[Insight-users] Fwd: Few questions about quad edge mesh filters

Andriy Fedorov fedorov at bwh.harvard.edu
Wed Mar 17 12:07:52 EDT 2010


Previous message bounced, I put the attachments here:
http://dl.dropbox.com/u/3770904/QuadEdgeScreenhots.tgz


---------- Forwarded message ----------
From: Andriy Fedorov <fedorov at bwh.harvard.edu>
Date: Wed, Mar 17, 2010 at 12:04
Subject: Few questions about quad edge mesh filters
To: ITK Users <insight-users at itk.org>


Hi,

I have a processing pipeline that starts from a binary image,
generates quad-edge marching cubes surface, performs decimation, and
computes the surface normals. I am trying to use quad edge filters for
this task.

I stumbled across two issues:

1) the normals produced by QuadEdgeMesh are always (nan,nan,nan). I
saw the comment in the documentation, that the input and output mesh
traits should be the same for this filter to work, but this does not
seem to help.

2) whenever I am using QuadEdgeMeshQuadricDecimation, the decimated
surface is obviously incorrect
(QuadEdgeMeshSquaredEdgeLengthDecimation generates very nice surface).
Attached are screenshots for the MC surface, quadrics-decimated
surface, and squared edge decimation result

Source code and test data: http://dl.dropbox.com/u/3770904/QuadEdgeTests.tgz
Command line to reproduce the first problem (second issue can be
reproduced trivially by swapping the decimator): "QuadEdgeMeshTests
blob1.nrrd output_mesh.vtk 10"

I see this problem both with the latest checkout of the 3.16 branch,
and with the current trunk build.

Am I doing something wrong, or is this a bug?

Thanks

Andriy Fedorov


More information about the Insight-users mailing list