#include <rtkADMMTotalVariationConjugateGradientOperator.h>

Inheritance diagram for rtk::ADMMTotalVariationConjugateGradientOperator< TOutputImage >:

Collaboration diagram for rtk::ADMMTotalVariationConjugateGradientOperator< TOutputImage >:

Public Types
using	BackProjectionFilterPointer = typename BackProjectionFilterType::Pointer

using	BackProjectionFilterType = rtk::BackProjectionImageFilter< TOutputImage, TOutputImage >

using	DisplacedDetectorFilterType = rtk::DisplacedDetectorImageFilter< TOutputImage >

using	DivergenceFilterType = rtk::BackwardDifferenceDivergenceImageFilter< GradientImageType, TOutputImage >

using	ForwardProjectionFilterPointer = typename ForwardProjectionFilterType::Pointer

using	ForwardProjectionFilterType = rtk::ForwardProjectionImageFilter< TOutputImage, TOutputImage >

using	GatingWeightsFilterType = rtk::MultiplyByVectorImageFilter< TOutputImage >

using	GradientFilterType = ForwardDifferenceGradientImageFilter< TOutputImage, typename TOutputImage::ValueType, typename TOutputImage::ValueType, GradientImageType >

using	GradientImageType = typename TOutputImage::template RebindImageType< VectorPixelType, TOutputImage::ImageDimension >

using	MultiplyFilterType = itk::MultiplyImageFilter< TOutputImage >

using	Pointer = itk::SmartPointer< Self >

using	Self = ADMMTotalVariationConjugateGradientOperator

using	SubtractFilterType = itk::SubtractImageFilter< TOutputImage >

using	Superclass = ConjugateGradientOperator< TOutputImage >

using	VectorPixelType = itk::CovariantVector< typename TOutputImage::ValueType, TOutputImage::ImageDimension >

Public Types inherited from rtk::ConjugateGradientOperator< TOutputImage >
using	Pointer = itk::SmartPointer< Self >

using	Self = ConjugateGradientOperator

using	Superclass = itk::ImageToImageFilter< TOutputImage, TOutputImage >

Public Member Functions
virtual ::itk::LightObject::Pointer	CreateAnother () const

const char *	GetNameOfClass () const override

void	SetBackProjectionFilter (BackProjectionFilterType *_arg)

virtual void	SetBeta (float _arg)

void	SetForwardProjectionFilter (ForwardProjectionFilterType *_arg)

void	SetGatingWeights (std::vector< float > weights)

virtual void	SetGeometry (ThreeDCircularProjectionGeometry *_arg)


virtual void	SetDisableDisplacedDetectorFilter (bool _arg)

virtual bool	GetDisableDisplacedDetectorFilter ()

Public Member Functions inherited from rtk::ConjugateGradientOperator< TOutputImage >
virtual ::itk::LightObject::Pointer	CreateAnother () const

const char *	GetNameOfClass () const override

virtual void	SetX (const TOutputImage *OutputImage)

Static Public Member Functions
static Pointer	New ()

Static Public Member Functions inherited from rtk::ConjugateGradientOperator< TOutputImage >
static Pointer	New ()

Protected Member Functions
	ADMMTotalVariationConjugateGradientOperator ()

void	GenerateData () override

void	VerifyInputInformation () const override

void	VerifyPreconditions () const override

	~ADMMTotalVariationConjugateGradientOperator () override=default


void	GenerateInputRequestedRegion () override

void	GenerateOutputInformation () override

Protected Member Functions inherited from rtk::ConjugateGradientOperator< TOutputImage >
	ConjugateGradientOperator ()

	~ConjugateGradientOperator () override=default

Protected Attributes
BackProjectionFilterPointer	m_BackProjectionFilter

float	m_Beta

bool	m_DisableDisplacedDetectorFilter

DisplacedDetectorFilterType::Pointer	m_DisplacedDetectorFilter

DivergenceFilterType::Pointer	m_DivergenceFilter

ForwardProjectionFilterPointer	m_ForwardProjectionFilter

std::vector< float >	m_GatingWeights

GatingWeightsFilterType::Pointer	m_GatingWeightsFilter

ThreeDCircularProjectionGeometry::Pointer	m_Geometry

GradientFilterType::Pointer	m_GradientFilter

bool	m_IsGated

MultiplyFilterType::Pointer	m_MultiplyFilter

SubtractFilterType::Pointer	m_SubtractFilter

MultiplyFilterType::Pointer	m_ZeroMultiplyProjectionFilter

MultiplyFilterType::Pointer	m_ZeroMultiplyVolumeFilter

Detailed Description

template<typename TOutputImage>
class rtk::ADMMTotalVariationConjugateGradientOperator< TOutputImage >

Implements the operator A used in the conjugate gradient step of ADMM reconstruction with total variation regularization.

This filter implements the operator A used in the conjugate gradient step of a reconstruction method based on compressed sensing. The method attempts to find the f that minimizes || sqrt(D) (Rf -p) ||_2^2 + alpha * TV(f), with R the forward projection operator, p the measured projections, D the displaced detector weighting matrix, and TV the total variation. Details on the method and the calculations can be found in

Mory, C., B. Zhang, V. Auvray, M. Grass, D. Schafer, F. Peyrin, S. Rit, P. Douek, and L. Boussel. "ECG-Gated C-Arm Computed Tomography Using L1 Regularization." In Proceedings of the 20th European Signal Processing Conference (EUSIPCO), 2728-32, 2012.

This filter takes in input f and outputs R_t D R f - beta div(grad(f)).

Test:: rtkadmmtotalvariationtest.cxx

Author: Cyril Mory

Definition at line 104 of file rtkADMMTotalVariationConjugateGradientOperator.h.