[ITK-users] Questions on the itkANTSNeighborhoodCorrelationImageToImageMetricv4GetValueAndDerivativeThreader.hxx

[Dženan Zukić]

Hi Jiancong,

I have not seen this class before, so I can't answer your questions.
However I can recommend you to run
itkANTSNeighborhoodCorrelationImageToImageMetricv4Test.cxx under debugger
and inspect the loop of interest. Then change ImageDimension to 3 and run
it again. It looks like the test was written with such experimentation in
mind, so it should run. You might also play with imageSize (currently 6) to
have a smaller or large image.

Hopefully this will help you to understand how that code works.

Regards

On Fri, Feb 3, 2017 at 11:50 PM, jiancong wang <10811130d at gmail.com> wrote:

> Hi guys,
> I am working on replicating this image registration metric on gpu and has
> some questions that I want to ask.
>
> In both of the UpdateQueue functions I saw a line like this:
> ......
> for ( SizeValueType indct = i; indct < hoodlen; indct += ( diameter +
> NumericTraits<SizeValueType>::OneValue() ) )
> ......
>
> Why the index is incremented by the diameter of the NCC window instead of
> by 1? Isn't this suppose accumulate local pixel values for reuse purpose?
> Also why there is only one for loop here? How do they handle 3D data in
> this case?
>
> I read the original paper and they are trying to cache the local sum value
> (for example for a 7 by 7 window they are supposed to accumulate 7by1
> column window, that each time it proceeds it only has to recompute the next
> 7by1 column instead of the full 7by7 window).
>
> Any help is greatly appreciated! Thanks!
>
> _____________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Kitware offers ITK Training Courses, for more information visit:
> http://www.kitware.com/products/protraining.php
>
> Please keep messages on-topic and check the ITK FAQ at:
> http://www.itk.org/Wiki/ITK_FAQ
>
> Follow this link to subscribe/unsubscribe:
> http://public.kitware.com/mailman/listinfo/insight-users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/insight-users/attachments/20170204/d97afe84/attachment.html>