[Insight-users] problem with mutual information metric in 3d registration

Aviv Hurvitz aviv.hurvitz at gmail.com
Thu Oct 23 12:41:21 EDT 2008


Here's one idea.

Make sure you call:
 optimizer->MaximizeOn() for MutualInformationImageToImageMetric and
 optimizer->MinimizeOn() for MattesMutualInformationImageToImageMetric

- Aviv



On Thu, Oct 23, 2008 at 11:25 AM, Giorgos Pichis <gpichis at gmail.com> wrote:

> Hello all,
> I am trying to implement an 3d registration algorithm using
> MutualInformationImageToImageMetric.
> The problem is that it throws an exception:
> " All the sample point mapped to outside of the moving image "
>
> Details about my program:
> 1. I have done the same algorithm with MattesMutualInformation and works
> fine. The only changes , are about the metric,
> and the preprocessing step for the images (Normalization and Smoothing)
> 2. I am trying to do this algorithm with VectorRigid3D and Similarity3D
> Transforms. In both cases Mattes works fine, but ViolaWells implementation
> throws the same error. (I am using VectorRigid3DTransformOptimizer and
> RegularStepGradientDescentOptimizer respectively)
> 3. The images I am using are from the BrainWeb Image Repository, and are MR
> T1 and T2
>
> Things I have tried but don't work:
> 1. Changing the initial maximum step length (values, 0.01, 0.1, 1, 2, 4...)
> 2. Changing Vector and Transform scales in the optimizer
> 3. Changing Image's Pixel Type (unsigned char / short,  or float)
>
> Do you have any clue of what should I do?
>
> Thank you for your time.
>
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-users/attachments/20081023/de2fdf0d/attachment.htm>


More information about the Insight-users mailing list