[Insight-developers] Image as an OrientedImage Progress

[Bill Lorensen]

Folks,

As you can see in this build,
http://www.cdash.org/CDash/viewTest.php?onlyfailed&buildid=173260
there are only 11 tests failing with
ITK_IMAGE_BEHAVES_AS_ORIENTED_IMAGE turned on.

The previous failures were fixed by:

BUG: need to initialize internal matrices if
ITK_IMAGE_BEHAVES_AS_ORIENTED_IMAGE is ON.
http://public.kitware.com/cgi-bin/viewcvs.cgi/Code/Common/itkImage.txx?root=Insight&sortby=date&view=log

BUG: directions were not initialized.
http://public.kitware.com/cgi-bin/viewcvs.cgi/Code/BasicFilters/itkExtractImageFilter.txx?root=Insight&sortby=date&view=log

BUG: After vector.resize() in  SetNumberOfDimensions, origin should be
set to 0 and spacing to 1, the defaults.
http://public.kitware.com/cgi-bin/viewcvs.cgi/Code/IO/itkImageIOBase.cxx?root=Insight&sortby=date&view=log

ENH: support for ITK_IMAGE_BEHAVES_AS_ORIENTED_IMAGE.
http://public.kitware.com/cgi-bin/viewcvs.cgi/Code/SpatialObject/itkImageSpatialObject.txx?root=Insight&view=log#rev1.57

BUG: spacing[0] = 0.0.
http://public.kitware.com/cgi-bin/viewcvs.cgi/Testing/Code/IO/itkReadWriteSpatialObjectTest.cxx?root=Insight&view=log#rev1.25

This is very encouraging. In fact 8 of the failing tests are due to
one ANalyze test program that is producing processing 2D Analyze files
that have {1 0} {0 0} direction cosines.

The other 3 are due to strange DICOM files I think.

I'll look at the DICOM failures, but someone from Iowa (Hans or Kent)
will need to look at the Analyze issue.

I have added better diagnostics to itkImage that produce more
informative errors (other than "Singular matrix"). This should help
find current and future test failures.

BTW, no regression test images were changed or injured in this exercise!

Bill