[vtkusers] Canny - TCL to C++ conversion

[Sean McInerney]

Hi Brian,

   I wrote an example application for vtkFLTK 
<http://vtkfltk.sourceforge.net/Examples/Canny.html> based on Canny.tcl. 
Even if you are not interested in using vtkFLTK, the source code should 
be of help. You can browse the source from the CVS repository by 
clicking 'Canny' above the screenshot.

-Sean

Bahr, Brian wrote:
> More specifically, I am trying to write a Canny filter in C++.  I am
> using the Canny.tcl file supplied by VTK at the following address...
> 
> http://public.kitware.com/cgi-bin/viewcvs.cgi/*checkout*/Graphics/Testin
> g/Tcl/Canny.tcl?root=VTK&content-type=text/plain
> 
> It seems that most of the classes and methods used in this TCL file map
> directly to C++ VTK classes and methods.  I just can't quite get it
> right.  Here is the C++ that I am working with...
> 
> 		vtkJPEGReader* imageInReader = vtkJPEGReader::New();
> 		imageInReader->SetFileName("MyJPEG.jpg");
> 		vtkImageLuminance* il = vtkImageLuminance::New();
> 		il->SetInput(imageInReader->GetOutput());
> 		il->Update();
> 
> 		vtkImageCast* ic = vtkImageCast::New();
> 		ic->SetOutputScalarTypeToFloat();
> 		ic->SetInput(il->GetOutput());
> 		ic->Update();
> 
> 		//smooth the image
> 		vtkImageGaussianSmooth* pGaussian =
> vtkImageGaussianSmooth::New();
> 		pGaussian->SetInput(ic->GetOutput());
> 		pGaussian->SetRadiusFactors(RFx,RFy,RFz);
> 		pGaussian->SetReleaseDataFlag(1);
> 		pGaussian->Update();
> 
> 		//gradient the image
> 		vtkImageGradient* pGradient = vtkImageGradient::New();
> 		pGradient->SetInput(pGaussian->GetOutput());
> 		pGradient->SetDimensionality(2);
> 		pGradient->SetReleaseDataFlag(1);
> 		pGradient->Update();
> 
> 		vtkImageMagnitude* pMagnitude =
> vtkImageMagnitude::New();
> 		pMagnitude->SetInput(pGradient->GetOutput());
> 		pMagnitude->Update();
> 
> 		vtkImageNonMaximumSuppression* pNonMax =
> vtkImageNonMaximumSuppression::New();
> 		pNonMax->SetMagnitudeInput(pMagnitude->GetOutput());
> 		pNonMax->SetVectorInput(pGradient->GetOutput());
> 		pNonMax->SetDimensionality(2);
> 		pNonMax->Update();
> 
> 		vtkImageConstantPad* pPad = vtkImageConstantPad::New();
> 		pPad->SetInput(pGradient->GetOutput());
> 		pPad->SetOutputNumberOfScalarComponents(3);
> 		pPad->SetConstant(0);
> 		pPad->Update();
> 		
> 		vtkImageToStructuredPoints* i2sp1 =
> vtkImageToStructuredPoints::New();
> 		i2sp1->SetInput(pNonMax->GetOutput());
> 		i2sp1->SetVectorInput(pPad->GetOutput());
> 		i2sp1->Update();
> 
> 		//link edgles
> 		vtkLinkEdgels* pLink = vtkLinkEdgels::New();
> 		pLink->SetInput((vtkImageData*)i2sp1->GetOutput());
> 		pLink->SetGradientThreshold(2);
> 		pLink->Update();
> 
> 		//threshold links
> 		vtkThreshold* pThreshold = vtkThreshold::New();
> 		pThreshold->SetInput((vtkDataSet*)pLink->GetOutput());
> 		pThreshold->ThresholdByUpper(10);
> 		pThreshold->AllScalarsOff();
> 		pThreshold->Update();
> 
> 		vtkGeometryFilter* pGeomFilter =
> vtkGeometryFilter::New();
> 		pGeomFilter->SetInput((vtkDataSet
> *)pThreshold->GetOutput());
> 		pGeomFilter->Update();
> 
> 		vtkImageToStructuredPoints* i2sp =
> vtkImageToStructuredPoints::New();
> 		i2sp->SetInput(pMagnitude->GetOutput());
> 		i2sp->SetVectorInput(pPad->GetOutput());
> 		i2sp->Update();
> 
> 		//subpixel them
> 		vtkSubPixelPositionEdgels* spe =
> vtkSubPixelPositionEdgels::New();
> 		spe->SetInput((vtkPolyData *)pGeomFilter->GetOutput());
> 		spe->SetGradMaps(i2sp->GetOutput());
> 		spe->Update();
> 
> 		vtkStripper* strip = vtkStripper::New();
> 		strip->SetInput(spe->GetOutput());
> 		strip->Update();
> 
> 		vtkPolyDataMapper* dsm = vtkPolyDataMapper::New();
> 		dsm->SetInput(strip->GetOutput());
> 		//dsm->SetStartVisibilityOff();
> 		dsm->Update();
> 
> 		vtkActor* actor = vtkActor::New();
> 		actor->SetMapper(dsm);
> //[planeActor GetProperty] SetAmbient 1.0
> //[planeActor GetProperty] SetDiffuse 0.0
> 
> 		//render
> 		vtkRenderer *ren1 = vtkRenderer::New();
> 		ren1->AddActor(actor);
> 		vtkRenderWindow *renWin = vtkRenderWindow::New();
> 		renWin->OffScreenRenderingOn();
> 		renWin->AddRenderer(ren1);
> 		renWin->Render();
> 
> 		//get contents from window into image
> 		vtkWindowToImageFilter *w2if =
> vtkWindowToImageFilter::New();
> 		w2if->SetInput(renWin); 
> 		this->pImage = w2if->GetOutput();  //where pImage is
> vtkImageData*
> 
> -----Original Message-----
> From: Daniel Sidobre [mailto:daniel.sidobre at laas.fr] 
> Sent: Friday, October 08, 2004 10:34 AM
> To: Bahr, Brian
> Cc: vtkusers at vtk.org
> Subject: [FILTERED] - [vtkusers] Canny - TCL to C++ conversion -
> Bayesian Filter detected spam
> 
> 
> Vous avez ecrit (dans votre message du 8 10 2004)
>  > I am trying to create a Canny filter in C++.  I can successfully run
> the
>  > Canny.tcl example, but I am having a hard time converting the Tcl
> code
>  > into a C++ equivalent.
> 
>  Hi,
> 
>  So, I suppose you have write some Tcl code to create some Canny
>  filter that you run successfully. And you want to translate it in c++.
> 
>  As a so general question the response is : it's probably impossible
>  as Tcl has many concept that doesn't exist in c++. For example you
>  can redefine all functions, you can evaluate expressions in differents
>  stack levels, ...
> 
>  But, if from the begening of your writing, you have had in mind that
>  you want to translate to c++, the job may be pretty easy. For each
>  line add some comas and brackets and suppress some dollars.
> 
>  Daniel
> 
> _______________________________________________
> This is the private VTK discussion list. 
> Please keep messages on-topic. Check the FAQ at: <http://public.kitware.com/cgi-bin/vtkfaq>
> Follow this link to subscribe/unsubscribe:
> http://www.vtk.org/mailman/listinfo/vtkusers
>