Em 18-03-2012 15:25, Luca Antiga escreveu:
Excellent, I'm happy to know you're on the right track.
Have fun!
Luca
On Mar 18, 2012, at 1:30 PM, Lia Silva wrote:
Em 18-03-2012 11:56, Luca Antiga escreveu:
Hi Lia,
that's not quite what I intended. I'll you a few more explicit
suggestions:
1) all the algorithms in vmtk are already in C++. The fact that
you're using them
in vmtk scripts is just for convenience. Note that this does not
impose *any* overhead,
speed-wise.
2) make sure you understand why you need to have a C++-only code. Is
it for speed only? See point 1). Is if because you want to
incorporate vmtk in
a purely C++ program? This would make more sense.
3) every algorithm in vmtk is a VTK class. If you know how to
program in C++ using VTK,
you know how to program in C++ using vmtk. So, make sure you are
proficient writing
VTK code before you move on with vmtk. There's a great resource of
examples here:
http://www.vtk.org/Wiki/VTK/Examples/Cxx
4) Once you have addressed 1-3, have a look into vmtk/vmtkScripts
and identify the scripts
you're interested in and take a look at what classes are used. For
instance, the code for
vmtksurfacetriangle is
cleaner = vtk.vtkCleanPolyData()
cleaner.SetInput(self.Surface)
cleaner.Update()
triangleFilter = vtk.vtkTriangleFilter()
triangleFilter.SetInput(cleaner.GetOutput())
triangleFilter.Update()
self.Surface = triangleFilter.GetOutput()
which in C++ would become
vtkCleanPolyData* cleaner = vtkCleanPolyData::New();
cleaner->SetInput(surface);
cleaner->Update();
vtkTriangleFilter* triangleFilter = vtkTriangleFilter::New();
triangleFilter->SetInput(cleaner->GetOutput());
triangleFilter->Update();
surface = triangleFilter->GetOutput();
As you see, the high-level code is almost exactly the same, and in both
cases, the call to Update triggers code that is already written in C++.
Hope this helps clarifying, but most of all, what you should clarify
is point 2, i.e. what
is the need you're trying to address when you say you want pure C++
code.
Best regards
Luca
On Mar 17, 2012, at 6:41 PM, Lia Silva wrote:
Em 17-03-2012 07:40, Luca Antiga escreveu:
Dear Lia,
vmtk code presents itself as Python code, but it is in fact a
tiny wrapper
on top of C++ code. All algorithms are already implemented in C++ (see
vmtk/vtkVmtk directory). If you look for code optimizations, you
should look
into the C++ code in that directory.
Best regards
Luca
--
Luca Antiga, PhD
Cofounder and Principal Scientist, Orobix Srl
via L.A. Muratori 3, 24123 Bergamo, Italy
orobix: www.orobix.com <http://www.orobix.com/>
home: lantiga.github.com <http://lantiga.github.com/>
twitter: twitter.com/lantiga <http://twitter.com/lantiga>
On Mar 17, 2012, at 12:50 AM, Lia Silva wrote:
Hello all!
How can i convert a vmtk code into c++ code, so i can run it
faster? The
code in vmtk it's attached.
Thank you in advance
<tudo>------------------------------------------------------------------------------
This SF email is sponsosred by:
Try Windows Azure free for 90 days Click Here
http://p.sf.net/sfu/sfd2d-msazure_______________________________________________
vmtk-users mailing list
vmtk-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/vmtk-users
Não foram detectados vírus nesta mensagem.
Verificado por AVG - www.avg.com <http://www.avg.com/>
Versão: 2012.0.1913 / Base de dados de Vírus: 2114/4874 - Data de
Lançamento: 03/16/12
So, it means that i have to implement all in c++ since the
beginning? For example, i used the /vmtkimagereader/, so in c++
i've to use/Dicomseriereadprinttagswrite/? And for the
/vmtkimagevoiselector /i have to use the
/RegionofInterestImageFilter/? Is that it? I'm so confused... Isn't
there another way easier and faster? I thought i could convert
directly into c++..
Thank you for all your pacience
------------------------------------------------------------------------------
This SF email is sponsosred by:
Try Windows Azure free for 90 days Click Here
http://p.sf.net/sfu/sfd2d-msazure_______________________________________________
vmtk-users mailing list
vmtk-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/vmtk-users
Não foram detectados vírus nesta mensagem.
Verificado por AVG - www.avg.com <http://www.avg.com/>
Versão: 2012.0.1913 / Base de dados de Vírus: 2114/4878 - Data de
Lançamento: 03/18/12
Dear Luca,
Thank you for the answer. It was very helpfull. In fact, the main
reason i want to use just C++ isn't just to make it faster. This is
also a reason, but i dont know how exactly vmtk works, and i need to
do a automatic/semi-automatic tool for measure the aorta diameter. I
found vmtk by accident and i liked it very much cause its easier to
work with, but i dont know if i can make a automatic tool using it (i
dont know how to automatize the process) . In my circle, nobody knows
vmtk, and my thesis' supervisor just work with c++, so i'm going
forward using vmtk (cause i dont have much experience with c++ and
vmtk scripts are easier for me) and later, i pretend to convert into
c++ so my supervisor can correct it.
You gave me an excellent help and now i'm doing the step 4 by seeing
the vmtk scripts. If i have any addicional doubt, i hope i can count
on you.
Thank you very much for your pacience and dedication.
Best regards
Lia
Não foram detectados vírus nesta mensagem.
Verificado por AVG - www.avg.com <http://www.avg.com>
Versão: 2012.0.1913 / Base de dados de Vírus: 2114/4878 - Data de
Lançamento: 03/18/12
Dear Luca,
I decided not to convert into C++, for now. It's not worth it.
Now i'm facing two problems: is there a way to automatize some
parameters? I would like to know if i can change the python code, so i
dont need to input the parameters manually at the console. In the
/vmtkimagevoiselector/ i can specify the coordinates of my cube, and
if i could insert them in the code it became easier. Same thing with the
/vmtklevelsetsegmentation/. The parameters are always the same, so can i
insert them into the code? And how do i do that?
The other issue is: I want to measure the diameter of the ascending
aorta in nine specifics points. I computed the centerline that gave me
the central points and the radius in all points since the source till
the target! But i just need nine points. Is there a way to do that?
For this project i'm doing a lot of steps.. If i could avoid the most of
them, it would be easier to automatize the process and to decrease the
processing time. I mean, for compute the centerline its really necessary
to run the /vmtkimagevoiselector, vmtklevelsetsegmentation/ and
/vmtkmarchingcubes/?
I'm sorry for the amount of questions, but i'm new at vmtk and programming..
Best regards,
Lia Silva
------------------------------------------------------------------------------
This SF email is sponsosred by:
Try Windows Azure free for 90 days Click Here
http://p.sf.net/sfu/sfd2d-msazure
_______________________________________________
vmtk-users mailing list
vmtk-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/vmtk-users
