brains:RemoveOrientImageFilter

From NITRC Wiki

FILES using itk:OrientImageFilter:

BRAINSTools/GTRACT/Cmdline/gtractCopyImageOrientation.cxx

Don't know what to do with this. It tries to take an image and put it in the same orientation as a reference image.

First of all, if all processing is in physical space, then this program should never be used. Second, to do it correctly would involve computing the input to reference transform and apply it to the input image, which would require resampling, which I presume is a bad thing.

BRAINSTools/GTRACT/Cmdline/gtractCoRegAnatomyBspline.cxx

Does the same thing as gtractCopyImageOrientation -- if BRAINSFit is working in physical space now, this can just disappear.

BRAINSTools/GTRACT/Cmdline/gtractCoRegAnatomyPrimary.cxx

Looks to be a hacked up version of gtractCoRegAnatomyBspline. Or vice versa. Same thing applies.

BRAINSTools/GTRACT/Cmdline/gtractImageConformity.cxx

This program is almost exactly the same as gtractCopyImageOrientation and I don't know why it's still here.

BRAINSTools/GTRACT/Cmdline/gtractResampleCodeImage.cxx

This makes no sense except as a symptom of blind block copying between programs; it uses OrientImageFilter to put the input image into the same orientation as a reference image, and THEN applies a transform read from a file. I think it's guaranteed to always do something wrong.

BRAINSTools/GTRACT/Cmdline/gtractResampleFibers.cxx

Pretty much the same story as theother programs here. According to Joy, most of these are no longer relevant.

src/iplProg/BRAINSMarchingCubes/vtkITKArchetypeImageSeriesReader.cxx

Has a concept of 'desired coordinate orientation' and uses OrientImageFilter to achieve it. Don't know if this option is ever used or if we can kill it.

src/iplProg/BRAINSMarchingCubes/vtkITKArchetypeImageSeriesScalarReader.cxx

Same as above

src/iplProg/DebugImageViewer/DebugImageViewerClient.h

This has to do with preserving assumptions in the DebugImageViewer.

src/iplProg/GenusZeroImageFilter/vtkITK/vtkITKArchetypeImageSeriesReader.cxx src/iplProg/GenusZeroImageFilter/vtkITK/vtkITKArchetypeImageSeriesScalarReader.cxx src/iplProg/GenusZeroImageFilter/vtkITK/vtkITKArchetypeImageSeriesVectorReaderFile.cxx src/iplProg/GenusZeroImageFilter/vtkITK/vtkITKArchetypeImageSeriesVectorReaderSeries.cxx

These all appear to be code from the same source as the ones in BRAINSMarchingCubes, but again who knows how it's used, except it probably should never be.

src/iplProg/InsightApplications/MRIBiasCorrection/imageutils.h

All of these usages are applied to masks being read in. I don't know if mask processing happens in physical or index space, if the latter, then this stuff should never work.

src/iplProg/PPD/code/DTIBsplineRegistration.cxx

based on a flag OrientAnatomicalImage, it reorients the moving and fixed images. Don't know if the actual registration works in physical space, but I sure hope so.

src/iplProg/surface/ClipLabelMap/clip.cxx src/iplProg/surface/SurfaceLabel/ClipLabelMap/clip.cxx src/iplProg/surface/TopologyCorrection/07312008/GraphAll.cxx src/iplProg/surface/TopologyCorrection/08182008/GraphAll.cxx

none of the surface subdirectory looks like it's production code, and if it isn't maybe it needs to go away.

src/iplProg/Tensor_Transform/TensorTransformation/DTIBsplineRegistration.cxx src/iplProg/Tensor_Transform/TensorTransformation/DTIRigidRegistration.cxx Both of these programs reorient everything based on the OrientAnatomicalImage command line flag. This is duplicate code with src/ioplProg/PPD/code/DTIBsplineRegistration. If this isn't production code, it needs to go away.

FILES using itkUtil::ReadImageAndOrient

BRAINS3/src/iplProg/DebugImageViewer/vtkKWImageViewerWidget.cxx BRAINS3/src/iplProg/ImageEval2/ImageEval-Old/vtkKWMyWindow.cxx

Both of these programs need re-writing, and should figure out some more copacetic way to deal with image orientation -- the same solution as should be used in BRAINSTracer.

BRAINS3/src/iplProg/itkBRAINSClassify/ReadMask.h

This is code that's the same as ReadMask.h used elsewhere.

BRAINS3/src/iplProg/itkBRAINSClassify/itkBRAINSClassifyTrainingClassesPrimary.cxx

Uses ReadImageAndOrient for some reason I know not what. Given that nothing else in this project does this, I think it's probably code copied from elsewhere that should have been more carefully vetted.

BRAINSTracerQT

needs to not reorient everything it works on. This has the unfortunate effect of having all points traced potentially being junk.

FILES using itkUtil::OrientImage

BRAINS3/BRAINSTools/BRAINSConstellationDetector/src/landmarkIO.cxx

appears to only be used to create a DebugImage, and I don't know whether it's even relevant.