[ITK-users] [ITK-Users] Multiple Inputs of different types

Matt McCormick matt.mccormick at kitware.com
Wed Jan 21 15:11:02 EST 2015


Hi Michael,

The ImageToImageFilter will still need to be templated over
TInputImage and TOutputImage.

For example filters that have multiple inputs to study, see, e.g.,

ConvolutionImageFilterBase [1]  (has a KernalImage input)
MaskedFFTNormalizedCorrelationImageFilter [2] (has MaskImage input)
HistogramThresholdImageFilter [3] (has MaskImage input)

HTH,
Matt

[1] http://itk.org/Doxygen/html/classitk_1_1ConvolutionImageFilterBase.html

[2]  http://itk.org/Doxygen/html/classitk_1_1MaskedFFTNormalizedCorrelationImageFilter.html#a238b7f7d871a81849e66a12e0564825a

[3] http://itk.org/Doxygen/html/classitk_1_1HistogramThresholdImageFilter.html#a19a692e8043422ba6721815f94f04944

On Wed, Jan 21, 2015 at 2:44 PM, Michael Pinkert <mpinkert at wisc.edu> wrote:
> Hi all,
>
> I'm trying to develop a Vector Connected Component Filter/pipeline that
> checks the direction of the eigenvector as well as whether or not the pixels
> are touching.  I am doing this by modifying the three following filters that
> build off of each other.
>
> itkSymmetricEigenAnalysisImageFilter.h so that it returns the largest
> eigenvector rather than the eigenvalues
>
> itkBinaryImageToLabelMapFilter.h to check that the eigenvectors for each
> pixel are aligned after checking that they are touching
>
> itkBinaryShapeKeepNObjectsImageFilter.h to simply pass the hessian into the
> LabelMapFilter.
>
> The problem is that I need to include a image that contains the hessian for
> each pixel along with the normal input image, so that the KeepNObjects
> filter can send it down to the LabelMapFilter and then to the
> EigenAnalysisFilter.  However, none of the methods I have tried to get the
> hessian image into the filters work, and I do not understand how to allow
> for this second input.
>
> So far I have tried modifying the template as follows
>
> class BinaryVectorConnectedShapeKeepNObjectsImageFilter:
>   public ImageToImageFilter< TInputImage, THessianImage, TInputImage >
>
> Combined with using the SetInputX functions as seen in other filters as
> follows
>
>   void SetInput1(const TInputImage *image1)
>   {
>           this->SetNthInput(0, const_cast <TInputImage * >( image1));
>   }
>
> This produces an error of "Too many template arguments".
>
> I have tried to make the hessian image a member variable and to use the
> normal Get/Set macros, but that also produces errors of being unable to
> convert between types when the function is used in main.
>
> I have also tried to simply define a new function to set the Hessian, like
> such
>
>   void SetHessian(HessianImageType::Pointer hessianImage)
>   {
>      m_HessianImage = hessianImage;
>   }
>
> And alternate versions with different input arguments, which include "const
> HessianImage *hessianImage".  These also run into errors where it won't
> covert between types when I use the function in main.
>
> I have looked around at different filters that have multiple inputs, and all
> of them seem to be based on Functors/the BinaryFunctorImageFilter, but I am
> very unsure of how I would go about modifying the filters, and have very
> little experience with Functors.  Also, the arguments for the other filters
> based on Functors all seem to include an operator as a fourth argument, and
> I do not want to perform any operations on the Hessian until the
> SymmetricAnalysis filter.
>
> Does anyone know how I could set this second input simply, or could point me
> to an example that does similar?
>
> Best regards,
> Michael Pinkert
>
>
>
>
>
>
> --
> View this message in context: http://itk-insight-users.2283740.n2.nabble.com/ITK-Users-Multiple-Inputs-of-different-types-tp7586810.html
> Sent from the ITK Insight Users mailing list archive at Nabble.com.
> _____________________________________
> 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.php
>
> 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://public.kitware.com/mailman/listinfo/insight-users


More information about the Insight-users mailing list