#include <itkSparseFieldFourthOrderLevelSetImageFilter.h>

Inheritance diagram for itk::SparseFieldFourthOrderLevelSetImageFilter< TInputImage, TOutputImage >:

Collaboration diagram for itk::SparseFieldFourthOrderLevelSetImageFilter< TInputImage, TOutputImage >:

Public Types
typedef SmartPointer< const Self >	ConstPointer
typedef Superclass::IndexType	IndexType
typedef Superclass::LayerType	LayerType
typedef LevelSetFunctionWithRefitTerm < OutputImageType, SparseImageType >	LevelSetFunctionType
typedef Superclass::NeighborhoodScalesType	NeighborhoodScalesType
typedef NormalBandNode < OutputImageType >	NodeType
typedef ImplicitManifoldNormalVectorFilter < OutputImageType, SparseImageType >	NormalVectorFilterType
typedef NormalVectorDiffusionFunction < SparseImageType >	NormalVectorFunctionType
typedef NodeType::NodeDataType	NormalVectorType
typedef Superclass::OutputImageType	OutputImageType
typedef SmartPointer< Self >	Pointer
typedef Superclass::RadiusType	RadiusType
typedef SparseFieldFourthOrderLevelSetImageFilter	Self
typedef NeighborhoodIterator < SparseImageType >	SparseImageIteratorType
typedef SparseImage< NodeType, itkGetStaticConstMacro(ImageDimension) >	SparseImageType
typedef SparseFieldLevelSetImageFilter < TInputImage, TOutputImage >	Superclass
typedef Superclass::ValueType	ValueType
Public Member Functions
virtual const ValueType &	GetCurvatureBandWidth ()
virtual const unsigned int &	GetMaxNormalIteration ()
virtual const unsigned int &	GetMaxRefitIteration ()
unsigned int	GetMinimumNumberOfLayers () const
virtual const char *	GetNameOfClass () const
virtual const ValueType &	GetNormalProcessConductance ()
virtual const int &	GetNormalProcessType ()
virtual const bool &	GetNormalProcessUnsharpFlag ()
virtual const ValueType &	GetNormalProcessUnsharpWeight ()
virtual const ValueType &	GetRMSChangeNormalProcessTrigger ()
virtual void	SetCurvatureBandWidth (ValueType _arg)
void	SetLevelSetFunction (LevelSetFunctionType *lsf)
virtual void	SetMaxNormalIteration (unsigned int _arg)
virtual void	SetMaxRefitIteration (unsigned int _arg)
virtual void	SetNormalProcessConductance (ValueType _arg)
virtual void	SetNormalProcessType (int _arg)
virtual void	SetNormalProcessUnsharpFlag (bool _arg)
virtual void	SetNormalProcessUnsharpWeight (ValueType _arg)
virtual void	SetNumberOfLayers (const unsigned int n)
virtual void	SetRMSChangeNormalProcessTrigger (ValueType _arg)
Static Public Attributes
static const unsigned int	ImageDimension = Superclass::ImageDimension
Protected Member Functions
bool	ActiveLayerCheckBand () const
ValueType	ComputeCurvatureFromSparseImageNeighborhood (SparseImageIteratorType &neighborhood) const
void	ComputeCurvatureTarget (const OutputImageType distanceImage, SparseImageType sparseImage) const
void	ProcessNormals ()
Private Member Functions
void	operator= (const Self &)
	SparseFieldFourthOrderLevelSetImageFilter (const Self &)
Private Attributes
bool	m_ConvergenceFlag
ValueType	m_CurvatureBandWidth
LevelSetFunctionType *	m_LevelSetFunction
unsigned int	m_MaxNormalIteration
unsigned int	m_MaxRefitIteration
ValueType	m_NormalProcessConductance
int	m_NormalProcessType
bool	m_NormalProcessUnsharpFlag
ValueType	m_NormalProcessUnsharpWeight
unsigned int	m_RefitIteration
ValueType	m_RMSChangeNormalProcessTrigger
Static Private Attributes
static const ValueType	m_DimConst
static const SizeValueType	m_NumVertex
virtual void	InitializeIteration ()
	typedef (Concept::HasNumericTraits< ValueType >) OutputHasNumericTraitsCheck
	SparseFieldFourthOrderLevelSetImageFilter ()
	~SparseFieldFourthOrderLevelSetImageFilter ()
virtual void	PrintSelf (std::ostream &os, Indent indent) const

Detailed Description

template<class TInputImage, class TOutputImage>
class itk::SparseFieldFourthOrderLevelSetImageFilter< TInputImage, TOutputImage >

This class implements the fourth order level set PDE framework.

: This class adds a ProcessNormals method to SparseFieldLevelSetImageFilter class. The ProcessNormals method uses the ImplicitManifoldNormalDiffusionFilter class to generate a SparseImage of filtered normal vectors. We make a copy of the current state of the output image (also referred to as level set image) for this class and pass it to ImplicitManifoldNormalDiffusionFilter. That class computes the normal vectors to the level set image and filters them. The output is in the form of a sparse image templated with the NormalBandNode type. We then compute curvatures from that output and store them in the SparseImage as well. This SparseImage is passed onto the LevelSetFunctionWithRefitTerm filter class to be used as a target in the propagation term.

INPUT and OUTPUT: Same as SparseFieldLevelSetImageFilter

PARAMETERS: MaxRefitIteration sets the maximum number of allowable iterations between calls to ProcessNormals. The decision of when to call the ProcessNormals method is made in InitializeIteration according to a few criteria one of which is this maximum number of iterations.

: MaxNormalIteration sets the maximum number of diffusion iterations on the normals to be performed by the ImplicitManifoldNormalDiffusionFilter class. Please read the documentation for that class.

: CurvatureBandWidth determines the width of the band to be processed in ImplicitManifoldNormalDiffusionFilter.

: RMSChangeNormalProcessTrigger provides another mechanism in InitializeIteration for calling the ProcessNormals method. Whenever the RMS change reported by SparseFieldLevelSetImageFilter falls below this parameter ProcessNormals is called regardless of whether MaxRefitIteration has been reached. This parameter could be used to speed up the algorithm; however, it can also effect the results. Use with caution. Default is 0 which does nothing.

IMPORTANT: Defaults for above parameters are set in the constructor. Users should not change these unless they have a good understanding of the algorithm.

OTHER PARAMETERS: NormalProcessType tells ImplicitManifoldNormalVectorFilter whether to use isotropic or anisotropic diffusion. A value of 0 means isotropic whereas a value of 1 means anisotropic diffusion. If this parameter is set to 1, NormalProcessConductance determines the level of detail preservation. Please read the documentation for ImplicitManifoldNormalVectorFilter and AnisotropicFourthOrderLevelSetImageFilter.

: NormalProcessUnsharpFlag turns unsharp masking on/off. If this parameter is turned on, then NormalProcessUnsharpWeight should be set. Please read the documentation for ImplicitManifoldNormalVectorFilter.

IMPORTANT: Users of this class must define the Halt function.

Definition at line 157 of file itkSparseFieldFourthOrderLevelSetImageFilter.h.