[Insight-users] time step in itkLevelSetFunction

[Luis Ibanez]

Hi Valerie,

The basic heuristic for selecting the time step
of the Level Set Function is that during one
iteration of the LevelSet field computation,
the LevelSet front should never advance more
than one pixel.

Note that the time step could have always been
taken smaller, at the cost of requiring more
iterations for a LevelSet to advance a certain
distance in the physical space of the image.


    Regards,


       Luis


--------------------
Valérie Duay wrote:
> Hi,
> 
> I would like to reask the same question as an old mail sent to this 
> mailing list last summer (see below).  Apparently, it didn't receive any 
> answer. I notice that m_DT is linked to curvature-based evolution terms 
> and  m_WaveDT  to the other evolution terms. Is it because 
> curvature-based terms is more sensitive to numerical instabilities than 
> the others and could need a smaller dt?
> 
> How did you choose the following values for  m_WaveDT  and m_DT? Is 
> there another justification than the one given by Sethian?
> template< class TImageType >
> double LevelSetFunction<TImageType>::m_WaveDT = 1.0/(2.0 * ImageDimension);
> 
> template < class TImageType >
> double LevelSetFunction<TImageType>::m_DT     = 1.0/(2.0 * ImageDimension);
> 
> Thanks in advance for your help,
> 
> Valerie
> 
> "
> /Sat Aug  6 13:15:37 EDT 2005/
> Can someone explain why the time step is computed the way it is in
> the function ComputeGlobalTimeStep in itkLevelSetFunction.txx?
> I mean the mathematical justification. What are m_WaveDT and m_DT? 
> In Level Set Methods, Sethian only mentions (page 67, section 6.4.1) 
> 
> max F \Delta T <= \Delta x
> 
> A pointer to the appropriate theory would suffice. 
> 
> Thank you!
> "
> 
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users