[Insight-users] Problems in segmenting abdominal aorta from CT images

[杨帆]

Hi, all,

I' ve been struggling to solve a medical image segmentation task in the lab
for a few weeks.

The task is trying to segment the abdominal aorta from a set of CT images
and then to reconstruct the result to form the vascular structure.

I' ve tried the connected threshold region growing method given in ITK and
embedded it into the following demo:


InsightApplications\Auxiliary\vtk\itkReadITKImage3DSegmentShowVTK.cxx

The problem is it is hard to segment abdonimal aorta without the spinal
bones due to the anatomical structure.

I added seeds in the area of the vessels for each CT image. That made the 2D
segmentation results looked fine. Unfortunately, the reconstruct result was
still poor
that the spinal bones still appeared with the vascular structure and there
were lots of blobs and spikes shown in the vessel cavity.

Below are my parameters in the program:

CurvatureFlowImageFilter

numberOfIterations = 3;
timeStep = 0.0625;

ConnectedThresholdImageFilter

upperThreshold = 101;
lowerThreshold = 230;
replaceValue = 255

Can anyone tell how to overcome the effect of the spinal bones in my CT
dataset?

Thanks in advance!

Regards,

Fan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-users/attachments/20110624/15a4f4e6/attachment.htm>