[vtkusers] Re: Compute principle curvaturs

[Philipp.Batchelor]

Hello

yes, it's something I have experienced on datasets coming from e.g. 
medical images, via marching cubes (we used to actually use this 
criterion, i .e. positivity of H^2-K), to test datasets. The original 
idea I used for the computation of discrete mean curvatures might not be 
the best (although in my tests it performed ok). In any case, I suspect 
more the computation of the mean curvature than Gauss.

Some more sophisticated methods exist for direct computation of 
principal curvatures, e.g. fit a surface locally, and compute its mean 
curvatures.
I'am afraid it's been quite a while I haven't come back to this, but I 
think an update of Curvatures is welcome.

Btw, there was also once a discussion about curvatures of a line 
(curvature and torsion),with a class  vtkPolylineCurvature being 
circulated. It had some little bugs, which I corrected, I'll post it 
here if anyone is interested (but about to catch a flight so will have 
to wait a week at least).

Ph
Goodwin Lawlor wrote:with a class
> I've updated vtkCurvatures to calculate the principal curvatures... there
> seems to be alot of undefined points (especially with a vtkSphereSouce as
> the geometry). Perhaps the discrete methodology used to calculate the
> Gaussian and Mean curvatures either overestimate the Gaussian or
> underestimates the Mean. It seems to be dependant on the area of the
> individual triangles- with larger triangles producing undefined vertices.
> 
> The updated source is here:
> http://www.bioengineering-research.com/vtk/vtkCurvatures.cxx
> http://www.bioengineering-research.com/vtk/vtkCurvatures.h
> 
> Goodwin
> 
> 
> 
> "Moti Freiman" <freiman at cs.huji.ac.il> wrote in message
> news:000101c45d02$22c4d820$2a284184 at cas02...
> 
>>Hello!
>>I need to compute the principle curvatures at my software.
>>I saw that the vtkcurvatures class just compute the mean and Gaussian
>>curvatures. When I used the formula for extracting the principle
>>curvatures from mean and Gaussian,  I got undefind situation (negative
>>sqrt). Do someone know how I can get the principle curvatures with vtk?
>>Thank you
>>Moti freiman
>>
>>_______________________________________________
>>This is the private VTK discussion list.
>>Please keep messages on-topic. Check the FAQ at:
> 
> <http://public.kitware.com/cgi-bin/vtkfaq>
> 
>>Follow this link to subscribe/unsubscribe:
>>http://www.vtk.org/mailman/listinfo/vtkusers
>>
> 
> 
> 
> 
> _______________________________________________
> This is the private VTK discussion list. 
> Please keep messages on-topic. Check the FAQ at: <http://public.kitware.com/cgi-bin/vtkfaq>
> Follow this link to subscribe/unsubscribe:
> http://www.vtk.org/mailman/listinfo/vtkusers
>