[Insight-developers] Vote on a name for real-to-complex and complex-to-real FFT filters

[Cory Quammen]

Dear developers,

I've posted this patch to make the concrete FFT filters output images
consistent regardless of which underlying FFT library implementation
is used: http://review.source.kitware.com/#change,2625

For the main forward (inverse) FFT filter, the output (input) is going
to be the full size complex image, including the redundant half you
get whenever you take the discrete Fourier transform of a real image.

To satisfy performance-minded users, I've also added specialized
classes that exploit the conjugate symmetry in the DFT of a real image
by producing an output image about half the size of the full complex
image where the missing half would contain only redundant pixel
values. In the patch, these are currently called
itkRealToComplexForwardFFTImageFilter and
itkComplexToRealInverseFFTImageFilter.

The problem is that nobody seems to like those names (besides me). To
move this patch along, I'd like to take a poll on the names proposed
so far. Please vote for your favorite (or the one you would look for
on the doxygen page if you wanted to use these filters in your code).


itkHalfForwardFFTImageFilter / itkHalfInverseFFTImageFilter

itkHalfMatrixForwardFFTImageFilter / itkHalfMatrixInverseFFTImageFilter

itkRealToComplexForwardFFTImageFilter / itkComplexToRealInverseFFTImageFilter

itkRealToHalfComplexForwardFFTImageFilter /
itkHalfComplexToRealInverseFFTImageFilter

itkRealToConjugateSymmetricForwardFFTImageFilter /
itkConjugateSymmetricToRealInverseFFTImageFilter


Vote now or forever be annoyed at how the selected filter names are terrible.

Thanks,
Cory

-- 
Cory Quammen
Research Associate
Department of Computer Science
The University of North Carolina at Chapel Hill