[Insight-users] Plotting values of Mutual Information Metric, BSplineTransform

[Luis Ibanez]

Hi Christian,

You could plot the values of the metric with respect to the parameters
of a BSpline Transform.... but... the challenge is that a BSpline transform
typically has hundreds of parameters.

For example a 10x10 grid in a a 2 image will have 200 parameters....

Therefore you will have a cost function that should be plotted
in 200-dimensional space...

The example in the Software Guide is for a 2D Translation transform,
that conveniently only have two parameters, and therefore can be easily
plotted as a 2D surface in a 3D space.


You could take selected pairs of parameters out of the N from your
BSpline transform, and plot the metric for only that pair....

It may still be interesting, since it could show you how the modification
of a particular node in the grid will affect the overall value of the
Metric.



      Regards


             Luis


-----------------------------------------------
On Fri, Jul 24, 2009 at 8:35 AM, c23562 <c.popko at gmx.de> wrote:

>
> Hi all,
>
>
> In ItkSoftwareGuide (page 418) is an example for plotting Mean Squares
> Metric. I would like to plot the metric values of Mattes Mutual
> Information.
> But the transform is not a translation but a BSplineTransform. Is it
> possible too? With the plot i hope to find better parameters for
> regularStepGradientDescentOptimizer.
>
>
> Best regards,
> Christian.
> --
> View this message in context:
> http://www.nabble.com/Plotting-values-of-Mutual-Information-Metric%2C-BSplineTransform-tp24643762p24643762.html
> Sent from the ITK - Users mailing list archive at Nabble.com.
>
> _____________________________________
> 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-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-users/attachments/20090728/3d7a03cc/attachment-0001.htm>