From sean at rogue-research.com  Mon Oct  2 11:56:14 2017
From: sean at rogue-research.com (Sean McBride)
Date: Mon, 2 Oct 2017 11:56:14 -0400
Subject: [ITK-dev] [ITK]  Patches for 4.12.2 release
In-Reply-To: <1C12646C-C41C-4B67-9692-8FA6295E27AA@mail.nih.gov>
References: <CALzTN-QZJctXB69bte8GZk=pdaW_1rpbwzmuy3dMa4wHHfoHmQ@mail.gmail.com>
 <20170929153939.2106354230@mail.rogue-research.com>
 <1C12646C-C41C-4B67-9692-8FA6295E27AA@mail.nih.gov>
Message-ID: <20171002155614.338261949@mail.rogue-research.com>

On Fri, 29 Sep 2017 17:28:36 +0000, Lowekamp, Bradley (NIH/NLM/LHC) [C] said:

>Could you explicitly include nigti1_io.h in you code?

Brad,

<facepalm>.  Thanks.  I had mistakenly assumed that the symbols were made private and the header was no longer installed.  I see now that it's merely no longer #included implicitly.

Sorry for the noise.

Cheers,

Sean



From matt.mccormick at kitware.com  Thu Oct  5 19:19:15 2017
From: matt.mccormick at kitware.com (Matt McCormick)
Date: Thu, 5 Oct 2017 19:19:15 -0400
Subject: [ITK-dev] Proper ThirdParty (GDCM specifically) update
	procedure?
In-Reply-To: <CALzTN-TLrmVXb_o+V19AFQuO9oy5+Pi61cPw1+r5Z7cWLbhjiQ@mail.gmail.com>
References: <20170925174021.217863588@mail.rogue-research.com>
 <CALzTN-TLrmVXb_o+V19AFQuO9oy5+Pi61cPw1+r5Z7cWLbhjiQ@mail.gmail.com>
Message-ID: <CALzTN-TB8uwKz=D4X+-AJcebX6-pS-df=1Z2n7Ye-SWiFbB7Vg@mail.gmail.com>

Jon has created an UpdatingThirdParty.md doc:

  http://review.source.kitware.com/#/c/22670/

On Mon, Sep 25, 2017 at 4:32 PM, Matt McCormick
<matt.mccormick at kitware.com> wrote:
> Hi Sean,
>
> Thanks for helping to update these third party libraries.
>
>> I'm going to try to update ITK's GDCM... are there docs for this procedure?
>
> Essentially, run the UpdateFromUpstream.sh script, but we should add
> some instructions in comments at the top of that file to address your
> questions.
>
>
>> I know about the ITK/Modules/ThirdParty/GDCM/UpdateFromUpstream.sh script, but I'm wondering things like:
>>  - do I run it on the master branch, or should I first create a branch of my own?
>
> Yes, a local branch can be created first, like GDCMUpdate.
>
>
>>  - running the script creates a branch called 'upstream-gdcm'. After the script runs, there are conflicts.  After I just resolve them manually, which branch do I commit to?
>
> After the merge conflicts are resolved, run
>
>   git add file/with/conflict
>   git commit
>   git gerrit-push
>
>
> Thanks,
> Matt

From nasil122002 at yahoo.de  Sun Oct  8 14:53:20 2017
From: nasil122002 at yahoo.de (Yusuf OEZBEK)
Date: Sun, 8 Oct 2017 18:53:20 +0000 (UTC)
Subject: [ITK-dev] itk::Fem::Solver update problem
References: <1535649976.6328478.1507488800156.ref@mail.yahoo.com>
Message-ID: <1535649976.6328478.1507488800156@mail.yahoo.com>

Dear ITK Users,
According the mail?https://public.kitware.com/pipermail/insight-users/2007-November/024272.html?, I have converted my vtkUnstructuredGrid, wich I extracted from a vtkPolyData and contains TRIANGLE_CELL's to the itkMesh and than from itkMesh to itk::Fem::Object to deform my mesh:My problem is: If I call "m_FemSolver->Update();" ,I get a segmentation fault, which I really not understand.?

Program received signal SIGSEGV, Segmentation fault.
0x00007fffb1086dd5 in itk::fem::FEMObject<2u>::DeepCopy (this=0x4cc5ee0,?
??? Copy=0x4cc1780)
??? at /home/toolkits/itk-4.7.2/Modules/Numerics/FEM/include/itkFEMObject.hxx:157
157??? ??? a->UnRegister();


void itkMeshToFemMesh::updateItkMeshSlot(itk::Mesh<double, 3, MeshTraitsType> *itkMesh)
{
? typedef itk::fem::FEMObject<3> FemObjectType;
? FemObjectType::Pointer m_FemObject= FemObjectType::New();

? typedef itk::fem::Solver<3> SolverType;
? typedef SolverType::Pointer SolverPointerType;
? SolverPointerType m_FemSolver= SolverType::New();

? // Define the Element Material properties
? typedef itk::fem::MaterialLinearElasticity MaterialType;
? MaterialType::Pointer material= MaterialType::New();
? material->SetGlobalNumber(0);
? material->SetYoungsModulus(200E6);
? material->SetCrossSectionalArea(1.0);
? material->SetThickness(1.0);
? material->SetMomentOfInertia(1.0);
? material->SetPoissonsRatio(0.2);
? material->SetDensityHeatProduct(1.0);
? m_FemObject->AddNextMaterial(material);

? // Define the element type
? typedef itk::fem::Element2DC0LinearTriangularStrain TriangularElementType;
? TriangularElementType::Pointer triangularElement= TriangularElementType::New();
? triangularElement->SetMaterial(material.GetPointer());

? // Convert mesh points into nodes
? VectorType point(3);
? PointType*ptr;
? PointType pt;
? ptr= &pt;

? int numOfPoints= itkMesh->GetNumberOfPoints();
? cout<<"itkMesh numOfPoints: "<<numOfPoints<<endl;

? int numberOfCells= itkMesh->GetNumberOfCells();
? cout<<"itkMesh numberOfCells: "<<numberOfCells<<endl;

? CellsContainerPointer cellIterator= itkMesh->GetCells();
? CellIterator cells= cellIterator->Begin();

? bool flag= true;

? for(int k=0; k< numberOfCells; k++)
? {
??? CellType *cellPtr= cells.Value();
??? cout<<"Cell Value: "<< cells.Value() << " Cell Type= " << cellPtr->GetType()<<endl;

??? switch(cellPtr->GetType())
??? {
????? case CellType::TRIANGLE_CELL:
????? {
??????? if(flag== true) // To make sure that the nodes are created just once
??????? {
??? ??? ? for(int j= 0; j< numOfPoints; j++)
??? ??? ? {
??? ??? ??? itkMesh->GetPoint(j, ptr);

??? ??? ??? typedef TriangularElementType::Node TriangularNodeType;
??? ??? ??? TriangularNodeType::Pointer triangularNode= TriangularNodeType::New();
??? ??? ??? point[0]= -1.0;
??? ??? ??? point[1]= 2.0;
??? ??? ??? point[2]= 3.0;
??? ??? ??? triangularNode->SetCoordinates(point);
??? ??? ??? triangularNode->SetGlobalNumber(j);
??? ??? ??? m_FemObject->AddNextNode(triangularNode.GetPointer());
??? ??? ? }
??? ??? ? flag= false;
??????? }
??????? PointIdIterator pointIt= cellPtr->PointIdsBegin();
??????? int i= 0;

??????? while(pointIt!= cellPtr->PointIdsEnd())
??????? {
??? ???? ? triangularElement->SetNode(i, m_FemObject->GetNode(*pointIt));
??? ????? pointIt++;
??? ????? i++;
??? ??? }
??? ??? cells++;
??? ??? triangularElement->SetGlobalNumber(k);
??? ??? m_FemObject->AddNextElement(triangularElement.GetPointer());
??? ??? break;
????? }
??? }
? }

? // Define some Load
? LoadNodePointerType loadNode =LoadNodeType::New();
? loadNode->SetElement(triangularElement);
? loadNode->SetGlobalNumber(3);
? loadNode->SetNode(1);

? vnl_vector<double> force(2);
? force[0]= 20.0;
? force[1]= -20.0;? force[2]= 10.0;
? loadNode->SetForce(force);
? m_FemObject->AddNextLoad(loadNode);

? //Solve for displacements.
? m_FemObject->FinalizeMesh();
? m_FemSolver->SetInput(m_FemObject);
? m_FemSolver->Update();???? // SEGMENTATION FAULT

? cout<< "Fem Solver Output: "<< m_FemSolver->GetOutput()<<endl;

? const unsigned int invalidId= itk::fem::Element::InvalidDegreeOfFreedomID;
? int numberOfNodes= m_FemSolver->GetInput()->GetNumberOfNodes();

? for(int i= 0; i< numberOfNodes; i++)
? {
??? itk::fem::Element::Node::Pointer node= m_FemSolver->GetInput()->GetNode(i);
??? cout<<"Fem Node : "<< node->GetGlobalNumber()<<endl;

??? for(unsigned int j= 0, dof; (dof= node->GetDegreeOfFreedom(j))!= invalidId; j++)
??? {
????? cout <<"Solver GetSolution: "<< m_FemSolver->GetSolution(dof)<<endl;
??? }
? }

? // Write the deformed mesh
? typedef itk::FEMObjectSpatialObject<3> FemObjectSpatialObjectType;
? FemObjectSpatialObjectType::Pointer m_FemSpatialObject= FemObjectSpatialObjectType::New();
? m_FemSpatialObject->SetFEMObject(m_FemSolver->GetOutput());

? typedef itk::FEMSpatialObjectWriter<3>????? ??? ??? ??? ??? FemSpatialObjectWriterType;
? FemSpatialObjectWriterType::Pointer femSpatialWriter= FemSpatialObjectWriterType::New();
? femSpatialWriter->SetInput(m_FemSpatialObject);
? femSpatialWriter->SetFileName("deformedMesh.meta");
? femSpatialWriter->Update();
}

Thank you for any help!


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/insight-developers/attachments/20171008/e5991fef/attachment.html>

From DLRdave at aol.com  Wed Oct 11 12:58:48 2017
From: DLRdave at aol.com (David Cole)
Date: Wed, 11 Oct 2017 12:58:48 -0400
Subject: [ITK-dev] HDF5 library naming
Message-ID: <CAKME8fETnafGad7RYU=DmvjbSvsBxwqP+GqKwpxFrY-HMu2xGw@mail.gmail.com>

I see the difference in release and debug library names for the HDF5
libraries has already been discussed about a year ago:

    https://itk.org/pipermail/insight-developers/2016-October/024995.html

However, the answer seems unsatisfactory to me.

Why can't ITK install these libraries named according to its own
convention for 3rd party libraries...? (i.e. -- why is "upstream did
it that way and we just updated to that" acceptable as a response
here?)

I use a non-CMake build system consuming ITK libraries where it is
much easier to deal with libraries named consistently (and using the
same names for debug and release builds). Especially when you are
dealing with static libraries only, there does not seem to be a good
reason to allow this deviation to persist long term.

Would a proposed patch correcting the naming (conforming to ITK
version naming convention like other ITK 3rd party libs, and named the
same in Release and Debug) be accepted?


Thanks,
David C.


details:

On Windows, BUILD_SHARED_LIBS=OFF,
... snippet of 'ls -la' in the Debug ITK 'lib' install directory:
...
itkvcl-4.12.lib
itkvnl-4.12.lib
itkvnl_algo-4.12.lib
itkzlib-4.12.lib
libhdf5.settings
libitkhdf5_D.lib
libitkhdf5_cpp_D.lib


Proposed new names for two hdf5 libraries for both Debug and Release builds:

itkhdf5-4.12.lib
itkhdf5_cpp-4.12.lib

From matt.mccormick at kitware.com  Wed Oct 11 13:14:17 2017
From: matt.mccormick at kitware.com (Matt McCormick)
Date: Wed, 11 Oct 2017 13:14:17 -0400
Subject: [ITK-dev] HDF5 library naming
In-Reply-To: <CAKME8fETnafGad7RYU=DmvjbSvsBxwqP+GqKwpxFrY-HMu2xGw@mail.gmail.com>
References: <CAKME8fETnafGad7RYU=DmvjbSvsBxwqP+GqKwpxFrY-HMu2xGw@mail.gmail.com>
Message-ID: <CALzTN-QFUP-nk9WuERhUr+2P=uS+mShv-HYFtKUo=XeHT8q48g@mail.gmail.com>

Hi David,

On Wed, Oct 11, 2017 at 12:58 PM, David Cole via Insight-developers
<insight-developers at itk.org> wrote:
> I see the difference in release and debug library names for the HDF5
> libraries has already been discussed about a year ago:
>
>     https://itk.org/pipermail/insight-developers/2016-October/024995.html
>
> However, the answer seems unsatisfactory to me.
>
> Why can't ITK install these libraries named according to its own
> convention for 3rd party libraries...? (i.e. -- why is "upstream did
> it that way and we just updated to that" acceptable as a response
> here?)

Because we do not have infinite resources to maintain forks and
support all build systems.


> I use a non-CMake build system consuming ITK libraries where it is
> much easier to deal with libraries named consistently (and using the
> same names for debug and release builds). Especially when you are
> dealing with static libraries only, there does not seem to be a good
> reason to allow this deviation to persist long term.
>
> Would a proposed patch correcting the naming (conforming to ITK
> version naming convention like other ITK 3rd party libs, and named the
> same in Release and Debug) be accepted?

A patch would be accepted assuming that responsibility will be taken
by the author to maintain the fork in the future.

Thanks,
Matt

From jchris.fillionr at kitware.com  Wed Oct 11 15:10:46 2017
From: jchris.fillionr at kitware.com (Jean-Christophe Fillion-Robin)
Date: Wed, 11 Oct 2017 15:10:46 -0400
Subject: [ITK-dev] HDF5 library naming
In-Reply-To: <CALzTN-QFUP-nk9WuERhUr+2P=uS+mShv-HYFtKUo=XeHT8q48g@mail.gmail.com>
References: <CAKME8fETnafGad7RYU=DmvjbSvsBxwqP+GqKwpxFrY-HMu2xGw@mail.gmail.com>
 <CALzTN-QFUP-nk9WuERhUr+2P=uS+mShv-HYFtKUo=XeHT8q48g@mail.gmail.com>
Message-ID: <CANLz8R6bhAzRjSC7nF2rn7YqMkLxFVgwK0RimZ67XXK9cFcoVg@mail.gmail.com>

Hi,

What about patching the upstream system to allow configuration of library
output name and install dir ?

Hth
Jc

On Wed, Oct 11, 2017 at 1:14 PM, Matt McCormick <matt.mccormick at kitware.com>
wrote:

> Hi David,
>
> On Wed, Oct 11, 2017 at 12:58 PM, David Cole via Insight-developers
> <insight-developers at itk.org> wrote:
> > I see the difference in release and debug library names for the HDF5
> > libraries has already been discussed about a year ago:
> >
> >     https://itk.org/pipermail/insight-developers/2016-October/
> 024995.html
> >
> > However, the answer seems unsatisfactory to me.
> >
> > Why can't ITK install these libraries named according to its own
> > convention for 3rd party libraries...? (i.e. -- why is "upstream did
> > it that way and we just updated to that" acceptable as a response
> > here?)
>
> Because we do not have infinite resources to maintain forks and
> support all build systems.
>
>
> > I use a non-CMake build system consuming ITK libraries where it is
> > much easier to deal with libraries named consistently (and using the
> > same names for debug and release builds). Especially when you are
> > dealing with static libraries only, there does not seem to be a good
> > reason to allow this deviation to persist long term.
> >
> > Would a proposed patch correcting the naming (conforming to ITK
> > version naming convention like other ITK 3rd party libs, and named the
> > same in Release and Debug) be accepted?
>
> A patch would be accepted assuming that responsibility will be taken
> by the author to maintain the fork in the future.
>
> Thanks,
> Matt
> The ITK community is transitioning from this mailing list to
> discourse.itk.org. Please join us there!
> ______________________________________
> 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://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-developers
>



-- 
+1 919 869 8849 <(919)%20869-8849>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/insight-developers/attachments/20171011/5a51cb70/attachment.html>

From nasil122002 at yahoo.de  Thu Oct 19 16:26:39 2017
From: nasil122002 at yahoo.de (Yusuf OEZBEK)
Date: Thu, 19 Oct 2017 20:26:39 +0000 (UTC)
Subject: [ITK-dev] itk FEM Solver
References: <1776029806.849749.1508444799457.ref@mail.yahoo.com>
Message-ID: <1776029806.849749.1508444799457@mail.yahoo.com>

Dear ITK community,

I want to deform a 3D model using itk?s FEM module. For this I have implemented the following workflow:

1. Read CT data-set and create 3D iso-surface using vtkMarchingCubes (output is polydata)
2. Extracted the interested surface using vtkHardwareSelector (output is 2D polydata)
3. Converted the extracted surface to the unstructured gird using vtkUnstructuredGrid and vtkDelaunay3D in order to get a 3D tetrahedron as unstructured grid (output is 3D unstructured grid)
4. Converted the unstructured grid to the itk mesh, which contains tetrahedron cells (output is itkMesh object)
5. To convert the itk mesh to the FEM model, I have created FEM object (FEMObject<3>), which contains a material (MaterialLinearElasticity), elements and nodes (Element3DC0LinearTetrahedronStrain), boundary conditions within force (LoadBC) and to generate a solution for the formulation, I have created a solver (FEMSolver<3>).

My problem is:
If I call the update of solver ?solver->Update()? I get the warnings ?/itk-4.7.2/Modules/ThirdParty/VNL/src/vxl/core/vnl/algo/vnl_qr.txx: vnl_qr::solve() : matrix is rank-deficient by 3? and generated solutions are just zeros.

The test code can be found under the link:

https://www.dropbox.com/s/v8q5qdqg9ads33n/vtkUGridToFemMesh.zip?dl=0

Thank you for any help

From nasil122002 at yahoo.de  Mon Oct 23 10:25:27 2017
From: nasil122002 at yahoo.de (Yusuf OEZBEK)
Date: Mon, 23 Oct 2017 14:25:27 +0000 (UTC)
Subject: [ITK-dev] Convert itk::FEM Object to vtk after solve
References: <1051321742.3378574.1508768727809.ref@mail.yahoo.com>
Message-ID: <1051321742.3378574.1508768727809@mail.yahoo.com>

Hello,
Is there any possibility to convert the deformed fem object to the vtk (e.g. polydata), without saving it as meta object?
Thanks!?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/insight-developers/attachments/20171023/9905d1cd/attachment.html>