#include <itkMovingHistogramImageFilterBase.h>

Inheritance diagram for itk::MovingHistogramImageFilterBase< TInputImage, TOutputImage, TKernel >:

Collaboration diagram for itk::MovingHistogramImageFilterBase< TInputImage, TOutputImage, TKernel >:

Classes
class	DirectionCost
Public Types
typedef SmartPointer< const Self >	ConstPointer
typedef TInputImage::IndexType	IndexType
typedef TInputImage	InputImageType
typedef KernelType::ConstIterator	KernelIteratorType
typedef TKernel	KernelType
typedef std::list< OffsetType >	OffsetListType
typedef std::map< OffsetType, OffsetListType, typename OffsetType::LexicographicCompare >	OffsetMapType
typedef TInputImage::OffsetType	OffsetType
typedef Superclass::OutputImageRegionType	OutputImageRegionType
typedef TOutputImage	OutputImageType
typedef TOutputImage::PixelType	OutputPixelType
typedef TInputImage::PixelType	PixelType
typedef SmartPointer< Self >	Pointer
typedef KernelType::SizeType	RadiusType
typedef TInputImage::RegionType	RegionType
typedef MovingHistogramImageFilterBase	Self
typedef TInputImage::SizeType	SizeType
typedef KernelImageFilter < TInputImage, TOutputImage, TKernel >	Superclass
Public Member Functions
virtual ::itk::LightObject::Pointer	CreateAnother (void) const
virtual const char *	GetNameOfClass () const
virtual SizeValueType	GetPixelsPerTranslation () const
void	SetKernel (const KernelType &kernel)
Static Public Member Functions
static Pointer	New ()
Static Public Attributes
static const unsigned int	ImageDimension = TInputImage::ImageDimension
Protected Member Functions
void	GetDirAndOffset (const IndexType LineStart, const IndexType PrevLineStart, OffsetType &LineOffset, OffsetType &Changes, int &LineDirection)
	MovingHistogramImageFilterBase ()
void	PrintSelf (std::ostream &os, Indent indent) const
	~MovingHistogramImageFilterBase ()
Protected Attributes
OffsetMapType	m_AddedOffsets
FixedArray< int, itkGetStaticConstMacro(ImageDimension) >	m_Axes
OffsetListType	m_KernelOffsets
SizeValueType	m_PixelsPerTranslation
OffsetMapType	m_RemovedOffsets
Private Member Functions
	MovingHistogramImageFilterBase (const Self &)
void	operator= (const Self &)

Detailed Description

template<class TInputImage, class TOutputImage, class TKernel>
class itk::MovingHistogramImageFilterBase< TInputImage, TOutputImage, TKernel >

Implements a generic moving histogram algorithm.

This filter is a base class to implement efficiently many neighborhood filters. Instead of visiting all the neighbors of a pixel, the set of pixels in the neighborhood is updated when the filter is moving to a new pixel. The number of pixels read for each pixel can be very smaller than the number of pixels read by a basic algorithm.

This filter moves the neighborhood over all the pixels of the output requested region, and pass the pixel added and removed of the neighborhood to the an histogram class. This filter doesn't implement the histogram class - it must be implement and passed as template parameter. The histogram class is not necessary a real histogram. It can be implemented in many ways, and only has to provide the methods described below.

This filter takes 4 template parameters: the input and output image type, the structuring element (or kernel) type, and the histogram type. The input and output image must have the same number of dimension.

The histogram type is a class which has to implements seven methods: + a default constructor which takes no parameter. + void AddPixel( const InputPixelType &p ) is called when a new pixel is added to the histogram. + void RemovePixel( const InputPixelType &p ) is called when a pixel is removed of the histogram. + void AddBoundary() is called when a pixel outside the image is added. No value is provided: it's the responsability to the histogram class to get it if needed. This method can be kept empty to ignore the boundary pixels. + void RemoveBoundary() is called to when a pixel outside the image is removed. No value is provided: it's the responsability to the histogram class to get it if needed. This method can be kept empty to ignore the boundary pixels. + AType GetValue() is called to set the value of the output image. AType must be the output pixel type, or a type castable to the output pixel type.

MovingHistogramImageFilterBase add the new pixels before removing the old ones, so, if AddBoundary() is implemented and/or the kernel is symetric, it is safe to consider that the histogram will never be empty.

One histogram is created for each thread by the method NewHistogram(). The NewHistogram() method can be overiden to pass some parameters to the histogram.

The neighborhood is defined by a structuring element, and must a itk::Neighborhood object or a subclass. The structuring element is assumed to be composed of binary values (zero or one). Only elements of the structuring element having values > 0 are candidates for affecting the center pixel.

See also:: MovingWindowMeanImageFilter, RankImageFilter, MaskedMovingHistogramImageFitler,; MovingHistogramMorphologicalGradientImageFilter

Author:: Gaetan Lehmann; Richard Beare

Definition at line 90 of file itkMovingHistogramImageFilterBase.h.