[Insight-users] ConfidenceConnectedImageFilter question

[Miller, James V (Research)]

Bill Lorensen recently added a filter to ITK call
itkOtsuThresholdImageFilter.

It uses the OtsuThresholdCalculator and a BinaryThresholdImageFilter
internally.
When applied to CT images of the thorax, it does a pretty good job of
picking a 
threshold automatically that separates the lung field from the body. But it
also ends up labeling the region outside the body and outside the field of
view with
the same label as the lung field.  You can run
ConnectedComponentsImageFilter to give the lungs a different label than the
outside of the body, then cull away any region that touches the boundary of
the image. (Or use ConnectedThresholdImageFilter to pull out the lungs from
the label map or remove the corners of the image).

Jim




-----Original Message-----
From: Jacob Boomgaarden [mailto:boomgaar at arsc.edu]
Sent: Friday, July 30, 2004 5:18 PM
To: insight-users at itk.org
Subject: [Insight-users] ConfidenceConnectedImageFilter question


Hello All,

I am working with a simple segmentation problem of extracting the lungs from
a CT
image file. So far I have been using one of the region growing algorithms to
define
the segmented area (ConfidenceConnectedImageFilter) but am curious about a
question
that arose.

Q) To extract both the left and right lung segments, the region growing
algorithm
must overcome other anatomical structures. What is the best method for doing
this? I
am under the impression that the ConfidenceConnectedImageFilter does not
allow for
multiple seeds to be issued, so what other ideas can you throw at me to try!

Thanks for the help,



-- 
Jacob Boomgaarden
ARSC Intern
West Ridge Research Building, Room 011
Phone:  (907)-455-3350
Office: (907)-450-8701
Email: boomgaar at arsc.edu


_______________________________________________
Insight-users mailing list
Insight-users at itk.org
http://www.itk.org/mailman/listinfo/insight-users