered the same issue on Mac OS X Mojave 10.14

Windows wheels have been uploaded to PyPI for Python 3.6 :sparkler: . Please try

python -m pip install itk-vtkglue

Hi Matt,

Thank you so much for uploading the windows wheel, which is amazing!

Extra question if you don’t mind: I started rewriting the vtkWarpTransform method vtkWarpInverseTransformPoint for Matlab, but couldn’t quite find how TemplateTransformPoint(inverse,deltaP,derivative) (line 145) worked. The input for this function is a point (inverse) and the displacement vector at that point, the output the displaced point (deltaP) and the derivative, if I understand correctly. But how is this derivative calculated? I hope you don’t mind me rewriting it to Matlab, I will of course cite the VTK.

Thanks for your amazing work!

RK

vtkWarpTransform is part of VTK. If the docs don’t answer your question, you should ask for advice on one of VTK mailing lists.

Thank you, I will!

Note that there are similar classes in ITK:

Studying these classes and re-implementing their functionality in Matlab is a good exercise to understand the operation. And, it will help build an appreciation of why ITK’s computations in physical space, which are difficult in Matlab, and ITK’s performance make it preferable over Matlab ;-).

1 Like