Uses a flooding algorithm to set all voxels that are not a regional extrema to the max or min of the pixel type. More...

#include <itkValuedRegionalExtremaImageFilter.h>

Inheritance diagram for itk::ValuedRegionalExtremaImageFilter< TInputImage, TOutputImage, TFunction1, TFunction2 >:

Collaboration diagram for itk::ValuedRegionalExtremaImageFilter< TInputImage, TOutputImage, TFunction1, TFunction2 >:

List of all members.

Public Types
typedef SmartPointer< const Self >	ConstPointer
typedef InputImageType::ConstPointer	InputImageConstPointer
typedef InputImageType::PixelType	InputImagePixelType
typedef InputImageType::Pointer	InputImagePointer
typedef InputImageType::RegionType	InputImageRegionType
typedef TInputImage	InputImageType
typedef InputImageType::SizeType	ISizeType
typedef OutputImageType::ConstPointer	OutputImageConstPointer
typedef OutputImageType::PixelType	OutputImagePixelType
typedef OutputImageType::Pointer	OutputImagePointer
typedef OutputImageType::RegionType	OutputImageRegionType
typedef TOutputImage	OutputImageType
typedef SmartPointer< Self >	Pointer
typedef ValuedRegionalExtremaImageFilter	Self
typedef ImageToImageFilter < TInputImage, TOutputImage >	Superclass
Public Member Functions
virtual ::itk::LightObject::Pointer	CreateAnother (void) const
virtual bool	GetFlat () const
virtual const char *	GetNameOfClass () const
	typedef (Concept::HasNumericTraits< InputImagePixelType >) InputHasNumericTraitsCheck
	typedef (Concept::HasPixelTraits< InputImagePixelType >) InputHasPixelTraitsCheck

virtual void	SetFullyConnected (bool _arg)
virtual const bool &	GetFullyConnected ()
virtual void	FullyConnectedOn ()
virtual void	FullyConnectedOff ()

virtual void	SetMarkerValue (typename TInputImage::PixelType _arg)
virtual const TInputImage::PixelType &	GetMarkerValue ()
Static Public Member Functions
static Pointer	New ()
Static Public Attributes

static const unsigned int	InputImageDimension = TInputImage::ImageDimension
static const unsigned int	OutputImageDimension = TOutputImage::ImageDimension
Protected Member Functions
void	EnlargeOutputRequestedRegion (DataObject *)
void	GenerateData ()
void	GenerateInputRequestedRegion ()

	ValuedRegionalExtremaImageFilter ()
	~ValuedRegionalExtremaImageFilter ()
void	PrintSelf (std::ostream &os, Indent indent) const
Private Types
typedef ConstShapedNeighborhoodIterator < InputImageType >	ConstInputIterator
typedef std::stack< OutIndexType >	IndexStack
typedef InputImageType::IndexType	InIndexType
typedef ShapedNeighborhoodIterator < OutputImageType >	NOutputIterator
typedef OutputImageType::IndexType	OutIndexType
Private Member Functions
void	operator= (const Self &)
	ValuedRegionalExtremaImageFilter (const Self &)
Private Attributes
bool	m_Flat
bool	m_FullyConnected
TInputImage::PixelType	m_MarkerValue

Detailed Description

template<class TInputImage, class TOutputImage, class TFunction1, class TFunction2>
class itk::ValuedRegionalExtremaImageFilter< TInputImage, TOutputImage, TFunction1, TFunction2 >

Uses a flooding algorithm to set all voxels that are not a regional extrema to the max or min of the pixel type.

This is the class used by ValuedRegionalMinimaImageFilter and ValuedRegionalMaximaImageFilter. There is no supression of regional minima based on dynamics, as available in HMinimaImageFilter. This flooding algorithm is a very simple one, but I'm not sure where it came from - I certainly didn't invent it.

Let's consider the case of regional minima. The basic algorithm is: Boundary conditions are such that the image is logically surrounded by a border that is either maximal or minimal for the pixel type. An optimized version could explicitly set the border to avoid the need for boundary checks. For regional minima the boundary is set to the maximal value for the pixel type.

Pixels are visited in raster order. The neighbors of each pixel are examined. If any neighbor is greater than the centre, then the centre pixel cannot be a regional minima. The centre pixel is part of a flat region (consisting of at least one pixel) that is therefore not a regional minima either. This region is set to the maximum value for the pixel type using a flooding algorithm.

There are some minor complications that prevent pixels being examined more than once -- basically check that the output value is less than the maximum for the pixel type.

The implementation uses the functor model from itkMaximumImageFilter.

This code was contributed in the Insight Journal paper: "Finding regional extrema - methods and performance" by Beare R., Lehmann G. http://hdl.handle.net/1926/153 http://www.insight-journal.org/browse/publication/65

Author:: Richard Beare. Department of Medicine, Monash University, Melbourne, Australia.

See also:: ValuedRegionalMinimaImageFilter, ValuedRegionalMaximaImageFilter,; HMinimaImageFilter

Definition at line 77 of file itkValuedRegionalExtremaImageFilter.h.