[Insight-developers] Re: [Insight-users] Feedback wanted for additions to NeighborhoodIterator classes

Joshua Cates cates@sci.utah.edu
Mon, 25 Nov 2002 11:53:57 -0700 (MST)


Hi Luis,

Yes I would like to add this "arbitrarily-shaped ND neighborhood"
functionality.  There are many decisions to be made regarding the API,
which I am afraid will end up being very complicated.  I'll post a
proposed API for such an object soon. In the meantime if you, (or anyone
else!) has API ideas, please let me know.

Josh.

______________________________
 Josh Cates			
 School of Computer Science	
 University of Utah
 Email: cates@sci.utah.edu
 Phone: (801) 587-7697
 URL:   www.cs.utk.edu/~cates


On Mon, 25 Nov 2002, Luis Ibanez wrote:

> Hi Josh,
> 
> 
> How about adding a way of selecting the connectivity.
> 
> Several users have asked about using
> 
>    26-connectedness
>    18-connectedness
>     6-connectedness
> 
> in 3D, for example.
> 
> It could be nice to be able to feed a kind of list of
> clicks to the Neigborhood iterator. This will be like
> the set of neigbors to be visited. This will be useful
> for doing advance morphomath too.
> 
> It will also facilitate to write cellular automatas
> to run in itkImages. After all, the PDE solvers, the
> convolution operators and the mathematical morphology
> operations are all subsets of cellular automata.
> 
> 
> 
>      Luis
> 
> 
> 
> ==================================================
> 
> 
> Joshua Cates wrote:
> 
> > Hello all ITK developers and users:
> > 
> > I am planning to make some simple additions to the API of the neighborhood
> > iterator classes (and base classes) and would like some feedback.
> > 
> > I'm particularly interested in suggestions for naming of methods and
> > classes and any additional functionality that people would like to see in
> > the NeighborhoodIterators and NeighborhoodOperators.
> > 
> 
> .... snip
> 
> > 
> > 
> > 3) Create a new class which only manages pointers to the 2N-connected
> > neighbor pixels as it iterates in N dimensions.  This should give some
> > speed advantage to simple algorithms in higher dimensions.
> > 
> > itk::CityBlockConnectedSmartNeighborhoodIterator
> > 
> > 
> > 4) Other iterator types?  I'm open to suggestions.
> > 
> > Thanks,
> > 
> > Josh.
> > 
> > 
> > ______________________________
> >  Josh Cates			
> >  School of Computer Science	
> >  University of Utah
> >  Email: cates@sci.utah.edu
> >  Phone: (801) 587-7697
> >  URL:   www.cs.utk.edu/~cates
> > 
> > 
> 
> 
>