[Insight-users] canny level set segmentation

[Luis Ibanez]

Hi Jesse,

Canny edges may not be a good method for segmenting blood
vessels, specially is these vessels are very fine.

The Canny edges detect derivatives of intensity, and therefore
they need a support of more than a couple of pixels. That is,
the intensity transition should happen in a band.

Small vessels may not be able to produce significant Canny
edges on both sides of the vessel.

You may find useful to look at the many methods available
for detecting vasculature.

In particular, you should look at the applications:


     InsightApplications/
            Curves2DExtractor
            Curves3DExtractor
            DuctExtractor


as well as the classes:

           Insight/Code/Review/
itkHessianToObjectnessMeasureImageFilter.h
itkMultiScaleHessianBasedMeasureImageFilter.h



------


About your second question:

The output of the CannySegmentationLevelSetImageFilter is
a level set.  For a description of the representation of Level Sets
please refer to the ITK Software Guide:

             http://www.itk.org/ItkSoftwareGuide.pdf

in the "Image Segmentation" chapter.


     Regards,


           Luis


-----------------------------------------------------------------------------------
On Fri, Mar 26, 2010 at 9:51 PM, Jesse Stokum <jessestokum at gmail.com> wrote:
> Hello,
>
> I've been trying to use the CannySegmentationLevelSetImageFilter which
> has been met with limited success. Firstly, does anyone have
> suggestions for parameters to pick out fine vasculature - the vessels
> in my segmentation are somewhat "bubbly" or "chunky" rather than
> snakelike. I'm running 300 iterations of this level set after a
> threshold segmentation.
>
> Secondly, the output from this filter (I've been using the code from
> the examples by the way) seems to be a speed image when visualized
> with itk snap; It's not a simple binary back/foreground image. Does
> someone know what I'm looking at? This is the image from calling
> GetOutput(). Any help is appreciated and thanks in advance.
>
> Jesse
> _____________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Kitware offers ITK Training Courses, for more information visit:
> http://www.kitware.com/products/protraining.html
>
> Please keep messages on-topic and check the ITK FAQ at:
> http://www.itk.org/Wiki/ITK_FAQ
>
> Follow this link to subscribe/unsubscribe:
> http://www.itk.org/mailman/listinfo/insight-users
>