[Insight-users] Big-O time complexity of level set segmentation?

[Zachary Pincus]

Hello,

Does anyone know what the approximate time complexity of the level set 
segmentation methods in ITK are?

I ask because I have a task that involves segmenting multiple cells per 
image field. I can either segment each cell separately (grow a new 
level set for one cell, then the next, etc) or segment in parallel 
(seed the level set at, say, the centroids of every cell at once). 
Which methods is preferable depends on the big-O time complexity of 
level set segmentation.

For example, if the level set methods are, say, O(n log n), where n is 
the size of the front, then it clearly makes sense to segment each cell 
individually. If the complexity is O(n) then it doesn't matter whether 
you do each cell individually, or all in parallel. And of course, if 
it's O(log n) then there is substantial time savings to doing the image 
at once.

Does anyone have an idea about the complexity of these methods, as 
implemented by ITK?

Thanks,

Zach Pincus

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