[Insight-users] Mattes Mutual Information

[brian avants]

there is a brief summary of the v4 framework here:

https://github.com/stnava/ITKv4Documentation/blob/wbir2012/ITKv4RefactoringNotes.pdf

the most salient points:

* all metrics are multi-threaded

* all metrics may be computed from sparse or dense samples

* all metrics may be used with low or high-dimensional (e.g.
diffeomorphic, syn) transformations + generic pixel types

* optimization / transformation framework reworked to be memory
efficient i.e. avoid copying (potentially large) parameter vectors

* new metrics for vector pixels (e.g. RGB) use same base classes

* optimizers employ gradient step sizes with physical units, line
search and have the ability to automate parameter scaling

* composite transforms are now trivial.

* "fixed" images can have transforms too (useful in unbiased
registration strategies)

* point set metrics can be combined with image metrics via
itkObjectToObjectMultiMetricv4

* some global optimization tricks are available in the
MultiStartOptimizer ... very useful.

there is also a GPU registration framework though that is a separate beast.


brian




On Wed, Jan 9, 2013 at 7:49 AM, Matt McCormick
<matt.mccormick at kitware.com> wrote:
> Hi Tim,
>
> The v4 registration framework slightly different, and the
> MattesMutualInformationImageToImageMetric can only be used with the v3
> framework and the MattesMutualInformationImageToImageMetricv4 can only
> be used with the v4 framework (the v4 has many additional features).
> Additionally, I believe the implementations may be slightly different.
>
> Thanks,
> Matt
>
> On Wed, Jan 9, 2013 at 7:35 AM, Tim Allman <dr.tim.allman at gmail.com> wrote:
>> There are two classes which implement the Mattes mutual information metric,
>> MattesMutualInformationImageToImageMetric and
>> MattesMutualInformationImageToImageMetricv4. There is little documentation
>> or indication why I would want to use one or the other or why the v4 version
>> was created. Could someone give me a little guidance?
>>
>> Thanks,
>> Tim
>>
>> --
>> Tim Allman, Ph.D.
>> 35 Margaret Street,
>> Guelph Ontario N1E 5R6
>> Canada
>>
>>
>> _____________________________________
>> 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://www.itk.org/mailman/listinfo/insight-users
>>
> _____________________________________
> 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://www.itk.org/mailman/listinfo/insight-users