[ITK-users] cannot acces virtual void itk::Rigid3DTransform< TScalar >::SetMatrix (const MatrixType & matrix, double tolerance ) from the Rigid3D transform class.

[Matt McCormick]

Hi Yago,

The tolerance parameter was added in a very recent version of ITK.
Make sure you are using the latest version of ITK.

HTH,
Matt

On Fri, Feb 6, 2015 at 2:32 PM, Yago Diez <yagodiezdonoso at gmail.com> wrote:
>
>
> Hi all,
>
> I want to store a 3D rotation matrix that I have computed in some format
> that itk can understand (and use to transform points and images, not to as
> the starting point of any optimization process). The problem is that the
> matrix is not "as orthogonal as it should be" due to numerical problems:
>
>
> So far I am using the Rigid3DTransform which looks the best way to go (but
> am willing to change if this is wrong). The numerical problems express like
> this:
>
> 1) if I use the "MatrixIsOrthogonal(,)" function I get a "yes" for a 1e-5
> tolerance but not for the default "1e-10" tolerance.
> 2) Consequently, if I use the SetMatrix() function I get an exception.
>
> I am pretty confident that I can get away with what I need to do if I
> convince the class to just swallow the "not very orthogonal" matrix (by
> using a less restrictive tolerance). I have seen that there exists a
> function to do just that but I am having problems using it.
>
> Specifically, I cannot seem to access the
>
> itk::Rigid3DTransform< TScalar >::SetMatrix     (const MatrixType &
> matrix, double      tolerance      )      function
>
> From the Rigid3D transform class.
>
> Consequently, the following code returns a compilation error :
>
>> error: no matching function for call to
>> ‘itk::Rigid3DTransform<double>::SetMatrix(smallMatrixType&, double)’
>> XXXXXXXXXXXX: note: candidate is:
>> PATH/Code/Common/itkRigid3DTransform.txx:93:1: note: void
>> itk::Rigid3DTransform<TScalarType>::SetMatrix(const MatrixType&) [with
>> TScalarType = double, itk::Rigid3DTransform<TScalarType>::MatrixType =
>> itk::Matrix<double, 3u, 3u>]
>> PATH/Code/Common/itkRigid3DTransform.txx:93:1: note:   candidate expects 1
>> argument, 2 provided
>
>
> I have checked that the itkRigid3DTransform.txx file has indeed only the
> definition of the setmatrix method with one parameter.
>
> I assume I am looking for the setmatrix function with two parameters "in the
> wrong place", but have not yet managed to find the right one, any ideas?
>
>
> Here is the detailed code
>
>
>> typedef itk::Matrix<double, 3, 3> smallMatrixType;
>> smallMatrixType auxMatrix;
>>
> // matrix initialization, matrix M1 is orthogonal but I have numerical
> problems
>
> auxMatrix[0][0]=M1[0][0];
> auxMatrix[0][1]=M1[0][1];
> auxMatrix[0][2]=M1[0][2];
> auxMatrix[1][0]=M1[1][0];
> auxMatrix[1][1]=M1[1][1];
> auxMatrix[1][2]=M1[1][2];
> auxMatrix[2][0]=M1[2][0];
> auxMatrix[2][1]=M1[2][1];
> auxMatrix[2][2]=M1[2][2];
>
>
> typedef itk::Rigid3DTransform<double>  TransformType;
> TransformType::Pointer  rotation = TransformType::New();
>
> // rotation->MatrixIsOrthogonal(auxMatrix,1e-10) outputs 0, but
> rotation->MatrixIsOrthogonal(auxMatrix,1e-5) outputs 1
>
> rotation->SetMatrix(auxMatrix,1e-5); // this produces a compilation error
>
>
> //rotation->SetMatrix(auxMatrix); // THIS RESULTS IN AN EXCEPTION:
> itk::ERROR: Rigid3DTransform(0x1dec800): Attempting to set a non-orthogonal
> rotation matrix
>
>
> Thanks you very much for any help that you can provide.
>
>
>
> Yago Diez Donoso (PhD)
> Computer Vision and Robotics Group http://vicorob.udg.es/
> E-mail: yago at eia.udg.es; yagodiezdonoso at gmail.com
> Phone: (int. code) 34 972418013
> University of Girona
>
> _____________________________________
> 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
>