[Insight-users] CT Artifact Segmentation

[Luis Ibanez]

Hi Neil,

This refinement of the classification looks like a natural task for
Markov Random Fields, where you can express the expectation
of neighborhoods between the different classes in the image.

You may want to look at:
http://www.itk.org/Doxygen/html/classitk_1_1MRFImageFilter.html


     Luis


---------------------------------------------------------------------------
On Fri, May 20, 2011 at 4:47 AM, Neil Panjwani <paniwani at gmail.com> wrote:
> Hi,
> I need some help with a CT imaging artifact. I'm working on a stool
> subtraction project of the colon in which the stool is tagged near 700 HU
> (bright), air is -1025 HU (dark) and tissue is grey in the range of -300-400
> HU (depending on the proximity to stool).
> I'm trying to separate stool-air boundaries from stool-tissue-air
> boundaries. The problem arises because of the partial volume effect of CT
> imaging. At the stool-air boundary, the intensity is blurred and mimics
> tissue as the intensity goes from very negative to very positive. The same
> occurs during a stool-tissue-air boundary, however the change is less abrupt
> due to the inner tissue layer.
> I've tried using a connected component from known tissue but I erroneously
> connect stool-air as tissue in the process. I've tried intensity and
> gradient magnitude thresholds but the intensity-gradient relationship for
> the two cases are nearly identical. The gradient is very sensitive to my
> choice of sigma; values too large lose the tissue altogether and values too
> low keep too much noise. I've also tried directional gradients, moving
> outward from air (dark), but again the choice of sigma makes it extremely
> sensitive.
> I have also tried using local Haralick texture features (in the OTB
> toolbox), but haven't seen good results. I've only tried a few offset pairs.
> I try to avoid computing all offsets and averaging because it is incredibly
> slow to generate all the GLCM histograms.
> I've attached the original input and my voxel map, based on simple intensity
> and gradient thresholds, as well as zoomed in versions illustrating the
> problem. Here, orange represents tissue, black is stool, blue is air, and
> white is unclassified. As can be seen in map_zoom.png, tissue which is
> bordering both stool and air in the center of the image is left
> unclassified.
> How can I distinguish these two boundaries in order to recover the
> unclassified tissue?
> _____________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Kitware offers ITK Training Courses, for more information visit:
> http://www.kitware.com/products/protraining.html
>
> Please keep messages on-topic and check the ITK FAQ at:
> http://www.itk.org/Wiki/ITK_FAQ
>
> Follow this link to subscribe/unsubscribe:
> http://www.itk.org/mailman/listinfo/insight-users
>
>