[ITK-users] Open surface mesh construction from feature map
Will Schroeder
will.schroeder at kitware.com
Wed Sep 7 11:47:25 EDT 2016
Wei-
FYI- I have an initial implementation in a VTK remote module
(vtkPointCloud) of an isosurface algorithm that can generate open surfaces
(typically from truncated distance functions). If you have some data I can
use it for testing and debugging purposes....
Best,
W
On Wed, Sep 7, 2016 at 11:43 AM, Wei Liu <weiliu620 at gmail.com> wrote:
> Dženan, thanks for the reply. I have no experience in mesh generation or
> surface construction, so I try to use filters in ITK/VTK.
>
> vtkDelaunay3D seems for closed surface. Do you know any filter for open
> surface?
>
> From your email it seems straightforward to generate mesh given point
> clouds with normal vectors. I don't trust my own implementation, but I'll
> give it a try if I couldn't find any filters. Do you know any papers, or
> example implementations, either inside or outside of ITK community?
>
> Thanks again,
>
> On Tue, Sep 6, 2016 at 9:17 AM, Dženan Zukić <dzenanz at gmail.com> wrote:
>
>> Hi Wei,
>>
>> you should calculate the peak of your probability so you have "thin"
>> layer of voxels representing your surface. Coupled with the knowledge of
>> the normal, it would then be only the matter of determining connectivity
>> for the voxels to generate triangles of the mesh. And yes, you should use
>> VTK for mesh storage and later filtering (such as smoothing).
>>
>> You could look at Canny filter
>> <https://itk.org/Doxygen/html/classitk_1_1CannyEdgeDetectionImageFilter.html>'s
>> source code for an example in ITK of how to do thinning.
>>
>> HTH,
>> Dženan
>>
>> On Mon, Sep 5, 2016 at 5:25 PM, Wei Liu <weiliu620 at gmail.com> wrote:
>>
>>> Hi ITK users,
>>>
>>> I'm trying to detect (possibly multiple) open surface in a 3D volume.
>>> Suppose I have a filter that give me a 3D feature map (e.g. the probability
>>> of each voxel being on the surface), where should I go from this feature
>>> map to construct smooth open surfaces?
>>>
>>> I did some search. It seems most of existing examples construct
>>> isosurface mesh from a binary volume (or distance map from it). The binary
>>> volume usually has a closed surface on the object boundary, so it is
>>> different from my problem of open surface.
>>>
>>> I can probably calculate the normal of the surface at each voxel, if
>>> that helps.
>>>
>>> I guess I should do some skeleton extraction and get a 'thin' set of
>>> voxels on the surface, and use some filters in ITK or VTK(new to VTK but
>>> willing to learn if necessary) to convert binary volumes to surface meshes,
>>> and do some smoothing, filling holes etc.
>>>
>>> Thanks!
>>>
>>>
>>>
>>> _____________________________________
>>> 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://www.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-users
>>>
>>>
>>
>
> _____________________________________
> 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://www.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-users
>
>
--
William J. Schroeder, PhD
Kitware, Inc. - Building the World's Technical Computing Software
28 Corporate Drive
Clifton Park, NY 12065
will.schroeder at kitware.com
http://www.kitware.com
(518) 881-4902
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/insight-users/attachments/20160907/0b1e802c/attachment.html>
More information about the Insight-users
mailing list