Implementation of the N4 bias field correction algorithm. More...

#include <sitkN4BiasFieldCorrectionImageFilter.h>

Inheritance diagram for itk::simple::N4BiasFieldCorrectionImageFilter:

Collaboration diagram for itk::simple::N4BiasFieldCorrectionImageFilter:

Public Types
typedef RealPixelIDTypeList	PixelIDTypeList

typedef N4BiasFieldCorrectionImageFilter	Self

Public Types inherited from itk::simple::ImageFilter< 0 >
typedef ImageFilter	Self

Public Types inherited from itk::simple::ProcessObject
typedef ProcessObject	Self

Public Member Functions
Image	Execute (const Image &image, const Image &maskImage)

Image	Execute (const Image &image, const Image &maskImage, double convergenceThreshold, std::vector< uint32_t > maximumNumberOfIterations, double biasFieldFullWidthAtHalfMaximum, double wienerFilterNoise, uint32_t numberOfHistogramBins, const std::vector< uint32_t > &numberOfControlPoints, uint32_t splineOrder)

double	GetBiasFieldFullWidthAtHalfMaximum () const

double	GetConvergenceThreshold () const

std::vector< uint32_t >	GetMaximumNumberOfIterations () const

std::string	GetName () const

std::vector< uint32_t >	GetNumberOfControlPoints () const

uint32_t	GetNumberOfHistogramBins () const

uint32_t	GetSplineOrder () const

double	GetWienerFilterNoise () const

	N4BiasFieldCorrectionImageFilter ()

Self &	SetBiasFieldFullWidthAtHalfMaximum (double BiasFieldFullWidthAtHalfMaximum)

Self &	SetConvergenceThreshold (double ConvergenceThreshold)

Self &	SetMaximumNumberOfIterations (std::vector< uint32_t > MaximumNumberOfIterations)

Self &	SetNumberOfControlPoints (const std::vector< uint32_t > &NumberOfControlPoints)

Self &	SetNumberOfControlPoints (uint32_t value)

Self &	SetNumberOfHistogramBins (uint32_t NumberOfHistogramBins)

Self &	SetSplineOrder (uint32_t SplineOrder)

Self &	SetWienerFilterNoise (double WienerFilterNoise)

std::string	ToString () const

	~N4BiasFieldCorrectionImageFilter ()

Public Member Functions inherited from itk::simple::ImageFilter< 0 >
	ImageFilter ()

virtual	~ImageFilter ()=0

Public Member Functions inherited from itk::simple::ProcessObject
virtual void	Abort ()

virtual int	AddCommand (itk::simple::EventEnum event, itk::simple::Command &cmd)
	Add a Command Object to observer the event. More...

virtual float	GetProgress () const
	An Active Measurement of the progress of execution. More...

virtual bool	HasCommand (itk::simple::EventEnum event) const
	Query of this object has any registered commands for event. More...

	ProcessObject ()

virtual void	RemoveAllCommands ()
	Remove all registered commands. More...

virtual	~ProcessObject ()

virtual void	DebugOn ()

virtual void	DebugOff ()

virtual bool	GetDebug () const

virtual void	SetDebug (bool debugFlag)

virtual void	SetNumberOfThreads (unsigned int n)

virtual unsigned int	GetNumberOfThreads () const

Private Types
typedef Image(Self::*	MemberFunctionType) (const Image image, const Image maskImage)

Private Member Functions
template<class TImageType1 , class TImageType2 >
Image	DualExecuteInternal (const Image image, const Image maskImage)

Private Attributes
double	m_BiasFieldFullWidthAtHalfMaximum

double	m_ConvergenceThreshold

nsstd::auto_ptr< detail::DualMemberFunctionFactory< MemberFunctionType > >	m_DualMemberFactory

std::vector< uint32_t >	m_MaximumNumberOfIterations

std::vector< uint32_t >	m_NumberOfControlPoints

uint32_t	m_NumberOfHistogramBins

uint32_t	m_SplineOrder

double	m_WienerFilterNoise

Friends
struct	detail::DualExecuteInternalAddressor< MemberFunctionType >

Additional Inherited Members
Static Public Member Functions inherited from itk::simple::ProcessObject
static bool	GetGlobalDefaultDebug ()

static void	GlobalDefaultDebugOff ()

static void	GlobalDefaultDebugOn ()

static void	SetGlobalDefaultDebug (bool debugFlag)

static void	GlobalWarningDisplayOn ()

static void	GlobalWarningDisplayOff ()

static void	SetGlobalWarningDisplay (bool flag)

static bool	GetGlobalWarningDisplay ()

static void	SetGlobalDefaultNumberOfThreads (unsigned int n)

static unsigned int	GetGlobalDefaultNumberOfThreads ()

static double	GetGlobalDefaultCoordinateTolerance ()
	Access the global tolerance to determine congruent spaces. More...

static void	SetGlobalDefaultCoordinateTolerance (double)
	Access the global tolerance to determine congruent spaces. More...

static double	GetGlobalDefaultDirectionTolerance ()
	Access the global tolerance to determine congruent spaces. More...

static void	SetGlobalDefaultDirectionTolerance (double)
	Access the global tolerance to determine congruent spaces. More...

Protected Member Functions inherited from itk::simple::ProcessObject
virtual unsigned long	AddITKObserver (const itk::EventObject &, itk::Command *)

virtual itk::ProcessObject *	GetActiveProcess ()

virtual void	OnActiveProcessDelete ()

virtual void	onCommandDelete (const itk::simple::Command *cmd) throw ()

virtual void	PreUpdate (itk::ProcessObject *p)

virtual void	RemoveITKObserver (EventCommand &e)

Protected Member Functions inherited from itk::simple::NonCopyable
	NonCopyable ()

Static Protected Member Functions inherited from itk::simple::ImageFilter< 0 >
static void	FixNonZeroIndex (TImageType *img)

Static Protected Member Functions inherited from itk::simple::ProcessObject
template<class TImageType >
static TImageType::ConstPointer	CastImageToITK (const Image &img)

template<class TImageType >
static Image	CastITKToImage (TImageType *img)

template<class TPixelType , unsigned int VImageDimension, unsigned int VLength, template< typename, unsigned int > class TVector>
static Image	CastITKToImage (itk::Image< TVector< TPixelType, VLength >, VImageDimension > *img)

static const itk::EventObject &	GetITKEventObject (EventEnum e)

template<typename T >
static std::ostream &	ToStringHelper (std::ostream &os, const T &v)

static std::ostream &	ToStringHelper (std::ostream &os, const char &v)

static std::ostream &	ToStringHelper (std::ostream &os, const signed char &v)

static std::ostream &	ToStringHelper (std::ostream &os, const unsigned char &v)

Detailed Description

Implementation of the N4 bias field correction algorithm.

The nonparametric nonuniform intensity normalization (N3) algorithm, as introduced by Sled et al. in 1998 is a method for correcting nonuniformity associated with MR images. The algorithm assumes a simple parametric model (Gaussian) for the bias field and does not require tissue class segmentation. In addition, there are only a couple of parameters to tune with the default values performing quite well. N3 has been publicly available as a set of perl scripts (http://www.bic.mni.mcgill.ca/ServicesSoftwareAdvancedImageProcessingTools/HomePage )

The N4 algorithm, encapsulated with this class, is a variation of the original N3 algorithm with the additional benefits of an improved B-spline fitting routine which allows for multiple resolutions to be used during the correction process. We also modify the iterative update component of algorithm such that the residual bias field is continually updated

Notes for the user:

Since much of the image manipulation is done in the log space of the intensities, input images with negative and small values (< 1) can produce poor results.

The original authors recommend performing the bias field correction on a downsampled version of the original image.

A binary mask or a weighted image can be supplied. If a binary mask is specified, those voxels in the input image which correspond to the voxels in the mask image are used to estimate the bias field. If a UseMaskLabel value is set to true, only voxels in the MaskImage that match the MaskLabel will be used; otherwise, all non-zero voxels in the MaskImage will be masked. If a confidence image is specified, the input voxels are weighted in the b-spline fitting routine according to the confidence voxel values.

The filter returns the corrected image. If the bias field is wanted, one can reconstruct it using the class itkBSplineControlPointImageFilter. See the IJ article and the test file for an example.

The 'Z' parameter in Sled's 1998 paper is the square root of the class variable 'm_WienerFilterNoise'.

The basic algorithm iterates between sharpening the intensity histogram of the corrected input image and spatially smoothing those results with a B-spline scalar field estimate of the bias field.

Author: Nicholas J. Tustison

Contributed by Nicholas J. Tustison, James C. Gee in the Insight Journal paper: https://hdl.handle.net/10380/3053

REFERENCE

J.G. Sled, A.P. Zijdenbos and A.C. Evans. "A Nonparametric Method for Automatic Correction of Intensity Nonuniformity in Data" IEEE Transactions on Medical Imaging, Vol 17, No 1. Feb 1998.

N.J. Tustison, B.B. Avants, P.A. Cook, Y. Zheng, A. Egan, P.A. Yushkevich, and J.C. Gee. "N4ITK: Improved N3 Bias Correction" IEEE Transactions on Medical Imaging, 29(6):1310-1320, June 2010.

See also: itk::simple::N4BiasFieldCorrection for the procedural interface; itk::N4BiasFieldCorrectionImageFilter for the Doxygen on the original ITK class.

Definition at line 72 of file sitkN4BiasFieldCorrectionImageFilter.h.

Member Typedef Documentation

typedef Image(Self::* itk::simple::N4BiasFieldCorrectionImageFilter::MemberFunctionType) (const Image *image, const Image *maskImage)

private

Setup for member function dispatching

Definition at line 178 of file sitkN4BiasFieldCorrectionImageFilter.h.

typedef RealPixelIDTypeList itk::simple::N4BiasFieldCorrectionImageFilter::PixelIDTypeList

Define the pixels types supported by this filter

Definition at line 84 of file sitkN4BiasFieldCorrectionImageFilter.h.

typedef N4BiasFieldCorrectionImageFilter itk::simple::N4BiasFieldCorrectionImageFilter::Self

Definition at line 74 of file sitkN4BiasFieldCorrectionImageFilter.h.

Constructor & Destructor Documentation

itk::simple::N4BiasFieldCorrectionImageFilter::N4BiasFieldCorrectionImageFilter ( )

Default Constructor that takes no arguments and initializes default parameters

itk::simple::N4BiasFieldCorrectionImageFilter::~N4BiasFieldCorrectionImageFilter ( )

Destructor

Member Function Documentation

template<class TImageType1 , class TImageType2 >

Image itk::simple::N4BiasFieldCorrectionImageFilter::DualExecuteInternal	(	const Image *	image,
		const Image *	maskImage
	)

private

Image itk::simple::N4BiasFieldCorrectionImageFilter::Execute	(	const Image &	image,
		const Image &	maskImage
	)

Execute the filter on the input image

Image itk::simple::N4BiasFieldCorrectionImageFilter::Execute	(	const Image &	image,
		const Image &	maskImage,
		double	convergenceThreshold,
		std::vector< uint32_t >	maximumNumberOfIterations,
		double	biasFieldFullWidthAtHalfMaximum,
		double	wienerFilterNoise,
		uint32_t	numberOfHistogramBins,
		const std::vector< uint32_t > &	numberOfControlPoints,
		uint32_t	splineOrder
	)

Execute the filter on the input image with the given parameters

double itk::simple::N4BiasFieldCorrectionImageFilter::GetBiasFieldFullWidthAtHalfMaximum ( ) const

inline

Get the full width at half maximum parameter characterizing the width of the Gaussian deconvolution. Default = 0.15.

Definition at line 116 of file sitkN4BiasFieldCorrectionImageFilter.h.

double itk::simple::N4BiasFieldCorrectionImageFilter::GetConvergenceThreshold ( ) const

inline

Get the convergence threshold. Convergence is determined by the coefficient of variation of the difference image between the current bias field estimate and the previous estimate. If this value is less than the specified threshold, the algorithm proceeds to the next fitting level or terminates if it is at the last level.

Definition at line 96 of file sitkN4BiasFieldCorrectionImageFilter.h.

std::vector<uint32_t> itk::simple::N4BiasFieldCorrectionImageFilter::GetMaximumNumberOfIterations ( ) const

inline

Get the maximum number of iterations specified at each fitting level. Default = 50.

Definition at line 106 of file sitkN4BiasFieldCorrectionImageFilter.h.

std::string itk::simple::N4BiasFieldCorrectionImageFilter::GetName ( ) const

inlinevirtual

Name of this class

Implements itk::simple::ProcessObject.

Definition at line 162 of file sitkN4BiasFieldCorrectionImageFilter.h.

std::vector<uint32_t> itk::simple::N4BiasFieldCorrectionImageFilter::GetNumberOfControlPoints ( ) const

inline

Get the control point grid size defining the B-spline estimate of the scalar bias field. In each dimension, the B-spline mesh size is equal to the number of control points in that dimension minus the spline order. Default = 4 control points in each dimension for a mesh size of 1 in each dimension.

Definition at line 150 of file sitkN4BiasFieldCorrectionImageFilter.h.

uint32_t itk::simple::N4BiasFieldCorrectionImageFilter::GetNumberOfHistogramBins ( ) const

inline

Get number of bins defining the log input intensity histogram. Default = 200.

Definition at line 136 of file sitkN4BiasFieldCorrectionImageFilter.h.

uint32_t itk::simple::N4BiasFieldCorrectionImageFilter::GetSplineOrder ( ) const

inline

Get the spline order defining the bias field estimate. Default = 3.

Definition at line 160 of file sitkN4BiasFieldCorrectionImageFilter.h.

double itk::simple::N4BiasFieldCorrectionImageFilter::GetWienerFilterNoise ( ) const

inline

Get the noise estimate defining the Wiener filter. Default = 0.01.

Definition at line 126 of file sitkN4BiasFieldCorrectionImageFilter.h.

Self& itk::simple::N4BiasFieldCorrectionImageFilter::SetBiasFieldFullWidthAtHalfMaximum ( double BiasFieldFullWidthAtHalfMaximum )

inline

Set the full width at half maximum parameter characterizing the width of the Gaussian deconvolution. Default = 0.15.

Definition at line 111 of file sitkN4BiasFieldCorrectionImageFilter.h.

Self& itk::simple::N4BiasFieldCorrectionImageFilter::SetConvergenceThreshold ( double ConvergenceThreshold )

inline

Set the convergence threshold. Convergence is determined by the coefficient of variation of the difference image between the current bias field estimate and the previous estimate. If this value is less than the specified threshold, the algorithm proceeds to the next fitting level or terminates if it is at the last level.

Definition at line 91 of file sitkN4BiasFieldCorrectionImageFilter.h.

Self& itk::simple::N4BiasFieldCorrectionImageFilter::SetMaximumNumberOfIterations ( std::vector< uint32_t > MaximumNumberOfIterations )

inline

Set the maximum number of iterations specified at each fitting level. Default = 50.

Definition at line 101 of file sitkN4BiasFieldCorrectionImageFilter.h.

Self& itk::simple::N4BiasFieldCorrectionImageFilter::SetNumberOfControlPoints ( const std::vector< uint32_t > & NumberOfControlPoints )

inline

Set the control point grid size defining the B-spline estimate of the scalar bias field. In each dimension, the B-spline mesh size is equal to the number of control points in that dimension minus the spline order. Default = 4 control points in each dimension for a mesh size of 1 in each dimension.

Definition at line 141 of file sitkN4BiasFieldCorrectionImageFilter.h.

Self& itk::simple::N4BiasFieldCorrectionImageFilter::SetNumberOfControlPoints ( uint32_t value )

inline

Set the values of the NumberOfControlPoints vector all to value

Definition at line 144 of file sitkN4BiasFieldCorrectionImageFilter.h.

Self& itk::simple::N4BiasFieldCorrectionImageFilter::SetNumberOfHistogramBins ( uint32_t NumberOfHistogramBins )

inline

Set number of bins defining the log input intensity histogram. Default = 200.

Definition at line 131 of file sitkN4BiasFieldCorrectionImageFilter.h.

Self& itk::simple::N4BiasFieldCorrectionImageFilter::SetSplineOrder ( uint32_t SplineOrder )

inline

Set the spline order defining the bias field estimate. Default = 3.