[Insight-users] watershed segment tree merging

[Joshua Cates]

Yes, that's the general procedure.

Josh.


On 20 Nov 2003, T.J. Rudge wrote:

> Thanks Joshua. Am I right in thinking that I should:
> 
> 1/ Modify the segment_t structure to include any new values about the 
> regions 2/ Change WatershedSegmenter to calculate the additional values 3/ 
> Change WatershedSegmentTreeGenerator to produce merges based on the new 
> values.
> 
> Tim
> 
> On Nov 20 2003, Joshua Cates wrote:
> 
> > Hi Tim,
> > 
> > I think this is a great idea. The WatershedImageFilter is constructed 
> > from a set of subfilters, one of which is called the 
> > SegmentTreeGenerator. This is the filter that performs region merging to 
> > generate the tree. You could start by copying this filter and modifying 
> > it, replacing the original with your modified filter in 
> > WatershedImageFilter. In fact, this was the purpose of making these 
> > modular subfilters.
> > 
> > You may find you need some additional information in the "SegmentTable",
> > which is the table of initial watershed transform regions generated from
> > the watershed::Segmenter object.  In this case, you will also need to
> > modify the SegmentTable object and code in the Segmenter filter.
> > 
> > Let me know how this goes and feel free to contact me with questions. We
> > should incorporate your additions into the ITK repository if you are
> > willing.
> > 
> > Josh.
> > 
> > ______________________________
> >  Josh Cates			
> >  Scientific Computing and Imaging Institute
> >  University of Utah
> >  Email: cates at sci . utah . edu
> >  Phone: (801) 587-7697
> >  URL:   http://www . sci . utah . edu/~cates
> > 
> > 
> > On 20 Nov 2003, T.J. Rudge wrote:
> > 
> > > Hi,
> > > 
> > > Im using the watershedImageFilter. I;d like to be able to use the 
> > > merge tree generated by the watershed system and add additional 
> > > criteria for the merge decision - based on properties of the objects, 
> > > e.g. max/min region size. This way merges could be forced or prevented 
> > > according to a model of the object region shape size etc. Do you think 
> > > this is a reasonable strategy? Any advice on the best way to go about 
> > > it?
> > > 
> > > many thanks,
> > > 
> > > Tim R.
> > > _______________________________________________
> > > Insight-users mailing list
> > > Insight-users at itk . org
> > > http://www . itk . org/mailman/listinfo/insight-users
> > > 
> > 
> >
>