Re: Error in axis alligned polygonization

New Message Reply Date view Thread view Subject view Author view

Robert Grzeszczuk (rg)
Mon, 14 Jun 1999 15:01:44 -0700 (PDT)


Manfred,

These are indeed voxel coordinates, and not the vertex coordinates that are
clipped to brick boundaries. This is intended and desirable. For example,
consider a 256^2 image (brick). Mapping this image onto a matching size
rectangle will map texel 255 onto the right edge of the image (255). The
implied mapping is an identity (255 -> 255). Now, suppose that we want to
stretch this image onto a rectangle twice as large. We have to use eplicit
mapping (255 -> 511) which will map the image onto the whole rectangle. If you
want to avoid stretching while mapping onto large area (e.g., a ROI that
extends to the outside of the image), you should continue using the identity
mapping (511 -> 511) but the voxel coordinates should be clipped (or repeated)
to brick boundary (255).

Does this clear things up? What are you trying to accomplish?

-rg

On Jun 9, 2:46pm, Manfred Weiler wrote:
> Subject: Error in axis alligned polygonization
>
> I think I discovered an error within the polzgonalization
> method. It occurs with 2D texture interpolation.
>
> When you specify explicit texture coordinats that
> differ from the coordinates of the vertices then
> voGeometryAction::polygonize only produces a view
> polygons and doesn't slice the whole geometry.
>
> I suppose that in this case not the vertex coordinates
> but the texture coordinates are clipped to the brick
> bounaries because when texture coordinates and
> vertex coordinates are the same everything works fine.
>
>
> Manfred.
>-- End of excerpt from Manfred Weiler


New Message Reply Date view Thread view Subject view Author view

This archive was generated by hypermail 2.0b2 on Mon Nov 01 1999 - 14:02:50 PST