[Insight-users] Watershed on DCM 3D images

[Luis Ibanez]

Hi Subbu,


1) Yes it is possible to use Watershed in 3D images.
    It doesn't really matter that they are CT or that
    they have been read from DICOM files.


2) You don't need to convert your data to RGB. The
    Watershed filter works on scalar images.

    The RGB conversion that you see in the Software Guide
    is done in order to color the many regions that are
    the labelled *output* of the watershed filter. This
    is done *only* for visualization purposes.


3) No, there are no standard values for the parameters.
    However the suggested values to start exploring is to
    use a water level of about 0.2 (the water level is
    already normalized to the highest value in the image),
    and a noise threshold value of about 0.01.  Again,
    those are just initial values from which you could
    start your exploration of parameters.



    Note that in practice, you probably want to pass
    as input not the CT image, but the gradient magnitude
    of this image, so that you have independent basins
    associated with anatomical structures in the image.


    Your pipeline maybe something like:


     Reader->GradientMagnitude->Watershed->ColoringFilter




   Regards,


      Luis


---------------------
K. Subburaj wrote:
> Hi all
> 
> 1. I would like to apply the watershed filter to 3D CT (DCM) images, is it
>    possible? In guide, it 's specified that RGB values are required.
> 
> 2. How can I do?, do we need to use any caster to convert dcm  scalar
> value to RGB values...,
> 
> 3. is there any standard for number of iterations and other values
> 
> i tried Examples/Segmentation/Watershed1.cxx on dcm 2d images it takes
> long time  to process , Can any one comment on this, why so?
> 
> Thanks / Subbu
> 
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users
>