[Insight-users] Segmenting Visible Human Data

[Luis Ibanez]

Hi Stefan,

4Gb is certainly a challenging image size.

For watershed, as well as for many other
segmentation algorithms you will have to
cut the image in pieces, process each piece
and put them together at the end.

Note that the native ITK support for streaming
may not be of great help here since it is
restricted to operations with a well defined
bounded support for every pixel computation.


Depending on how homogeneous is the anatomical
structure that you are trying to segment, you
could apply the region growin methods. In particular
you could do things like use the
ConfidenceConnected image filter for segmenting
the anatomy in one of the image pieces. Then
use the intensity range values computed by
ConfidenceConnected as parameters for the
ThresholdConnected filter, and apply this
filter in all the other pieces (with appropriate
seed points).

As usual, it is to expect that regions growing
methods will leak sonner or later. 4Gb bring
a lot of opportunities for finding a pixel
where the region can leak.  The fragility of
this methods is that *one bad* pixel is enough
for ruining the party.

Another good method to attempt is the Threshold
LevelSetSegmentation, which is the equivalent
of the ThresholdConnected region growing but
with the nice additional capabilities of the
level sets for preventing leaks by controlling
the scaling parameters of the speed function.

This will still require to cut your full image
in pieces and process each piece individually.
If you use level sets, you could post process
the joint regions between your contiguous pieces
in order to smooth out the potential differences
between the level set obtained in one region and
the level set obtained in the facing region.


One approach that you could considered is the
model based segmentation which actually is a way
of applying model-image registration for the sake
of segmentation.  The feasibility of this approach
in your case depend a lot on the shape complexity of
the anatomical structure that you seek to segment.



Regards,


  Luis

---------------------
Stefan Lindenau wrote:
> Hi,
> 
> I have to do a segmentation on a part of the female Dataset of the 
> Visible Human. I think this part will be as big as 4GB.
> After reading documentation and former Mailinglist postings I came to 
> the conclusion that the Watershed Algorithm in conjunction with the 
> SegmentationEditorFltkGUI would be an appropriate approach.
> 
> The goal is to use the segmented data in the VTK to render for example 
> the ribs of the Visible Woman. The SegmentationEditor provides a nice UI 
> to choose the appropriate elements in the segmented image and to 
> manipulate them.
> 
> My concern lies on the memory consumption of the processing. Is it 
> realistic to use this approach with a consumer computer? Maybe another 
> algorithm would fit better to my needs.
> 
> Thank you
> Stefan
> 
> 
> 
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk . org
> http://www . itk . org/mailman/listinfo/insight-users
>