Problem with vtkMarchingCubes - no isosurface

[Brent Burton]

Hello, all.

Maybe somebody will point out what I'm doing wrong with this volumetric
pipeline.  The TCL pipeline is short, so I'll just post it, then describe
the problem.  Summary: an isosurface for a volumetric dataset created
from 6 PGM images is not being generated by vtkMarchingCubes.

#---- start of volume.tcl
vtkPNMReader pnmreader
  pnmreader SetFilePrefix "1.pgm"
  pnmreader SetDataSpacing 0.6 0.6 3.5
  pnmreader SetDataExtent 0 511 0 511 1 6
  [pnmreader GetOutput] ReleaseDataFlagOn

vtkImageShrink3D shrinker
  shrinker SetInput [pnmreader GetOutput]
  shrinker SetShrinkFactors 4 4 1
  shrinker AveragingOn

vtkStructuredPointsWriter spwriter
  spwriter SetInput [shrinker GetOutput]
  spwriter SetFileName six-pts.vtk

puts "vtkMarchingCubes mcubes"
vtkMarchingCubes mcubes
  mcubes SetInput [shrinker GetOutput]
  mcubes ComputeScalarsOff
  mcubes ComputeGradientsOff
  mcubes ComputeNormalsOff
  mcubes SetNumberOfContours 1
  mcubes SetValue 0 0.627
  mcubes DebugOn
  [mcubes GetOutput] ReleaseDataFlagOn
  puts "mcubes Update"
  mcubes Update

#---- remainder of PolyData pipeline removed
#---- end of volume.tcl

The vtkPNMReader loads in 6 PGM images, each being 512x512.  Next, the
volume is reduced to 128x128x6 with the shrinker, and the reduced
StructuredPoints set is written out to six-pts.vtk.  The output of the
shrinker is also sent to the MarchingCubes to generate one isosurface with
value 0.627.

I looked at six-pts.vtk in Emacs, and the data appears normal, varying
from about 0.2 to 0.8 as expected.  The isovalue of 0.627 is a legal
value.

The problem first came up with vtk2.0p1, which I was using until
yesterday.  vtkMarchingCubes::Execute() would simply crash.  I updated to
VTK 2.4 on Linux (Redhat 5.1) and it crashes also (built with shared
libraries).  I rebuilt the entire VTK --without-shared and now it doesn't
crash.

Using gdb, I set a breakpoint at vtkMarchingCubes::Execute() and stepped
through it.  The internal data type is VTK_UNSIGNED_CHAR.  After
the call to vtkMarchingCubesComputeGradient() (a static template fcn)
returns, the number of newPts and newPolys are 0 and 0 (I.e., nothing
generated).

VTK dumps the following error five times, then reports the
number of points and triangles:

  Debug: In vtkProcessObject.h, line 117
  vtkMarchingCubes (0x8c5ced8): vtkMarchingCubes (0x8c5ced8): returning
  AbortExecute of 0
  [*5]

  Debug: In vtkMarchingCubes.cxx, line 577
  vtkMarchingCubes (0x8c5ced8): Created: 0 points, 0 triangles


So that's what I'm seeing.  Anybody have an idea about this?
I would sure appreciate hearing any tips or corrections to the
pipeline.

Thanks for reading, and have a good Thanksgiving.

-Brent Burton



-----------------------------------------------------------------------------
This is the private VTK discussion list.  Please keep messages on-topic.
Check the FAQ at: <http://www.automatrix.com/cgi-bin/vtkfaq>
To UNSUBSCRIBE, send message body containing "unsubscribe vtkusers" to
<majordomo at gsao.med.ge.com>.  For help, send message body containing
"info vtkusers" to the same address.     Live long and prosper.
-----------------------------------------------------------------------------