[Insight-users] itk::SparseFieldLevelSetImageFilter

[Koen Vermeer]

Hi Brady,

I'd reason like this: If 1 is true, then it would introduce spurious
zero level sets (resulting from the zero-crossings due to the sign
changes) and there would be no way to know which zero level set is the
'right' one. So only 2 gives meaningful results, hence, that's probably
the one that's implemented.

>From a more practical point of view: My experience with
itk::ThresholdSegmentationLevelSetImageFilter is that the sign is
meaningful at all locations.

Best,
Koen


On Tue, 2008-11-25 at 21:34 -0600, Brady McCary wrote:
> At every iteration I am interested in the sign of the level set
> function at every pixel, not just those close to the zero level set. I
> am not sure what the documentation is saying. It is probably saying
> one of these:
> 1. During all iterations, only pixels close to the zero level set have
> meaning values and meaningful signs.
> 2. During all iterations, only pixels close to the zero level set have
> meaning values, but all pixels have meaningful signs.
> Could someone clarify this point?