[Tubetk-developers] semi-automated vessel segmentation

[Stephen Aylward]

Hi Greg,

Thanks for the note.   TubeTK is still a work in progress, so expect
much improved segmentation and registration capabilities soon.

For interpolating between manual segmentations, we might be able to
help.    The tube data structure in TubeTK represents all tubes as
x,y,z,r points.   Where x,y,z are spatial coordinates, and r is the
radius.   So, it assumes a circular cross-section for the tubes.    If
that is sufficient for you, then you can store the x,y,z,r values of
your manual segmentations in a TubeTK tube file (i.e., and
itkTubeSpatialObject), and then we could provide spline interpolation
code to resample those points at whatever density you would like.   We
have code for this ("itkTubeToInterpolatedTubeFilter"), but we haven't
checked it into TubeTK yet.

So, is the assumption of having a circular cross-section acceptable for you?

Stephen

On Mon, Nov 26, 2012 at 1:22 PM, Greg ----- <gscizz1979 at gmail.com> wrote:
> First off: I love TubeTK!  Thanks for this great tool.
>
> Second: I have a random question about manual segmentation
>
>
> Say i outline a blood vessel in 2+ non-adjacent slices of a 3d volume.  What
> software do you recommend to fill in-between my manual segmentations.  Even
> if it's just a simple spline-fit.  I've made this for myself in matlab but
> would prefer to use a better tool.
>
>
>
> _______________________________________________
> Tubetk-developers mailing list
> Tubetk-developers at tubetk.org
> http://public.kitware.com/cgi-bin/mailman/listinfo/tubetk-developers
>



-- 

==============================
Stephen R. Aylward, Ph.D.
Senior Director of Operations - North Carolina
Kitware, Inc. - North Carolina Office
http://www.kitware.com
stephen.aylward (Skype)
(919) 969-6990 x300