[Insight-users] LaplacianSegmentationLevelSetImageFilter

[Luis Ibanez]

Hi Maria,

Most of the LevelSet segmentation filters in ITK
are based on solving differential equations by
iteratively computing finite differences.

The methods are considered to converge when the
difference between the field values of one iteration
and the previous iteration are below a certain
threshold. The value to compare against that threshold
is a RMS error (square root of the sum of squared errors).

The smaller you set the "MaximumRMSError" the more
demanding you will be with the convergence of the
level set, and therefore the more iterations it
may take to reach that level of stability.

It is usually a good practice to connect an observer
to this filter and print out (or plot) the current
RMS value versus the iteration number in order to
have an idea on how the filter is converging...

and whether you will have a
chance to have lunch or not  :-)

You will have many examples on how to connect
Observers to Filters in the ITK Software Guide


    http://www.itk.org/ItkSoftwareGuide.pdf



    Regards,


       Luis


-------------------------
Maria Cira Avvinto wrote:
> How can i undestand the meaning of RMSError in
> LaplacianSegmentationLevelSetImageFilter?
> Why MaximumRMSError is 0.002 ?
> Thanks,
> Maria
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users
>