[Paraview] [vtkusers] how to store label fields

[Dominik Szczerba]

On Thu, 2009-07-02 at 16:41 -0400, Berk Geveci wrote:
> OK, I think I understand now. This should work irrespective of ParaView version:
> 
> - Turn off "Interpolate Scalars"
> - Change Interpolation to Flat
> 

Yep - this only works in CVS and was a reason for my confusion. I still
wonder who noticed it as the ML archives go silent about it.

> It is possible that there was some bug that prevented something from
> working before.
> 
> If you have cell centered data, ParaView should never interpolate
> colors. Can you send me instructions to reproduce this?
> 

Use the dataset I sent you. Apply pointDataToCellData. Cut to you
liking. Observe the interpolation.

D0minik

> -berk
> 
> 
> On Thu, Jul 2, 2009 at 9:41 AM, Dominik Szczerba<dominik at itis.ethz.ch> wrote:
> > Berk,
> >
> > Example data: http://otoro.itis.ethz.ch/~dominik/tmp/head_eye.vti
> >
> > The data is point-centered. I was raising up a question if I should not
> > store my label fields (from segmentation) as cell centered StructureGrid
> > instead, because PV interpolates the colors regardless of the states of
> > 'interpolate colors / Goraud/Flat'. But the PV-CVS is no longer
> > interpolating integer point data (still insensitive to the mentioned
> > states). So something changed (for good), I was asking what and why.
> >
> > As to the contour filter - contouring such (point based) label field
> > results in an unstructured grid, so for arbitrary angles the voxel
> > structure is anyway lost, thus interpolation makes sense. If however
> > switched off it seems to correctly show (crippled) triangles.
> >
> > However, converting point data to cell data on my label field above
> > results in interpolation again, that can not be switched off - probably
> > because some interpolation is used during the transfer. So something is
> > probably still not OK here.
> >
> > So a bottom line: is it finally OK to store label fields as ImageData
> > (where only POINT_DATA is allowed) or do I take StructuredGrid instead
> > and store the labels explicitly as CELL_DATA?
> >
> > regards,
> > Dominik
> >
> > On Thu, 2009-07-02 at 09:09 -0400, Berk Geveci wrote:
> >> This confuses me. The slice filter (vtkCutter) is not supposed to
> >> interpolate any cell-centered data whether it is integer or double. To
> >> my knowledge, there has been no bug that changed this. Can you send me
> >> a small dataset that I can reproduce this with?
> >>
> >> -berk
> >>
> >> On Thu, Jul 2, 2009 at 8:12 AM, Dominik Szczerba<dominik at itis.ethz.ch> wrote:
> >> > I just checked that PV from CVS addresses exactly this issue, so I am
> >> > apparently not alone in my considerations. PV 3.4 will load a label
> >> > field and interpolate the colors when viewing slices (thus destroying
> >> > the image). PV CVS ignores interpolation settings and displays my labels
> >> > correctly (without interpolation). So I assume somebody realized the
> >> > problem recently too and corrected it. So I conclude that ImageData is
> >> > still the right choice for discrete label fields (and not
> >> > StructuredGrid), but one has to rely on the latest version. Please
> >> > correct me if I am wrong.
> >> >
> >> > -- Dominik
> >> >
> >> > On Wed, 2009-07-01 at 18:07 +0200, Dominik Szczerba wrote:
> >> >> When storing my label field as imagedata I get unwanted interpolation if
> >> >> I e.g. cut through the volume. Switching the interpolation off results
> >> >> in funny bad looking triangles. What's the best/intended/practiced way
> >> >> of storing (integer) label fields? Image data is point centered while
> >> >> label fields are clearly cell centered. As I see the only alternative is
> >> >> StructuredGrid, but functions like contouring or resampling require
> >> >> point data anyway. So what's better: store my labels as image (point)
> >> >> data and compromise display (unwanted interpolation) or go for
> >> >> structured grid and compromise size of the file and needed conversions
> >> >> to image data anyway?
> >> >>
> >> >> -- Dominik
> >> >>
> >> >> _______________________________________________
> >> >> Powered by www.kitware.com
> >> >>
> >> >> Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html
> >> >>
> >> >> Please keep messages on-topic and check the VTK FAQ at: http://www.vtk.org/Wiki/VTK_FAQ
> >> >>
> >> >> Follow this link to subscribe/unsubscribe:
> >> >> http://www.vtk.org/mailman/listinfo/vtkusers
> >> >>
> >> >
> >> > _______________________________________________
> >> > Powered by www.kitware.com
> >> >
> >> > Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html
> >> >
> >> > Please keep messages on-topic and check the ParaView Wiki at: http://paraview.org/Wiki/ParaView
> >> >
> >> > Follow this link to subscribe/unsubscribe:
> >> > http://www.paraview.org/mailman/listinfo/paraview
> >> >
> >>
> >
> >
>