Dear all,<br><br>I am recently perform a vessel segmentation on a series of CT images by the example code of itkFastMarchingImageFilter.<br><br>And there is a curious problem. Below is the original CT images and I want to segment the abdominal aorta(AA) from this image.<br>
<br><div style="margin-left: 240px;"><img title="1.2.840.113704.1.111.976.1298943214.4165-Frame1.jpg" alt="1.2.840.113704.1.111.976.1298943214.4165-Frame1.jpg" src="cid:ii_1319284607951e09" height="200" width="200"><br></div>
<br>When I run the itkFastMarchingImageFilter demo on it, the result seems good.<br>
<br><div style="margin-left: 240px;"><img title="1.2.840.113704.1.111.976.1298943214.4165_fm17-Frame1.jpg" alt="1.2.840.113704.1.111.976.1298943214.4165_fm17-Frame1.jpg" src="cid:ii_1319285b3f948cee" height="200" width="200"><br>
</div><br>The main parameters specified in this demo are:<br>
<br><ul><li>itkCurvatureAnisotropicDiffusionImageFilter</li></ul> numberOfIterations = 5;<br> timeStep = 0.125;<br> conductanceParameter = 0.1;<br><ul><li>itkGradientMagnitudeRecursiveGaussianImageFilter</li>
</ul> sigma = 0.98;<br><ul><li>itkSigmoidImageFilter</li></ul> outputMinimum = 0;<br> outputMaximum = 1;<br> alpha = 0.5;<br> beta = 21.5;<br><ul><li>itkFastMarchingImageFilter</li>
</ul> stoppingTime = 30;<br><ul><li>itkBinaryThresholdImageFilter</li></ul> lowerThreshold = 0;<br> upperThreshold = 35;<br> outsideValue = 0;<br> insideValue = 255;<br><br>The trial point are located at some point on the edge of the AA wall.<br>
<br>Then I modified this demo to enable 3D reconstruction segmentation
result of the whole CT dataset using vtkContourFilter and I got a poor result: there are pieces of snowflake-like object locating at the whole render window that I can't tell where is the segmented AA.<br><br>Anyone can explain this or can give me some advice?<br>
<br>Thanks in advance.<br>
<br><br><br>Regards, <br><font color="#888888"><br>Fan</font><br><br><br>