Hi cq,
ok I got the point. The only way to extract "real" centerlines without
specifying endpoints is to
1. have a surface that is open at the ends
2. have a way to specify endpoints automatically.
If 1. is not the case, then you have to open them somehow or resort to 2.
I have not seen your specific surface, but what you could do, if you're up for
some vtk coding, is
to use the output of vmtknetworkextractor as a hint, basically taking the ends
of the network as
seeds for the centerline. All you need is essentially to have a way to extract
the network endpoints
and pass them along to centerlines. This is a possibility, do you envision this
would work for you?
Luca
On Nov 17, 2010, at 6:25 PM, ququ wrote:
> Hi Luca,
> So far I can only get bpm files, that depends on other guys. But
> that’s fine, this part works.
>
> I got the centerline successfully using vmtknetworkextraction. But
> when I used this centerline as the input of vmtkbranchextractor, I didn’t get
> good result.
> vmtkbranchextractor -ifile xx.vtp -ofile xx.vtp -radiusarray Radius
> I can understand vmtknetworkextraction doesn’t know the start point,
> and it is not very accuracy.
> What I want to do is: I have a vessel tree with lot of branches, I
> want to get every branch’s radius, and their group. Just like
> ‘vmtkcenterlineviewer -ifile xxx.vtp -cellarray TractIds’. If I use
> vmtkcenterlines to create centerline, I can get good result.
> Any suggestion on this?
>
> Really appreciate your help.
>
> Thanks,
> -cq
> From: Luca Antiga [mailto:luca.ant...@gmail.com]
> Sent: Wednesday, November 17, 2010 4:49 AM
> To: ququ chen
> Cc: vmtk-users@lists.sourceforge.net
> Subject: Re: is it possible to create centerlines without choosing start and
> end point?
>
> Hi cq,
> ok. I find it "unusual" that you are working with bmp files. Usually they
> are 8bit, and you get to lose a lot of information compared to the images on
> the scanner. Don't you have the possibility of exporting images from the
> scanner directly in DICOM format? Just a suggestion, but I'd consider it.
>
> If you have a very complex tree and you don't need a lot of accuracy in
> extracting centerlines, you can use vmtknetworkextraction.
> To run it, please make sure that your input surface has at least one opening.
> In case it's close, just open it manually, say at the inlet, using
> vmtksurfaceclipper.
>
> Hope this helps
>
> Luca
>
>
> On Nov 16, 2010, at 11:09 PM, ququ chen wrote:
>
>
> Hi Luca,
> Thanks for your reply.
> It’s my dicom files’ issue. Python crashed because an exception was
> thrown from vtk class during reading slices.
> Now I read bmp files directly instead of converting bmp to dicom.
>
> I have another question about centerlines. I changed the subject to
> make this thread more clear.
> Now I have a vessel tree with lot of branches and I need the
> centerlines of whole tree. is there a way to get the centerlines without
> choosing start and end point?
>
> Thanks,
> -cq
>
>
> From: Luca Antiga [mailto:luca.ant...@gmail.com]
> Sent: Sunday, November 14, 2010 10:37 AM
> To: ququ
> Cc: vmtk-users@lists.sourceforge.net
> Subject: Re: [vmtk-users] vmtk - getting started
>
> Hi cq,
> the fact that python crashes is not normal, and very likely you won't solve
> your problem by using the C++ directly.
>
> You may want to first try out the both possible ways of reading DICOM files
> in vmtk:
>
> the first is
> vmtkimagereader -f dicom -d directory_with_dicom_files_in_it
>
> and the other is
> vmtkimagereader -ifile first_dicom_file_in_your_series.dcm
>
> The first uses vtkvmtkDICOMImageReader, while the second uses the ITK I/O
> mechanisms (which rely on GDCM, a dicom library).
>
> In any case, using a class from vmtk in C++ is no different than using a VTK
> class from C++, since all vmtk classes are derived from VTK.
> For the parameters to set, you can directly refer to how they are used in
> vmtkimagereader.py.
> In order to bulid against vmtk, just include the VMTKUse.cmake file that is
> in the bulid directory (or in the install directory) in your CMakeLists.txt
> file to successfully compile against vmtk. If you don't know about CMake or
> VTK, please refer to the relative manuals and mailing lists.
>
> Let me know if you need further directions
>
> Luca
>
>
>
> On Nov 12, 2010, at 5:48 PM, ququ wrote:
>
> Hi Luca,
> Could you please provide some c++ test classes for your c++ code? Now
> I try to read some dicom images using vmtkimagereader, python crashes every
> time. I want to use c++ code to call vtkvmtkDICOMImageReader directly.
> Thanks a lot.
>
> Regards,
> -cq
>
>
------------------------------------------------------------------------------
Beautiful is writing same markup. Internet Explorer 9 supports
standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3.
Spend less time writing and rewriting code and more time creating great
experiences on the web. Be a part of the beta today
http://p.sf.net/sfu/msIE9-sfdev2dev
_______________________________________________
vmtk-users mailing list
vmtk-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/vmtk-users
