Todo List

Shall we instead inhibit the declaration of the functions when GDAL version does not match?

We should detect whether the official C++11 std::unique_ptr<> is available instead of always using boost.interprocess.unique_ptr<>.

File is_contiguous.h

What licence to use ?

Member itk::AffineTransform< TScalarType, NDimensions >::Rotate2D (TScalarType angle, bool pre=0)

Find a way to generate a compile-time error is this is used with NDimensions != 2.

Member itk::AffineTransform< TScalarType, NDimensions >::Rotate3D (const OutputVectorType &axis, TScalarType angle, bool pre=0)

Find a way to generate a compile-time error is this is used with NDimensions != 3.

Class itk::AnisotropicDiffusionFunction< TImage >

Automatically generate the time step value from image dimensionality and order of the equations

Class itk::AzimuthElevationToCartesianTransform< TScalarType, NDimensions >

Is there any real value in allowing the user to template over the scalar type? Perhaps it should always be double, unless there's a compatibility problem with the Point class.

Derive this class from a yet undefined TransformBase class. Currently, this class derives from AffineTransform, although it is not an affine transform.

Class itk::BilateralImageFilter< TInputImage, TOutputImage >

Support color images

Support vector images

Class itk::Brains2MaskImageIO

Class itk::CannyEdgeDetectionImageFilter< TInputImage, TOutputImage >

Edge-linking will be added when an itk connected component labeling algorithm is available.

Member itk::CheckerBoardImageFilter< TImage >::PrintSelf (std::ostream &os, Indent indent) const

Add details about this class

Member itk::ColorTable< TPixel >::GetColorComponent (unsigned int colorId, char rgb)

Needs renaming.

Member itk::CurvatureFlowFunction< TImage >::ComputeGlobalTimeStep (void *GlobalData) const

compute timestep based on CFL condition.

Class itk::CurvatureNDAnisotropicDiffusionFunction< TImage >

References

Member itk::DeformationFieldSource< TOutputImage >::PrintSelf (std::ostream &os, Indent indent) const

Add details about this class

Member itk::EigenAnalysis2DImageFilter< TInputImage, TEigenValueImage, TEigenVectorImage >::MakeOutput (unsigned int idx)

Verify that MakeOutput is createing the right type of objects this could be the cause of the reinterpret_cast bug in this class

Member itk::ExponentialDeformationFieldImageFilter< TInputImage, TOutputImage >::PrintSelf (std::ostream &os, Indent indent) const

Add details about this class

Class itk::FiniteDifferenceFunction< TImageType >

Possibly subclass this object from Function. Stumbling blocks here are the specialized api of FiniteDifferenceFunction.

Member itk::i18n::I18nOpenForReading (const std::string &str)

check if cygwin has and needs the O_BINARY flag

Member itk::i18n::I18nOpenForWritting (const std::string &str, const bool append=false)

check if cygwin has and needs the O_BINARY flag

Class itk::ImageMomentsCalculator< TImage >

It's not yet clear how multi-echo images should be handled here.

Member itk::InverseDeformationFieldImageFilter< TInputImage, TOutputImage >::KernelTransformType

Check that input and output images have the same number of dimensions; this is required for consistency.

Member itk::InverseDeformationFieldImageFilter< TInputImage, TOutputImage >::PrintSelf (std::ostream &os, Indent indent) const

Add details about this class

Member itk::MRCImageIO::m_MetaDataHeaderName

Move to itkIOCommon with the other MetaDataDictionary keys, likely rename the symbol to something like ITK_MRCHHeader. (remember to fix class doc too)

Class itk::NeighborhoodIterator< TImage, TBoundaryCondition >

Better support for regions with negative indicies.

Add Begin() and End() methods?

Member itk::QuadEdgeMesh< TPixel, VDimension, TTraits >::Splice (QEPrimal *a, QEPrimal *b)

Compare the geometry of the two points and accept splicing when their geometry matches. We could fix an epsilon threshold distance above which the two points are considered distinct.

Member itk::QuadEdgeMeshConstFrontIterator< TMesh, TQE >::~QuadEdgeMeshConstFrontIterator ()

do we need here a : Superclass( mesh, start, seed ) { }

Member itk::QuadEdgeMeshEdgeMergeDecimationFilter< TInput, TOutput, TCriterion >::CheckOrientation (OutputQEType *iEdge, const OutputPointIdentifier &iId, const OutputPointType &iPt)

Finish to implement this method!

Member itk::QuadEdgeMeshPolygonCell< TCellInterface >::GetBoundaryFeature (int dimension, CellFeatureIdentifier cellId, CellAutoPointer &cell)

Member itk::ResampleImageFilter< TInputImage, TOutputImage, TInterpolatorPrecisionType >::PrintSelf (std::ostream &os, Indent indent) const

Add details about this class

Class itk::SimpleFilterWatcher

Allow any stream object to be used for the output (not just std::cout)

Class itk::Statistics::ImageToListGenerator< TImage, TMaskImage >

In future allow the filter to take a Spatial object as input so a generic spatial object like an ellipse etc can be used as a mask. Sure the ImageMaskSpatialObject can represent image masks too, so why not make SpatialObjects the default. I think the ImageMaskSpatialObject is slow in terms of inefficient iteration through the image.

Member itk::StreamingImageIOBase::ReadBufferAsBinary (std::istream &is, void *buffer, SizeType num)

Move this method to itk::ImageIOBase

Member itk::StreamingImageIOBase::WriteBufferAsBinary (std::ostream &is, const void *buffer, SizeType num)

Move this methods to itk::ImageIOBase

Member itk::TransformToDeformationFieldSource< TOutputImage, TTransformPrecisionType >::PrintSelf (std::ostream &os, Indent indent) const

Add details about this class

Member itk::VectorResampleImageFilter< TInputImage, TOutputImage, TInterpolatorPrecisionType >::PrintSelf (std::ostream &os, Indent indent) const

Add details about this class

Member itk::VectorResampleImageFilter< TInputImage, TOutputImage, TInterpolatorPrecisionType >::TransformType

Check that input and output images have the same number of dimensions; this is required by the current implementation of AffineTransform.

Member itk::VoronoiSegmentationRGBImageFilter< TInputImage, TOutputImage >::SetMeanPercentError (double x[6])

Document.

Member itkQEAccessorsMacro (st, pt, dt)

Should this macro be added to doxygen macros?

Class otb::DefaultGeometriesToGeometriesFilter< TransformationFunctor, FieldTransformationPolicy >

Find a better name

Class otb::EllipsoidAdapter

{Add the support for different ellipsoid models}

Member otb::FieldCopyTransformation::m_SourceToDestFieldIndicesMap

std::map may not be the fastest structure available => use a (sorted?) vector of pairs of ints, and search with a simple std::find (as we can expect the number of fields to be quite low).

Member otb::GDALImageIO::InternalWriteImageInformation (const void *buffer)

Why not keys[itkey] == MetadataKey::MetadataKey ?

Member otb::GeometriesSource::DoAllocateOutputs ()

Check whether we need to increase the count of the GeometriesSet as it'll get reelased at the end of function the scope.

Member otb::GeometriesToGeometriesFilter::DoDefineNewLayerSpatialReference (ogr::Layer const &source) const

Return a unique_ptr<OGRSpatialReference>.

Class otb::ImageOfVectorsToMonoChannelExtractROI< TInputImage, TOutputImage >

Document this class

Class otb::ImageView< TViewerModel >

: Rename ImageViewer when refactoring will be completed.

Class otb::NCCRegistrationFunction< TFixedImage, TMovingImage, TDeformationField >

Check if the bug has been corrected in the next ITK version.

Class otb::ogr::ImageReference< TPrecision >

See how mix it with the otb::ogr::DataSource wrapper as it was with VectorData.

Class otb::ogr::Layer

Find a way to be notified when the related OGRDataSource is released

Member otb::ogr::Layer::AlterFieldDefn (vcl_size_t fieldIndex, FieldDefn const &newFieldDefn, int nFlags)

Move to use otb::ogr::FieldDefn

Member otb::ogr::Layer::CreateField (FieldDefn const &field, bool bApproxOK=true)

Move to use otb::ogr::FieldDefn

Class otb::ogr::StringListConverter

Have DataSource constructor receive a std::vector<std::string>, or even more precide types.

Class otb::PersistentStreamingStatisticsMapFromLabelImageFilter< TInputVectorImage, TLabelImage >

Implement other statistics (min, max, stddev...)

Class otb::PixelDescriptionView< TModel >

: Rename PixelDescriptioner when refactoring will be completed.

Class otb::StreamingStatisticsMapFromLabelImageFilter< TInputVectorImage, TLabelImage >

Implement other statistics (min, max, stddev...)

Reimplement as a multi-threaded filter

Class otb::TransformationFunctorDispatcher< TransformationFunctor, TransformedElementType, FieldTransformationPolicy >

Add a specialization for ogr::Feature.

Class otb::VectorDataKeywordlist

add the accessor to enable modifying/updating the data.

Class otb::VectorDataModel

all cases are not handled currently, this class need some additions.

Class otb::WaveletFilterBank< TInputImage, TOutputImage, TWaveletOperator, TDirectionOfTransformation >

: At present version, there is not consideration on meta data information that can be transmited from the input(s) to the output(s)...

Group Spatial filter property

We'll see later if a Geometry capsule is defined, or a nondeletable<> pointer type.