[Insight-developers] So -- no concensus on smoothing re MultiResolutionPyramidImageFilter?

[Hans Johnson]

Since this is considered a bug for over a year
http://www.vtk.org/Bug/view.php?id=7002 we should consider the fact that
MultiResolutionPyramidImageFilter smooths for unit shrink factors a bug that
would allow us to fix it (even through numerical sameness/backwards
compatibility will be lost).

This is not the behavior that I would expect by reading the documentation on
how an Image Pyramid should be constructed (i.e. The top level should be an
exact copy of the reference images).

I think that this is a bug that also affects the
RecursiveMultiResolutionPyramidImageFilter because it would produce
different results for the unit shrink factor image if I change the shrink
schedule from being downward divisible to being non-downward divisible.
This is particularly nasty because given the same "shrink schedule
[8,4,2,1]" the unit shrink result will be different depending on the size of
the input image.

===============
If there are no objects today, then the solution that will be implemented
will be:

1) Add a test that fails when the unit shrink image is not the same as the
reference image (and the m_SmoothIfUnitShrinkFactors =false).
2) Add the m_SmoothIfUnitShrinkFactors conditional flag to allow explicit
smoothing of the unit shrink factors.
3) Set the default to "false" in all cases.
===============


Regards,
Hans


On 4/9/09 1:56 AM, "Tom Vercauteren" <tom.vercauteren at gmail.com> wrote:

> Hi Kent,
> 
> Even if there is no consensus about what the default should be, it
> could be add a flag that lets the user decide (as Stephen suggested).
> 
> Maybe something like:
>   m_SmoothIfUnitShrinkFactors
> 
> For backwards compatibility, the default for this flag would be:
>   - true for MultiResolutionPyramidImageFilter
>   - false for RecursiveMultiResolutionPyramidImageFilter
> 
> Now, I also would like to propose a NON-backward compatible change to
> fix something that might be very confusing to the user.
> 
> Since RecursiveMultiResolutionPyramidImageFilter wil call
> MultiResolutionPyramidImageFilter if the schedule is not downward
> divisible, I suggest that RecursiveMultiResolutionPyramidImageFilter
> sets
>   
> internalNonRecursivePyramid->SetSmoothIfUnitShrinkFactors(this->m_SmoothIfUnit
> ShrinkFactors)
> when the schedule is not downward divisible.
> 
> With this small non-backward compatible change, it would at least make
> things a bit clearer.
> 
> Otherwise we might need a three state flag for
> RecursiveMultiResolutionPyramidImageFilter:
>   - AlwaysSmoothIfUnitShrinkFactors
>   - NeverSmoothIfUnitShrinkFactors
>   - SmoothIfUnitShrinkFactorsAndScheduleDownardDivisible (this being
> the default for true backwards compatibility)
> 
> Thoughts anyone?
> 
> Tom
> 
> On Wed, Apr 8, 2009 at 23:27, kent williams <norman-k-williams at uiowa.edu>
> wrote:
>> It seems Hans assigned me yesterday to fix this 'bug':
>> 
>> http://www.vtk.org/Bug/view.php?id=7002
>> 
>> As is noted there, there still doesn't appear to be a consensus on how to
>> handle this problem.  So I've taken ownership of a bug with no clear way to
>> fix it.
>> 
>> If the issue is truly tabled, does it need to hang around as an open bug?
>> 
>> 
>> 
>> Notice: This UI Health Care e-mail (including attachments) is covered by the
>> Electronic Communications Privacy Act, 18 U.S.C. 2510-2521, is confidential
>> and may be legally privileged.  If you are not the intended recipient, you
>> are hereby notified that any retention, dissemination, distribution, or
>> copying of this communication is strictly prohibited.  Please reply to the
>> sender that you have received the message in error, then delete it.  Thank
>> you.
>> 
>> 
>> _______________________________________________
>> Powered by www.kitware.com
>> 
>> Visit other Kitware open-source projects at
>> http://www.kitware.com/opensource/opensource.html
>> 
>> 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://www.itk.org/mailman/listinfo/insight-developers
>> 
> _______________________________________________
> Powered by www.kitware.com
> 
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
> 
> 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://www.itk.org/mailman/listinfo/insight-developers