[vtkusers] vtkCutClosedSurface

[David Gobbi]

Hi Sebastian,

The filter does not move any points in the mesh (it just adds/removes
points).  Check the transform matrices for your actors, from the
picture it looks more like a scale difference than a shift but it is
hard to tell.

 - David


On Fri, Oct 7, 2011 at 7:42 AM, sebastian ordas
<sebastian.ordas at gmail.com> wrote:
> Thank you David!
>
> btw, using vtkClipClosedSurface, I´m getting a small shift between the
> original mesh and the closed clipped surface
> http://ScrnSht.com/byxsiw
>
> Is that a visualization issue only?
>
> regards,
> sebastian
>
>
> On 10/7/2011 3:06 AM, David Gobbi wrote:
>>
>> On Thu, Oct 6, 2011 at 4:06 PM, sebastian ordas
>> <sebastian.ordas at gmail.com>  wrote:
>>>
>>> Dear VTK developers,
>>>
>>> Following the explanation in
>>> http://www.vtk.org/Wiki/VTK/Closed_Surface_Clipping
>>> I was wondering about the current development state of the
>>> vtkCutClosedSurface class or a similar one.
>>
>> It has not been written.  Eventually I might need it, and when I do,
>> I'll write it... but that could be in a few months, a few years, or
>> never.
>>
>>> I´m basically needing the meshing of a contour, possibly with holes, but
>>> always closed.
>>
>> You can use vtkDelaunay2D for that, but be warned that some inputs
>> will cause vtkDelaunay2D to segfault (it is rare, but it does happen
>> occasionally).
>>
>>> I found this interesting:
>>>
>>> http://www.geometrictools.com/Documentation/TriangulationByEarClipping.pdf
>>
>> That method is very similar to the one that vtkClipClosedSurface uses.
>> It is messy but it is robust.  The fastest triangulation methods are
>> randomized Delaunay methods, but they are very sensitive to roundoff
>> error, and as a result it is hard to write a robust Delaunay algorithm.
>>
>>> Does anyone know a workaround (has to be fast, for a real time
>>> application)
>>> with existing VTK classes?
>>
>> For now, vtkDelaunay2D is the only class for doing triangulations with
>> holes.
>>
>>  - David
>
>