[Insight-users] Combining Segmentations

[Luis Ibanez]

Hi Lindsey,


Some questions:


A) Are the different segmentations being produced
    from the same original CT data set ?
    (e.g. by changing parameters in a segmentation method ?).

    or are they produced from CTs from different subjects ?


B) About the holes that you find in your segmentation...
    Where in the vertebrae are they located ?

    CT will only contain bright values in the regions of the
    vertebrae that are highly calcified. It is common to
    obtain holes if the low calcified regions, such as the
    top of the vertebral bodies.

    Holes may also be the result of having a CT dataset with
    a large interslice spacing (e.g. 2mm will be too large for
    capturing the thin bone layer at the top and bottom of the
    vertebral bodies.  What are the pixel dimensions of your
    CT dataset.


C) Are you familiar with the normal anatomical structure of C1 ?
    (sorry to ask this, but unfortunately not everybody in
     image analysis is familiar with the anatomy of the
     structures they work with...)

    C1 is a particularly small bony structure, and has two normal
    holes in the junction of the transverse process, those are
    called the transverse foramen.


    Maybe it will be helpful if you post a rendering of your
    segmentations where the holes are visible. (and point them
    out with some annotation).



Combining several segmentations could be done with the OR, AND
or MAX image filters, assuming that all the segmentations are
in the same reference system (e.g. taken from the same original
CT scan).  Otherwise you will have to solve first the registration
problem, and that in itself may result more challenging than just
improving your segmentation.


There are several techniques that you could use in order to fill
holes. The approach to use depend largely on the size of the holes
(both in diameter and depth) as well as the curvature of the structures
around the holes.

What segmentation technique did you used in order to get the current
segmentation with holes ?



   Please let us know,


      Thanks



         Luis




------------------------
Lindsey McAninch wrote:

> Hi all, I have a bunch of CT scans which I have made into volumes of the 
> cervical spine.  I am Segmenting out the C1 vertebrae and nothing else.  
> The problem is I am getting holes in my segmentation.  I was wondering 
> if there was a way to take several different segmentations and combine 
> the images (all of the same scan)  into one image to hopefully help to 
> fill the holes I get with each individual segmentation.  If anyone knows 
> how to do this your help is very much appreciated.  Thanks in advance!
> 
> Lindsey McAninch
> University of Utah
> SCI Institute
> 50 S. Central Campus Dr. Salt Lake City, UT 84112
> 
> lindsey at sci.utah.edu
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users
>