[Insight-users] MultiScaleHessianBasedMeasureImageFilter + HessianToObjectnessMeasureImageFilter, empty image
Sergio Vera
sergio.vera at alma3d.com
Mon Oct 11 06:22:02 EDT 2010
Hello all
Currently I use HessianRecursiveGaussianImageFilter Hessian3DToVesselness to
enhance coronary arteries in CTA images. I would like to test the multiscale
hessian enhancement using MultiScaleHessianBasedMeasureImageFilter and
HessianToObjectnessMeasureImageFilter, however, I'm only obtainig empty
(black) images as a result of the enhancement and the scales.
Here is a snippet of the code:
> typedef itk::ThresholdImageFilter<ITKImgInType> ThrFilter;
ThrFilter::Pointer thr = ThrFilter::New();
thr->SetInput(&a_imgIn);
thr->SetLower(80);
thr->SetUpper(10000); // remove other parts of the image with low values
> typedef itk::CastImageFilter<ITKImgInType, F1d3ITKImgType> CastFilter;
CastFilter::Pointer cast2float = CastFilter::New();
cast2float->SetInput(thr->GetOutput());
> typedef itk::SymmetricSecondRankTensor<float, 3> HessianPixelType;
typedef itk::Image<HessianPixelType, 3> HessianImage;
> typedef itk::MultiScaleHessianBasedMeasureImageFilter <F1d3ITKImgType,
> HessianImage, F1d3ITKImgType> HessianFilter;
HessianFilter::Pointer hessian = HessianFilter::New();
hessian->SetInput(cast2float->GetOutput());
hessian->SetNumberOfSigmaSteps(6);
hessian->SetSigmaStepMethodToEquispaced();
hessian->SetSigmaMinimum(1.1);
hessian->SetSigmaMinimum(2.4); // This sizes in mm?
hessian->SetGenerateHessianOutput(true);
hessian->SetGenerateScalesOutput(true);
> typedef itk::HessianToObjectnessMeasureImageFilter <HessianImage,
> F1d3ITKImgType> ObjectnessFilter;
ObjectnessFilter::Pointer vesselFilter = ObjectnessFilter::New();
vesselFilter->SetScaleObjectnessMeasure(false);
vesselFilter->SetBrightObject(true);
vesselFilter->SetAlpha(0.5);
vesselFilter->SetBeta(0.5);
vesselFilter->SetGamma(5.0);
vesselFilter->SetObjectDimension(1); // vessels
hessian->SetHessianToMeasureFilter(vesselFilter);
try {
hessian->Update();
} catch (const itk::ExceptionObject& e) {
std::cerr << __FUNCTION__ << ": itk::ExceptionObject caught: " <<
> e.what() << '\n';
return eErrorITKExceptionRunningSegmentation;
} catch (...) {
std::cerr << __FUNCTION__ << ": '...' exception caught.\n";
return eErrorUnknownExceptionStartingSegmentation;
}
spy::WriteFile<F1d3ITKImgType>(__FUNCTION__, "Vesselness.mhd",
> *(hessian->GetOutput()), true);
spy::WriteFile<F1d3ITKImgType>(__FUNCTION__, "Scales.mhd",
> *(hessian->GetScalesOutput()), true);
spy::WriteFile<F1d3ITKImgType>(__FUNCTION__, "Converted2Float.mhd",
> *(cast2float->GetOutput()), true);
The two first images written to disk are empty. Correct size and spacing but
only black voxels. Converted2Float.mhd seems to be correcttly converted
Thanks in advance
--
Sergio Vera
Alma IT Systems
C/ Vilana, 4B, 4º 1ª
08022 Barcelona
T. (+34) 932 380 592
www.alma3d.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-users/attachments/20101011/2e1efdc3/attachment-0001.htm>
More information about the Insight-users
mailing list