[Insight-users] level set stopping criteria

[Luca Antiga]

Hi Burak,
  level set segmentation in ITK uses the sparse field formulation,
in which the only front that's actually tracked (the active layer) is  
the one around
the zero level set. If you look at the  
itk::SparseFieldLevelSetImageFilter class, you'll
see that the rms change is only computed from the changes in the  
active layer.
Therefore, once your zero front stops, the RMS will drop.
Regards

Luca


--
Luca Antiga, PhD
Head, Medical Imaging Unit,
Bioengineering Department,
Mario Negri Institute
email: antiga at marionegri.it
web: http://villacamozzi.marionegri.it/~luca
mail: Villa Camozzi, 24020, Ranica (BG), Italy
phone: +39 035 4535-381


On Sep 12, 2007, at 10:30 AM, Burak Özkalaycı wrote:

> Dear all,
>
> I am making some level set evolution experiments using ITK and I could
> not figure out how the RMS change is calculate to stop iteration. I
> just initialized the level set with a signed distance function to a
> cube and assigned a velocity map which is a bigger cube inside all 0.5
> and outside -0.5. The initial cube grows to the bigger one and stops
> correctly. In the level set function formulation I just used the
> propagation term. The algorithm stops with a RMS error equal to zero.
> If the RMS error is just calculated by summing ( Phi(t+1)-Phi(t) )^2
> terms at all nodes, than the RMS error may not decrease I think.
> Because the other levels like Phi=1 or Phi=-1 should continue to
> propagate, if I am not wrong. Can anyone make clear how the RMS error
> is calculated? Thank you in advance.
>
>
>  Burak Ozkalayci
>
>  PhD Student
>  Electronic and Information Systems (ELIS-MEDISIP)
>  Ghent University Hospital - IBITECH
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users