[Insight-users] itkWeightedMeanSampleFilter is not calling GenerateData()

[Oscar Esteban]

Hi all,

I found out that itk::WeightedMeanSampleFilter< TSample > does not correctly
initializes the MeasurementVector when using a ListSample with
itk::VariableLengthVector.

I tried modifying Code/Review/Statistics/itkWeightedMeanSampleFilter.txx,
line 98, adding a new line after, with output.SetSize( measurementVectorSize
). But I get a compilation error (SetSize is not a member of...)

How should I correct this issue? (Or, how should I use this filter when
using VariableLengthVectors?)

Thanks. Regards,
Oscar Esteban


*______________________________________*
*Oscar Esteban*
PhD Student / Researcher

Biomedical Image Technologies (BIT), UPM
ETSI Telecomunicación Lab. C203, Av. Complutense s/n - E-28040 Madrid
(Spain)
+34 915 495 700 ext.4234

Signal Processing Laboratory (LTS5), EPFL-STI-IEL-LTS5
ELD 224 (Bâtiment ELD), Station 11, CH-1015 Lausanne, Switzerland



---------- Forwarded message ----------
From: Oscar Esteban <oesteban at die.upm.es>
Date: 3 October 2011 15:08
Subject: itkWeightedMeanSampleFilter is not calling GenerateData()
To: insight-users at itk.org


Hi all,

I'm trying to compute a weighted mean for each class based on samples
obtained from corresponding priors. This is the code I am using:

 for ( unsigned int i = 0; i<NumberOfClasses; i++) {
> if (i==0) continue;  // Background mean computation not yet implemented
>  // Weights array generation for this class
> typename ProbabilitySampleFilter::Pointer sampleFilter =
> ProbabilitySampleFilter::New();
>  sampleFilter->SetInput( m_Priors[i-1] );
> sampleFilter->SetMaskImage( m_MaskImage );
> sampleFilter->Update();
>  ProbabilitySampleConstPointer sample = sampleFilter->GetOutput();
> if ( sample->Size() != m_Sample->Size() )
>  itkExceptionMacro("Sample sizes of components and prior for class " <<
> i-1 << " do not match");
> typename MeanEstimatorType::WeightArrayType weights( sample->Size() );
>  weights.Fill(1.0);
> /*
> typename ProbabilitySampleType::ConstIterator it = sample->Begin();
>  typename ProbabilitySampleType::ConstIterator end = sample->End();
>
> while( it!=end ) {
>  weights[it.GetInstanceIdentifier()] = it.GetMeasurementVector()[0];
> ++it;
> }
> */
> // Weighted Mean estimation
> typename MeanEstimatorType::Pointer meanEst = MeanEstimatorType::New();
>  meanEst->SetInput(m_Sample);
> meanEst->SetWeights( weights );
> const typename MeanEstimatorType::MeasurementVectorDecoratedType * mean_out
> = meanEst->GetOutput();
>  try {
> meanEst->Update();
> } catch( itk::ExceptionObject & ex ) {
>  std::cerr << "Exception caught" << ex << std::endl;
> }
> typename MeanEstimatorType::MeasurementVectorType mean = mean_out->Get();
>  std::cout << "Estimated mean for class " << i-1 << "= ";
>
> for (unsigned int k; k< mean.GetSize(); k++) std::cout << mean[k] << ",";
>
> std::cout<<std::endl;
>           }


When debugging, I found that meanEst->Update() seems to be doing nothing.
Watching at the variable, mean has m_NumElements=0. Although, sample seems
to be OK. No exception is thrown, but I can't get the means value. I also
checked with weights.Fill(1.0) that the problem is not those weights.

Putting a breakpoint inside itkWeightedMeanSampleFilter::GenerateData() does
not stop the debugger, and it reaches end without apparently generate
output.

How could I resolve this problem?

Thank you very much in advance. Best regards,
Oscar Esteban

______________________________________
*Oscar Esteban*
PhD Student / Researcher

Biomedical Image Technologies (BIT), UPM
ETSI Telecomunicación Lab. C203, Av. Complutense s/n - E-28040 Madrid
(Spain)
+34 915 495 700 ext.4234

Signal Processing Laboratory (LTS5), EPFL-STI-IEL-LTS5
ELD 224 (Bâtiment ELD), Station 11, CH-1015 Lausanne, Switzerland
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-users/attachments/20111012/282bf421/attachment.htm>