[vtkusers] detecting overlapping cells in a polydata

[Kenneth Sloan]

On Jun 30, 2008, at 5:50 PM, Karl Merkley wrote:

>
> Anybody have any ideas on how to efficiently detect overlapping  
> cells in a polydata?
>
>    Thanks,
>        Karl
>
In my experience, the best way is to have a good model of the size/ 
shape of individual cells.  Overlapping cells are "too big" if the  
two cells appear to be merged, or one of them is "too small" if it is  
separate from, but obscured by, the other cell.  In both cases, the  
cell(s) are the "wrong shape".

That gives you a start on *detecting* overlapping cells.  Once  
detected, you need a way to hypothesize a configuration of cells that  
explains what you see in the image.

This takes you out of the area of straight "image processing", but  
(again, in my experience) it appears to be necessary to get good  
results.

--
Kenneth Sloan                                           
KennethRSloan at gmail.com
Computer and Information Sciences                        +1-205-934-2213
University of Alabama at Birmingham              FAX +1-205-934-5473
Birmingham, AL 35294-1170                http://KennethRSloan.com/