[Insight-users] 3D skeleton/centerline

[Ajay Sonar]

Hi Ruben,

I reffered to two papers to extract the centerlines,

A Robust Level-set algorithm for centerline extraction.
An augmented Fast marching method for computing skeletons and centerlines.

both from Joint EUROGRAPHICS - IEEE TCVG Symposium on Visualization (2002 -
2003).

I implimented this in matlab. I still have to figure out why i am not
getting a single voxel wide centerline and also want to see if i can fit a
line through the centerline voxels even if its not 1 voxel wide and use that
as reference.

Ajay.


On 5/24/07, Ruben Schilling <r.b.schilling at googlemail.com> wrote:
>
> Hi Ajay,
>
> I actually want to do the same thing and have some questions and some
> answers. I can share how I approach this problem at the minute. I use the
> Danielsson Distance map from ITK to calculate a distance transform. You
> would then have to extract local maxima from it, that correspond to your
> skeleton. This approach requires preprocessing to fill wholes, which can be
> done in itk via the iterative fill holes filter.
>
>
> My question:
>
>
> I would like to have 1 voxel thin lines or even better: a discrete graph
> extracted.
>
>
> Can anyone tell me (us) how to do this?
>
>
> Best
> Ruben
>
>
>
>
>
>  Am 24.05.2007 um 16:27 schrieb Ajay Sonar:
>
>  Hi All,
>
> I am new to ITK. Can anyone tell me if there is a way to extract the
> skeleton lines/center lines from a stack of image slices. I have a set of
> microCT images and i can get the surface or the voxel model from it in
> MATLAB. I need to extract the skeleton lines from it. I tried a few things
> but cannot get a single voxel width centerline.
>
>
> Thanks,
> Ajay.
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://public.kitware.com/pipermail/insight-users/attachments/20070524/06ebd40b/attachment.htm