[Insight-users] Watershed from marker

[Luis Ibanez]

Hi Lagaffe,

Thanks for your clarification.

Once you decide to actually use Markers (seed points), the operation
that you describe seems to be equivalent to using a region growing
method on the gradient magnitude of the image.


To that end you can simply use the filter:

http://www.itk.org/Insight/Doxygen/html/classitk_1_1ConnectedThresholdImageFilter.html

on the output of the GradientMagnitud filter such as


http://www.itk.org/Insight/Doxygen/html/classitk_1_1GradientMagnitudeImageFilter.html

or

http://www.itk.org/Insight/Doxygen/html/classitk_1_1GradientMagnitudeRecursiveGaussianImageFilter.html



Of course, you always have option (2) that you mention:
Implementing a specialized version of Watersheds.


Please let us know if you have further questions,


   Thanks


     Luis


---------------
Mr Gaffe wrote:

> Hi all,
>  
> By initials markers image I mean initials seeds. Is there are no intials 
> seeds or markers the watershed start growing from the local minima of 
> the image, which usually create an over segmentation even if you use a 
> diffusion filter ....The startegy is to impose the minima on the image 
> with your seeds inside and outside your tumor for example.
> So, you have two solutions:
> 1- Impose your minima on your image and use a normal watershed
> 2- Or basicaly, use a "special" watershed which start growing from the seeds
>  
> Lagaffe,
> 
> 
> */Luis Ibanez <luis.ibanez at kitware.com>/* wrote:
> 
> 
>     Hi Lagaffe,
> 
>     What do you mean by "Initial Marker Image" ?
> 
>     Please give us a better description of your problem.
> 
> 
>     Thanks
> 
> 
>     Luis
> 
> 
>     -----------------
>     Mr Gaffe wrote:
> 
>      > Hi all,
>      > It seems that there is no possibility to do a watershed from an
>     inital
>      > marker image ?
>      > So, the strategy is to impose the minima using the geodesic
>     function and
>      > then use the watershed filter ?
>      > thanks,
>      >
>      > Lagaffe.
>      >
>      >
> 
> 
>