#include <itkNarrowBandLevelSetImageFilter.h>

Inheritance diagram for itk::NarrowBandLevelSetImageFilter< TInputImage, TFeatureImage, TOutputPixelType, TOutputImage >:

Collaboration diagram for itk::NarrowBandLevelSetImageFilter< TInputImage, TFeatureImage, TOutputPixelType, TOutputImage >:

List of all members.

Public Types
typedef SmartPointer< const Self >	ConstPointer
typedef TFeatureImage	FeatureImageType
typedef Superclass::IndexType	IndexType
typedef Superclass::InputImageType	InputImageType
typedef TOutputImage	OutputImageType
typedef SmartPointer< Self >	Pointer
typedef SegmentationLevelSetFunction < OutputImageType, FeatureImageType >	SegmentationFunctionType
typedef NarrowBandLevelSetImageFilter	Self
typedef NarrowBandImageFilterBase < TInputImage, TOutputImage >	Superclass
typedef Superclass::TimeStepType	TimeStepType
typedef Superclass::ValueType	ValueType
typedef SegmentationFunctionType::VectorImageType	VectorImageType
Public Member Functions
virtual const SegmentationFunctionType::VectorImageType *	GetAdvectionImage () const
ValueType	GetAdvectionScaling () const
ValueType	GetCurvatureScaling () const
virtual FeatureImageType *	GetFeatureImage ()
unsigned int	GetMaximumIterations ()
virtual const char *	GetNameOfClass () const
ValueType	GetPropagationScaling () const
virtual SegmentationFunctionType *	GetSegmentationFunction ()
virtual const SegmentationFunctionType::ImageType *	GetSpeedImage () const
bool	GetUseNegativeFeatures () const
virtual void	SetInitialImage (InputImageType *f)
virtual void	SetMaximumRMSError (const double)
virtual void	SetSegmentationFunction (SegmentationFunctionType *s)
void	SetUseNegativeFeaturesOff ()
	typedef (Concept::HasNumericTraits< typename TOutputImage::PixelType >) OutputHasNumericTraitsCheck

virtual void	SetFeatureImage (const FeatureImageType *f)

void	SetUseNegativeFeaturesOn ()

void	SetUseNegativeFeatures (bool u)

virtual void	SetReverseExpansionDirection (bool _arg)
virtual bool	GetReverseExpansionDirection () const
virtual void	ReverseExpansionDirectionOn ()
virtual void	ReverseExpansionDirectionOff ()

void	SetFeatureScaling (ValueType v)

void	SetPropagationScaling (ValueType v)

void	SetAdvectionScaling (ValueType v)

void	SetCurvatureScaling (ValueType v)

void	SetMaximumIterations (unsigned int i)
Protected Types
typedef FastChamferDistanceImageFilter < OutputImageType, OutputImageType >	ChamferFilterType
typedef IsoContourDistanceImageFilter < OutputImageType, OutputImageType >	IsoFilterType
Protected Member Functions
virtual void	CreateNarrowBand ()
void	GenerateData ()
void	operator= (const Self &)
virtual void	PrintSelf (std::ostream &os, Indent indent) const

virtual	~NarrowBandLevelSetImageFilter ()
	NarrowBandLevelSetImageFilter ()
	NarrowBandLevelSetImageFilter (const Self &)

virtual void	InitializeIteration ()
Protected Attributes
ChamferFilterType::Pointer	m_ChamferFilter
IsoFilterType::Pointer	m_IsoFilter
bool	m_ReverseExpansionDirection
Private Attributes
SegmentationFunctionType *	m_SegmentationFunction

Detailed Description

template<class TInputImage, class TFeatureImage, class TOutputPixelType = float, class TOutputImage = Image< TOutputPixelType, ::itk::GetImageDimension< TInputImage >::ImageDimension >>
class itk::NarrowBandLevelSetImageFilter< TInputImage, TFeatureImage, TOutputPixelType, TOutputImage >

A base class which defines the API for implementing a special class of image segmentation filters using level set methods.

OVERVIEW: This object defines the framework for a class of segmentation filters which use level set methods. These filters work by constructing a ``feature image'' onto which the evolving level set locks as it moves. In the feature image, values that are close to zero are associated with object boundaries. An original (or preprocessed) image is given to the filter as the feature image and a seed for the level set is given as the input of the filter. The seed is converted into a level set embedding which propagates according to the features calculated from the original image.

TEMPLATE PARAMETERS: There are two required and two optional template parameter for these filters. Of the optional parameters, the last, TOutputImage, should not be changed from its default. It is only there to instantiate the parent class correctly.

TInputImage is the image type of the initial model you will input to the filter using SetInput() or SetInitialImage().

TFeatureImage is the image type of the image from which the filter will calculate the speed term for segmentation (see INPUTS).

TOutputPixelType is the data type used for the output image phi, the implicit level set image. This should really only ever be set as float (default) or double.

INPUTS: The input to any subclass of this filter is the seed image for the initial level set embedding. As with other subclasses of the SparseLevelSetImageFilter, the type of the input image is is not important. The (RequestedRegion) size of the seed image must, however, match the (RequestedRegion) size of the feature image.

You must identify the initial front (surface) in the input image. You do this by specifying its isovalue through the method SetIsoSurfaceValue(float f). The algorithm will then initialize its solution using the front represented by value f. Note that the front is always represented by isosurface zero in the output and not the isosurface you specified for the input. This is because, for simplicity, the filter will shift your input image so that the active front has zero values.

: Depending on the particular application and filter that you are using, the feature image should be preprocessed with some type of noise reduction filtering. The feature image input can be of any type, but it will be cast to floating point before calculations are done.

OUTPUTS: The output of any subclass of this filter is a level set embedding as described in SparseFieldLevelSetImageFilter. The zero crossings of the output image give the pixels closest to the level set boundary. By ITK convention, NEGATIVE values are pixels INSIDE the segmented region and POSITIVE values are pixels OUTSIDE the segmented region.

PARAMETERS: The MaximumRMSChange parameter is used to determine when the solution has converged. A lower value will result in a tighter-fitting solution, but will require more computations. Too low a value could put the solver into an infinite loop unless a reasonable NumberOfIterations parameter is set. Values should always be greater than 0.0 and less than 1.0.

: The NumberOfIterations parameter can be used to halt the solution after a specified number of iterations, overriding the MaximumRMSChange halting criteria.

: The standard convention for ITK level-set segmentation filters is that POSITIVE propagation (speed) and advection terms cause the surface to EXPAND while negative terms cause the surface to CONTRACT. When the ReverseExpansionDirection parameter is set to TRUE (on), it tells the function object to reverse the standard ITK convention so that NEGATIVE terms cause EXPANSION and positive terms cause CONTRACTION.

This parameter can be safely changed as appropriate for a particular application or data set to achieve the desired behavior.

: The FeatureScaling parameter controls the magnitude of the features calculated for use in the level set propagation and advection speeds. This value simply sets both parameters to equal values at once. Some filters may only use on of these two terms and this method is a generic way to set either or both without having to know which is in use.

: The CurvatureScaling parameter controls the magnitude of the curvature values which are calculated on the evolving isophote. This is important in controlling the relative effect of curvature in the calculation. Default value is 1.0. Higher values relative to the other level set equation terms (propagation and advection) will give a smoother result.

: The PropagationScaling parameter controls the scaling of the scalar propagation (speed) term relative to other terms in the level set equation. Setting this value will override any value already set by FeatureScaling.

: The AdvectionScaling parameter controls the scaling of the vector advection field term relative to other terms in the level set equation. Setting this value will override any value already set by FeatureScaling.

: See LevelSetFunction for more information.

Definition at line 147 of file itkNarrowBandLevelSetImageFilter.h.