[Insight-users] A couple of questions about level sets

[Zachary Pincus]

Hello all,

I'm still trying to figure out the "best practices" for implementing a 
level-set based segmentation system with ITK. I asked a couple of 
questions a few weeks ago, but they probably got lost in the shuffle 
with the 2.0 release and whatnot.

Anyhow, I've been wondering about the recommended way of creating a 
level set method that uses "custom" speed and edge advection images. It 
seems that one could either:

(a) create a subclass of SegmentationLevelSetImageFilter and 
SegmentationLevelSetFunction, where these subclasses create the custom 
speed and advection images by re-implementing GetSpeedImage(), etc. 
This is how the various methods like Curves or Shape Detection are 
implemented in ITK.

(b) create a SegmentationLevelSetImageFilter object (or an object of 
some particular subclass thereof) and then set custom advection and 
speed imaged with that class's methods for this purpose.

Are these approaches functionally equivalent? Are there caveats with 
one or the other?

Thanks,

Zach Pincus

Department of Biochemistry and Program in Biomedical Informatics
Stanford University School of Medicine