Unable to compile with VS2010 Win64

[heffter]

Unable to compile BRAINSTools HEAD with VS2010 Win64:

• teem branch external project does not compile, however latest teem works:
  1. in src\gage\stackBlur.c bzero is gcc only, memset is platform independent
  2. in src\meet\meetPull.c line 169 should move to 152

• ANTs library does not compile:

  1. in Examples\ImageMath.cxx nesting of code blocks exceeds the limit of 128 nesting levels, changing "else if" to if in line 14881 solves the issue.

  2. in ExternalApplications\MALF\CorrectiveLearning\segAdapter.cxx from line 501 should be modified to

     NeighborhoodIteratorType::OffsetType offset; offset[0] = j; offset[1] = k; offset[2] = l; X[C++] = autosegnit.GetPixel( offset ); for( size_t it = 0; it < ims.size(); it++ ) { X[C++] = imnits[it].GetPixel( offset ); }

  3. in ExternalApplications\MALF\JointFusion\LabelFusion.cxx line 330 should be

     int* tmap = new int[MaxV+1]; and line 342 delete[] tmap;

  4. in ExternalApplications\MALF\JointFusion\WeightedVotingLabelFusionImageFilter.txx from line 319:

     InputImagePixelType bestMatchSum = new InputImagePixelType[m_Modality]; InputImagePixelType bestMatchSSQ = new InputImagePixelType[m_Modality]; InputImagePixelType bestMatchMean = new InputImagePixelType[m_Modality]; InputImagePixelType bestMatchVar = new InputImagePixelType[m_Modality]; InputImagePixelType bestMatchSD = new InputImagePixelType[m_Modality]; InputImagePixelType MatchSum = new InputImagePixelType[m_Modality]; InputImagePixelType *MatchSSQ = new InputImagePixelType[m_Modality]; and from line 379: delete[] bestMatchSum; delete[] bestMatchSSQ; delete[] bestMatchMean; delete[] bestMatchVar; delete[] bestMatchSD; delete[] MatchSum; delete[] MatchSSQ;

  5. in Utilities\itkGeometricJacobianDeterminantImageFilter.hxx line 184, 187 and 190 vcl_sqrt( 3.0 ) should be used.

With these modification I was able to compile dependent libraries, however I still have a few problems I cannot figure out in BRAINSTools BRAINSCommonLib project:

1. Error 1 error C2244: 'ants::RegistrationHelper<TComputeType,VImageDimension>::SetFixedImageMask' : unable to match function definition to an existing declaration d:\devel\brainstools-x64\ants\examples\itkantsRegistrationHelper.hxx 827
2. Error 2 error C2244: 'ants::RegistrationHelper<TComputeType,VImageDimension>::SetMovingImageMask' : unable to match function definition to an existing declaration d:\devel\brainstools-x64\ants\examples\itkantsRegistrationHelper.hxx 838

[hjmjohnson]

@Chaircrusher Kent Please attempt to resolve these issues ASAP for BRAINSTools, ANTS, and Teem. If you need me to merge items into ANTS, please let me know.

I know that you can not build on windows, but try to resolve these windows based failures using the notes above on your mac/linux computers. Then request that @heffter rebuild.

Thanks, Hans

[heffter]

You can download VirtualBox (www.virtualbox.org) image from Microsoft for free from here: https://www.modern.ie/en-us/virtualization-tools You can also download VS2010 SP1 and use it for 30 days. Hope it helps.

[Chaircrusher]

I made the changes and they appear to compile. The problem I ran into fixing this is that in trying to compile, the current version of ANTs and ITK clashes with the ANTSRegistrationHelper.

The changes for ANTs: diff --git a/Examples/ImageMath.cxx b/Examples/ImageMath.cxx index c772e68..6bb773e 100644 --- a/Examples/ImageMath.cxx +++ b/Examples/ImageMath.cxx @@ -14877,7 +14877,7 @@ private: { MajorityVoting<3>(argc, argv); }

• else if( strcmp(operation.c_str(), "MostLikely") == 0 )
• if( strcmp(operation.c_str(), "MostLikely") == 0 ) { MostLikely<3>(argc, argv); } diff --git a/ExternalApplications/MALF/CorrectiveLearning/segAdapter.cxx b/ExternalApplications/MALF/CorrectiveLearning/segAdapter.cxx index 8f6828e..887e2b3 100644 --- a/ExternalApplications/MALF/CorrectiveLearning/segAdapter.cxx +++ b/ExternalApplications/MALF/CorrectiveLearning/segAdapter.cxx @@ -498,10 +498,14 @@ int main( int argc, char * * argv ) { for( l = -DZ; l < DZ + 1; l++ ) {
• X[C++] = autosegnit.GetPixel( (NeighborhoodIteratorType::OffsetType) {{j, k, l}});
• NeighborhoodIteratorType::OffsetType offset;
• offset[0] = j;
• offset[1] = k;
• offset[2] = l;
• X[C++] = autosegnit.GetPixel(offset); for( size_t it = 0; it < ims.size(); it++ ) {
• X[C++] = imnits[it].GetPixel( (NeighborhoodIteratorType::OffsetType) {{j, k, l}});
• X[C++] = imnits[it].GetPixel(offset); } } } diff --git a/ExternalApplications/MALF/JointFusion/LabelFusion.cxx b/ExternalApplications/MALF/JointFusion/LabelFusion.cxx index 86ae1fd..02df215 100644 --- a/ExternalApplications/MALF/JointFusion/LabelFusion.cxx +++ b/ExternalApplications/MALF/JointFusion/LabelFusion.cxx @@ -327,7 +327,7 @@ int lfapp(int argc, char *argv[]) cout<<"MaxV: "<<MaxV<<endl; if ((int)GroupSet.size()!=MaxV+1) {
• int tmap[MaxV+1];
• int tmap = new int[MaxV+1]; int tc=-1; for (set::iterator it = GroupSet.begin(); it != GroupSet.end(); ++it) { @@ -339,7 +339,7 @@ int lfapp(int argc, char argv[]) p.AtlasGroupID[i] = tmap[p.AtlasGroupID[i]]; cout<<p.AtlasGroupID[i]<<" : "; } -// delete tmap;
• delete [] tmap; } cout<<999<<endl; if (p.GroupWeight.size()!=GroupSet.size()) diff --git a/ExternalApplications/MALF/JointFusion/WeightedVotingLabelFusionImageFilter.txx b/ExternalApplications/MALF/JointFusion/WeightedVotingLabelFusionImageFilter.txx index 62e4cee..74720a7 100644 --- a/ExternalApplications/MALF/JointFusion/WeightedVotingLabelFusionImageFilter.txx +++ b/ExternalApplications/MALF/JointFusion/WeightedVotingLabelFusionImageFilter.txx @@ -4,7 +4,7 @@ Module: $Id: WeightedVotingLabelFusionImageFilter.txx,v 1.1 2012/09/05 14:39:51 zhuzhu771 Exp $ Language: C++ program Copyright (c) 2012 Paul A. Yushkevich, University of Pennsylvania

• + This file is part of ASHS

  ASHS is free software: you can redistribute it and/or modify @@ -15,22 +15,22 @@ This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the

• GNU General Public License for more details.
• GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program. If not, see http://www.gnu.org/licenses/.

- ===================================================================

• =================================================================== + CITATION:

  This program implements the method described in the paper

• H. Wang, J. W. Suh, S. Das, J. Pluta, C. Craige, P. Yushkevich,
• "Multi-atlas segmentation with joint label fusion," IEEE Trans.
• H. Wang, J. W. Suh, S. Das, J. Pluta, C. Craige, P. Yushkevich,

• "Multi-atlas segmentation with joint label fusion," IEEE Trans. on Pattern Analysis and Machine Intelligence, 35(3), 611-623, 2013

  =================================================================== */

• +

  include

  include

  include <vnl/vnl_matrix.h>

  @@ -102,9 +102,9 @@ template<class TInputImage, class TOutputImage> void WeightedVotingLabelFusionImageFilter<TInputImage, TOutputImage> ::ComputeOffsetTable(

• const InputImageType *image,
• const SizeType &radius,
• int **offset,
• const InputImageType *image,
• const SizeType &radius,
• int _offset, size_t &nPatch, int _manhattan) { @@ -118,7 +118,7 @@ WeightedVotingLabelFusionImageFilter<TInputImage, TOutputImage> iCenter[i] = r.GetIndex(i) + r.GetSize(i)/2; itTempPatch.SetLocation(iCenter);
• // Compute the offsets
• // Compute the offsets nPatch = itTempPatch.Size(); (*offset) = new int[nPatch]; if(manhattan) @@ -163,9 +163,9 @@ WeightedVotingLabelFusionImageFilter<TInputImage, TOutputImage> // Construct offset tables for all the images (these can be different because they // depend on the buffered region) size_t nPatch, nSearch;
• int *offPatchTarget,
• int _offPatchTarget, _offPatchAtlas = new int [n],
• *offPatchSeg = new int [n],
• _offPatchSeg = new int [n], _offSearchAtlas = new int [n], *manhattan;

@@ -209,7 +209,7 @@ WeightedVotingLabelFusionImageFilter<TInputImage, TOutputImage> m_PosteriorMap[*sit]->FillBuffer(0.0f); }

• std::cout<<m_RetainVotingWeight<<endl;
• std::cout<<m_RetainVotingWeight<<endl; if (m_RetainVotingWeight) { m_VotingWeight.clear(); @@ -220,7 +220,7 @@ WeightedVotingLabelFusionImageFilter<TInputImage, TOutputImage> m_VotingWeight[i]->SetRequestedRegion(this->GetOutput()->GetRequestedRegion()); m_VotingWeight[i]->SetBufferedRegion(this->GetOutput()->GetRequestedRegion()); m_VotingWeight[i]->Allocate();
• m_VotingWeight[i]->FillBuffer(0.0f);

• m_VotingWeight[i]->FillBuffer(0.0f);

  } } PosteriorImagePtr countermap = PosteriorImage::New(); @@ -252,7 +252,7 @@ WeightedVotingLabelFusionImageFilter<TInputImage, TOutputImage> apd[i] = new InputImagePixelType[nPatch * m_Modality];

  // Also an array of pointers to the segmentations of different atlases

• const InputImagePixelType *patchSeg = new const InputImagePixelType[n];

• const InputImagePixelType *patchSeg = new const InputImagePixelType[n];

  // Create an array for storing the normalized target patch to save more time InputImagePixelType _xNormTargetPatch = new InputImagePixelType[nPatch * mModality]; @@ -287,7 +287,7 @@ WeightedVotingLabelFusionImageFilter<TInputImage, TOutputImage> for(unsigned int i = 0; i < nPatch; i++) xNormTargetPatch[toff + i] = ((pTargetCurrent + offPatchTarget[i]) - mu) / sigma; }

• + // In each atlas, search for a patch that matches our patch for(int i = 0; i < n; i++) { @@ -296,7 +296,7 @@ WeightedVotingLabelFusionImageFilter<TInputImage, TOutputImage> int offPatch = offPatchAtlas[i], offSearch = offSearchAtlas[i];

  const InputImageType *tatlas = m_Atlases[toff];

• + // Get the requested region for the tatlas RegionType rr = tatlas->GetRequestedRegion();

@@ -316,8 +316,13 @@ WeightedVotingLabelFusionImageFilter<TInputImage, TOutputImage> const InputImagePixelType _bestMatchPtr = new const InputImagePixelType [m_Modality]; const InputImagePixelType _tMatchPtr = new const InputImagePixelType [m_Modality];

• InputImagePixelType bestMatchSum[m_Modality], bestMatchSSQ[m_Modality],bestMatchMean[m_Modality],bestMatchVar[m_Modality],bestMatchSD[m_Modality];
• InputImagePixelType MatchSum[m_Modality], MatchSSQ[m_Modality];
• InputImagePixelType *bestMatchSum = new InputImagePixelType[m_Modality];
• InputImagePixelType *bestMatchSSQ = new InputImagePixelType[m_Modality];
• InputImagePixelType *bestMatchMean = new InputImagePixelType[m_Modality];
• InputImagePixelType *bestMatchVar = new InputImagePixelType[m_Modality];
• InputImagePixelType *bestMatchSD = new InputImagePixelType[m_Modality];
• InputImagePixelType *MatchSum = new InputImagePixelType[m_Modality];

• InputImagePixelType *MatchSSQ =      new InputImagePixelType[m_Modality];

  int bestK = 0; for(unsigned int k = 0; k < nSearch; k++) @@ -377,15 +382,22 @@ WeightedVotingLabelFusionImageFilter<TInputImage, TOutputImage>

  // Store the best found neighborhood patchSeg[i] = (bestMatchPtr[0] - m_Atlases[toff]->GetBufferPointer()) + seg->GetBufferPointer();

• delete [] bestMatchSum;
• delete [] bestMatchSSQ;
• delete [] bestMatchMean;
• delete [] bestMatchVar;
• delete [] bestMatchSD;
• delete [] MatchSum;

• delete [] MatchSSQ;

  }

  // Allocate Mx MatrixType Mx(n, n);

  // Now we can compute Mx

• for(int i = 0; i < n; i++)
• for(int i = 0; i < n; i++) {
• for(int k = 0; k <= i; k++)

• for(int k = 0; k <= i; k++)
  {
  // Multiply through the apd arrays
  InputImagePixelType mxval = 0.0;

  @@ -393,7 +405,7 @@ WeightedVotingLabelFusionImageFilter<TInputImage, TOutputImage> mxval += apd[i][m] * apd[k][m];

  mxval /= (nPatch - 1);

• + if(m_Beta == 2) mxval *= mxval; else @@ -498,7 +510,7 @@ WeightedVotingLabelFusionImageFilter<TInputImage, TOutputImage> // Clear posterior maps if(!m_RetainPosteriorMaps) m_PosteriorMap.clear();

• + for(OutIter it(this->GetOutput(), this->GetOutput()->GetBufferedRegion()); !it.IsAtEnd(); ++it) { IndexType idx = it.GetIndex(); @@ -517,7 +529,7 @@ WeightedVotingLabelFusionImageFilter<TInputImage, TOutputImage> { for (int i=0; i < n; i++) m_VotingWeight[i]->SetPixel(idx,m_VotingWeight[i]->GetPixel(idx)/countermap->GetPixel(idx));

• }
• } } }

@@ -544,7 +556,7 @@ WeightedVotingLabelFusionImageFilter<TInputImage, TOutputImage>

/**

• This function computes similarity between a normalized patch (normtrg) and a patch
  • * that has not been normalized (psearch). It can be shown that the sum of squared
  • * that has not been normalized (psearch). It can be shown that the sum of squared
• differences between a normalized patch u and a unnormalized patch v is equal to *
• 2 [ (n-1) - (\Sum u_i v_i ) / \sigma_v ] @@ -559,10 +571,10 @@ template <class TInputImage, class TOutputImage> double WeightedVotingLabelFusionImageFilter<TInputImage, TOutputImage> ::PatchSimilarity(
  • const InputImagePixelType *psearch,
  • const InputImagePixelType psearch, const InputImagePixelType normtrg,
  • int offset,
  • size_t n,
  • int offset,
  • size_t n, int *offsets, InputImagePixelType &sum_psearch, InputImagePixelType &ssq_psearch) diff --git a/Utilities/itkGeometricJacobianDeterminantImageFilter.hxx b/Utilities/itkGeometricJacobianDeterminantImageFilter.hxx index 70753cc..9eb9add 100755 --- a/Utilities/itkGeometricJacobianDeterminantImageFilter.hxx +++ b/Utilities/itkGeometricJacobianDeterminantImageFilter.hxx @@ -181,13 +181,13 @@ GeometricJacobianDeterminantImageFilter<TInputImage, TRealType, TOutputImage> this->m_DeltaTriangularPointB.Fill( 0.0 ); this->m_DeltaTriangularPointC.Fill( 0.0 );
• this->m_DeltaTriangularPointA[1] = 0.25 * vcl_sqrt( 3 ) * spacing[1];

• this->m_DeltaTriangularPointA[1] = 0.25 * vcl_sqrt( 3.0 ) * spacing[1];

  this->m_DeltaTriangularPointB[0] = -0.5 * spacing[0];

• this->m_DeltaTriangularPointB[1] = -0.25 * vcl_sqrt( 3 ) * spacing[1];

• this->m_DeltaTriangularPointB[1] = -0.25 * vcl_sqrt( 3.0 ) * spacing[1];

  this->m_DeltaTriangularPointC[0] = 0.5 * spacing[0];

• this->m_DeltaTriangularPointC[1] = -0.25 * vcl_sqrt( 3 ) * spacing[1];
• this->m_DeltaTriangularPointC[1] = -0.25 * vcl_sqrt( 3.0 ) * spacing[1]; }

[Chaircrusher]

OK: So I checked in a change to BRAINSTools, which makes it pull the latest ANTs and ITKv4. I fixed the things mentioned in this issue for Visual Studio 2010, but the latest ANTs was incompatible with the ITKv4 version being pulled by BRAINSTools. So I also had to move to the latest ITKv4 version to get ANTs to compile.

[heffter]

Thank you for the quick fix. Could you also update the Teem library as well?

[heffter]

Can we reopen this issue? It still has not been resolved:

• Teem library compilation failed (need to update to latest, or resolve it as above mentioned)
• BRAINSCommonLib compilation failed with ANTSRegistrationHelper
  • 'ants::RegistrationHelper<TComputeType,VImageDimension>::SetFixedImageMask' : unable to match function definition to an existing declaration d:\devel\brainstools-new-x64\ants\examples\itkantsRegistrationHelper.hxx 827
  • 'ants::RegistrationHelper<TComputeType,VImageDimension>::SetMovingImageMask' : unable to match function definition to an existing declaration d:\devel\brainstools-new-x64\ants\examples\itkantsRegistrationHelper.hxx 838

[fedorov]

Tamas, this is not exactly the answer to your concerns, but you could try building BRAINS with ANTS disabled. It is not needed for our purposes. Also, you could look at the configuration options used for BRAINS by Slicer here https://github.com/Slicer/Slicer/blob/master/SuperBuild.cmake#L210-L248. We know that this configuration has no build issues on all three platforms: http://slicer.cdash.org/index.php?project=Slicer4 (correction: looks like the factory machines are using VS2008, not VS2010 that you reported, so not guaranteed the issues are the same). Hope this helps.