#include <itkGradientDescentOptimizerv4.h>

Detailed Description

template<typename TInternalComputationValueType>
class itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >

Gradient descent optimizer.

GradientDescentOptimizer implements a simple gradient descent optimizer. At each iteration the current position is updated according to

\[ p_{n+1} = p_n + \mbox{learningRate} \, \frac{\partial f(p_n) }{\partial p_n} \]

Optionally, the best metric value and matching parameters can be stored and retried via GetValue() and GetCurrentPosition(). See SetReturnBestParametersAndValue().

Gradient scales can be manually set or automatically estimated, as documented in the base class. The learning rate defaults to 1.0, and can be set in two ways: 1) manually, via SetLearningRate(). Or, 2) automatically, either at each iteration or only at the first iteration, by assigning a ScalesEstimator via SetScalesEstimator(). When a ScalesEstimator is assigned, the optimizer is enabled by default to estimate learning rate only once, during the first iteration. This behavior can be changed via SetDoEstimateLearningRateAtEveryIteration() and SetDoEstimateLearningRateOnce(). For learning rate to be estimated at each iteration, the user must call SetDoEstimateLearningRateAtEveryIteration(true) and SetDoEstimateLearningRateOnce(false). When enabled, the optimizer computes learning rate(s) such that at each step, each voxel's change in physical space will be less than m_MaximumStepSizeInPhysicalUnits.

 m_LearningRate =
   m_MaximumStepSizeInPhysicalUnits /
   m_ScalesEstimator->EstimateStepScale(scaledGradient)

where m_MaximumStepSizeInPhysicalUnits defaults to the voxel spacing returned by m_ScalesEstimator->EstimateMaximumStepSize() (which is typically 1 voxel), and can be set by the user via SetMaximumStepSizeInPhysicalUnits(). When SetDoEstimateLearningRateOnce is enabled, the voxel change may become being greater than m_MaximumStepSizeInPhysicalUnits in later iterations.

Note: Unlike the previous version of GradientDescentOptimizer, this version does not have a "maximize/minimize" option to modify the effect of the metric derivative. The assigned metric is assumed to return a parameter derivative result that "improves" the optimization when added to the current parameters via the metric::UpdateTransformParameters method, after the optimizer applies scales and a learning rate.

Examples: Examples/RegistrationITKv4/MultiStageImageRegistration1.cxx, and Examples/RegistrationITKv4/MultiStageImageRegistration2.cxx.

Definition at line 78 of file itkGradientDescentOptimizerv4.h.

Inheritance diagram for itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >:

Collaboration diagram for itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >:

Public Types
using	ConstPointer = SmartPointer< const Self >

using	InternalComputationValueType = TInternalComputationValueType

using	Pointer = SmartPointer< Self >

using	Self = GradientDescentOptimizerv4Template

using	Superclass = GradientDescentOptimizerBasev4Template< TInternalComputationValueType >

Public Types inherited from itk::GradientDescentOptimizerBasev4Template< TInternalComputationValueType >
using	ConstPointer = SmartPointer< const Self >

using	ConvergenceMonitoringType = itk::Function::WindowConvergenceMonitoringFunction< TInternalComputationValueType >

using	IndexRangeType = ThreadedIndexedContainerPartitioner::IndexRangeType

using	InternalComputationValueType = TInternalComputationValueType

using	MetricTypePointer = typename MetricType::Pointer

using	Pointer = SmartPointer< Self >

using	Self = GradientDescentOptimizerBasev4Template

using	Superclass = ObjectToObjectOptimizerBaseTemplate< TInternalComputationValueType >

Public Types inherited from itk::ObjectToObjectOptimizerBaseTemplate< TInternalComputationValueType >
using	ConstPointer = SmartPointer< const Self >

using	DerivativeType = typename MetricType::DerivativeType

using	MeasureType = typename MetricType::MeasureType

using	MetricType = ObjectToObjectMetricBaseTemplate< TInternalComputationValueType >

using	MetricTypePointer = typename MetricType::Pointer

using	NumberOfParametersType = typename MetricType::NumberOfParametersType

using	ParametersType = OptimizerParameters< TInternalComputationValueType >

using	Pointer = SmartPointer< Self >

using	ScalesEstimatorType = OptimizerParameterScalesEstimatorTemplate< TInternalComputationValueType >

using	ScalesType = OptimizerParameters< TInternalComputationValueType >

using	Self = ObjectToObjectOptimizerBaseTemplate

using	StopConditionDescriptionType = std::ostringstream

using	StopConditionReturnStringType = std::string

using	Superclass = Object

Public Types inherited from itk::Object
using	ConstPointer = SmartPointer< const Self >

using	Pointer = SmartPointer< Self >

using	Self = Object

using	Superclass = LightObject

Public Types inherited from itk::LightObject
using	ConstPointer = SmartPointer< const Self >

using	Pointer = SmartPointer< Self >

using	Self = LightObject

Public Member Functions
virtual void	EstimateLearningRate ()

virtual const TInternalComputationValueType &	GetConvergenceValue () const

const char *	GetNameOfClass () const override

void	ResumeOptimization () override

virtual void	SetConvergenceWindowSize (SizeValueType _arg)

virtual void	SetMinimumConvergenceValue (TInternalComputationValueType _arg)

void	StartOptimization (bool doOnlyInitialization=false) override

void	StopOptimization () override


virtual void	SetLearningRate (TInternalComputationValueType _arg)

virtual const TInternalComputationValueType &	GetLearningRate () const


virtual void	SetMaximumStepSizeInPhysicalUnits (TInternalComputationValueType _arg)

virtual const TInternalComputationValueType &	GetMaximumStepSizeInPhysicalUnits () const


virtual void	SetDoEstimateLearningRateAtEachIteration (bool _arg)

virtual const bool &	GetDoEstimateLearningRateAtEachIteration () const

virtual void	DoEstimateLearningRateAtEachIterationOn ()


virtual void	SetDoEstimateLearningRateOnce (bool _arg)

virtual const bool &	GetDoEstimateLearningRateOnce () const

virtual void	DoEstimateLearningRateOnceOn ()


virtual void	SetReturnBestParametersAndValue (bool _arg)

virtual const bool &	GetReturnBestParametersAndValue () const

virtual void	ReturnBestParametersAndValueOn ()

Public Member Functions inherited from itk::GradientDescentOptimizerBasev4Template< TInternalComputationValueType >
virtual const DerivativeType &	GetGradient () const

virtual const StopConditionObjectToObjectOptimizerEnum &	GetStopCondition () const

StopConditionReturnStringType	GetStopConditionDescription () const override

virtual void	ModifyGradientByScales ()

virtual void	ModifyGradientByLearningRate ()

Public Member Functions inherited from itk::ObjectToObjectOptimizerBaseTemplate< TInternalComputationValueType >
virtual bool	CanUseScales () const

virtual SizeValueType	GetCurrentIteration () const

virtual const MeasureType &	GetCurrentMetricValue () const

virtual const ParametersType &	GetCurrentPosition () const

const char *	GetNameOfClass () const override

virtual SizeValueType	GetNumberOfIterations () const

virtual const ThreadIdType &	GetNumberOfWorkUnits () const

virtual const ScalesType &	GetScales () const

virtual const bool &	GetScalesAreIdentity () const

bool	GetScalesInitialized () const

virtual const MeasureType &	GetValue () const

virtual const ScalesType &	GetWeights () const

virtual const bool &	GetWeightsAreIdentity () const

virtual void	SetNumberOfIterations (SizeValueType _arg)

virtual void	SetNumberOfWorkUnits (ThreadIdType number)

virtual void	SetScalesEstimator (ScalesEstimatorType *_arg)

virtual void	SetWeights (ScalesType _arg)

virtual void	SetMetric (MetricType *_arg)

virtual MetricType *	GetModifiableMetric ()

virtual void	SetScales (const ScalesType &scales)

virtual void	SetDoEstimateScales (bool _arg)

virtual const bool &	GetDoEstimateScales () const

virtual void	DoEstimateScalesOn ()

Public Member Functions inherited from itk::Object
unsigned long	AddObserver (const EventObject &event, Command *cmd) const

unsigned long	AddObserver (const EventObject &event, std::function< void(const EventObject &)> function) const

LightObject::Pointer	CreateAnother () const override

virtual void	DebugOff () const

virtual void	DebugOn () const

Command *	GetCommand (unsigned long tag)

bool	GetDebug () const

MetaDataDictionary &	GetMetaDataDictionary ()

const MetaDataDictionary &	GetMetaDataDictionary () const

virtual ModifiedTimeType	GetMTime () const

virtual const TimeStamp &	GetTimeStamp () const

bool	HasObserver (const EventObject &event) const

void	InvokeEvent (const EventObject &)

void	InvokeEvent (const EventObject &) const

virtual void	Modified () const

void	Register () const override

void	RemoveAllObservers ()

void	RemoveObserver (unsigned long tag) const

void	SetDebug (bool debugFlag) const

void	SetReferenceCount (int) override

void	UnRegister () const noexcept override

void	SetMetaDataDictionary (const MetaDataDictionary &rhs)

void	SetMetaDataDictionary (MetaDataDictionary &&rrhs)

virtual void	SetObjectName (std::string _arg)

virtual const std::string &	GetObjectName () const

Public Member Functions inherited from itk::LightObject
Pointer	Clone () const

virtual void	Delete ()

virtual int	GetReferenceCount () const

void	Print (std::ostream &os, Indent indent=0) const

Static Public Member Functions
static Pointer	New ()

Static Public Member Functions inherited from itk::Object
static bool	GetGlobalWarningDisplay ()

static void	GlobalWarningDisplayOff ()

static void	GlobalWarningDisplayOn ()

static Pointer	New ()

static void	SetGlobalWarningDisplay (bool val)

Static Public Member Functions inherited from itk::LightObject
static void	BreakOnError ()

static Pointer	New ()

Protected Member Functions
virtual void	AdvanceOneStep ()

	GradientDescentOptimizerv4Template ()

void	ModifyGradientByLearningRateOverSubRange (const IndexRangeType &subrange) override

void	ModifyGradientByScalesOverSubRange (const IndexRangeType &subrange) override

void	PrintSelf (std::ostream &os, Indent indent) const override

	~GradientDescentOptimizerv4Template () override=default

Protected Member Functions inherited from itk::GradientDescentOptimizerBasev4Template< TInternalComputationValueType >
	GradientDescentOptimizerBasev4Template ()

	~GradientDescentOptimizerBasev4Template () override=default

Protected Member Functions inherited from itk::ObjectToObjectOptimizerBaseTemplate< TInternalComputationValueType >
void	PrintSelf (std::ostream &os, Indent indent) const override

	ObjectToObjectOptimizerBaseTemplate ()

	~ObjectToObjectOptimizerBaseTemplate () override

Protected Member Functions inherited from itk::Object
	Object ()

bool	PrintObservers (std::ostream &os, Indent indent) const

virtual void	SetTimeStamp (const TimeStamp &timeStamp)

	~Object () override

Protected Member Functions inherited from itk::LightObject
virtual LightObject::Pointer	InternalClone () const

	LightObject ()

virtual void	PrintHeader (std::ostream &os, Indent indent) const

virtual void	PrintTrailer (std::ostream &os, Indent indent) const

virtual	~LightObject ()

Protected Attributes
ParametersType	m_BestParameters {}

TInternalComputationValueType	m_ConvergenceValue {}

MeasureType	m_CurrentBestValue {}

TInternalComputationValueType	m_LearningRate {}

TInternalComputationValueType	m_MinimumConvergenceValue {}

DerivativeType	m_PreviousGradient {}

bool	m_ReturnBestParametersAndValue { false }

Protected Attributes inherited from itk::GradientDescentOptimizerBasev4Template< TInternalComputationValueType >
ConvergenceMonitoringType::Pointer	m_ConvergenceMonitoring {}

SizeValueType	m_ConvergenceWindowSize {}

bool	m_DoEstimateLearningRateAtEachIteration {}

bool	m_DoEstimateLearningRateOnce {}

DerivativeType	m_Gradient {}

TInternalComputationValueType	m_MaximumStepSizeInPhysicalUnits {}

DomainThreader< ThreadedIndexedContainerPartitioner, Self >::Pointer	m_ModifyGradientByLearningRateThreader {}

DomainThreader< ThreadedIndexedContainerPartitioner, Self >::Pointer	m_ModifyGradientByScalesThreader {}

bool	m_Stop { false }

StopConditionObjectToObjectOptimizerEnum	m_StopCondition {}

StopConditionDescriptionType	m_StopConditionDescription {}

bool	m_UseConvergenceMonitoring {}

Protected Attributes inherited from itk::ObjectToObjectOptimizerBaseTemplate< TInternalComputationValueType >
SizeValueType	m_CurrentIteration {}

MeasureType	m_CurrentMetricValue {}

bool	m_DoEstimateScales {}

MetricTypePointer	m_Metric {}

SizeValueType	m_NumberOfIterations {}

ThreadIdType	m_NumberOfWorkUnits {}

ScalesType	m_Scales {}

bool	m_ScalesAreIdentity {}

ScalesEstimatorType::Pointer	m_ScalesEstimator {}

ScalesType	m_Weights {}

bool	m_WeightsAreIdentity {}

Protected Attributes inherited from itk::LightObject
std::atomic< int >	m_ReferenceCount {}

Member Typedef Documentation

◆ ConstPointer

template<typename TInternalComputationValueType>

using itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::ConstPointer = SmartPointer<const Self>

Definition at line 89 of file itkGradientDescentOptimizerv4.h.

◆ InternalComputationValueType

template<typename TInternalComputationValueType>

using itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::InternalComputationValueType = TInternalComputationValueType

It should be possible to derive the internal computation type from the class object.

Definition at line 99 of file itkGradientDescentOptimizerv4.h.

◆ Pointer

template<typename TInternalComputationValueType>

using itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::Pointer = SmartPointer<Self>

Definition at line 88 of file itkGradientDescentOptimizerv4.h.

◆ Self

template<typename TInternalComputationValueType>

using itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::Self = GradientDescentOptimizerv4Template

Standard class type aliases.

Definition at line 86 of file itkGradientDescentOptimizerv4.h.

◆ Superclass

template<typename TInternalComputationValueType>

using itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::Superclass = GradientDescentOptimizerBasev4Template<TInternalComputationValueType>

Definition at line 87 of file itkGradientDescentOptimizerv4.h.

Constructor & Destructor Documentation

◆ GradientDescentOptimizerv4Template()

template<typename TInternalComputationValueType>

itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::GradientDescentOptimizerv4Template ( )

protected

Default constructor

◆ ~GradientDescentOptimizerv4Template()

template<typename TInternalComputationValueType>

itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::~GradientDescentOptimizerv4Template ( )

overrideprotecteddefault

Destructor

Member Function Documentation

◆ AdvanceOneStep()

template<typename TInternalComputationValueType>

virtual void itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::AdvanceOneStep ( )

protectedvirtual

Advance one step following the gradient direction. Includes transform update.

Reimplemented in itk::LBFGS2Optimizerv4Template< TInternalComputationValueType >, itk::QuasiNewtonOptimizerv4Template< TInternalComputationValueType >, itk::RegularStepGradientDescentOptimizerv4< TInternalComputationValueType >, itk::GradientDescentLineSearchOptimizerv4Template< TInternalComputationValueType >, and itk::ConjugateGradientLineSearchOptimizerv4Template< TInternalComputationValueType >.

◆ DoEstimateLearningRateAtEachIterationOn()

template<typename TInternalComputationValueType>

virtual void itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::DoEstimateLearningRateAtEachIterationOn ( )

virtual

Option to use ScalesEstimator for learning rate estimation at each iteration. The estimation overrides the learning rate set by SetLearningRate(). Default is false.

See also: SetDoEstimateLearningRateOnce(); SetScalesEstimator()

◆ DoEstimateLearningRateOnceOn()

template<typename TInternalComputationValueType>

virtual void itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::DoEstimateLearningRateOnceOn ( )

virtual

Option to use ScalesEstimator for learning rate estimation only once, during first iteration. The estimation overrides the learning rate set by SetLearningRate(). Default is true.

See also: SetDoEstimateLearningRateAtEachIteration(); SetScalesEstimator()

◆ EstimateLearningRate()

template<typename TInternalComputationValueType>

virtual void itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::EstimateLearningRate ( )

virtual

Estimate the learning rate based on the current gradient.

Reimplemented in itk::RegularStepGradientDescentOptimizerv4< TInternalComputationValueType >.

◆ GetConvergenceValue()

template<typename TInternalComputationValueType>

virtual const TInternalComputationValueType& itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::GetConvergenceValue ( ) const

virtual

Get current convergence value. WindowConvergenceMonitoringFunction always returns output convergence value in 'TInternalComputationValueType' precision.

Reimplemented in itk::LBFGS2Optimizerv4Template< TInternalComputationValueType >.

◆ GetDoEstimateLearningRateAtEachIteration()

template<typename TInternalComputationValueType>

virtual const bool& itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::GetDoEstimateLearningRateAtEachIteration ( ) const

virtual

Option to use ScalesEstimator for learning rate estimation at each iteration. The estimation overrides the learning rate set by SetLearningRate(). Default is false.

See also: SetDoEstimateLearningRateOnce(); SetScalesEstimator()

◆ GetDoEstimateLearningRateOnce()

template<typename TInternalComputationValueType>

virtual const bool& itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::GetDoEstimateLearningRateOnce ( ) const

virtual

Option to use ScalesEstimator for learning rate estimation only once, during first iteration. The estimation overrides the learning rate set by SetLearningRate(). Default is true.

See also: SetDoEstimateLearningRateAtEachIteration(); SetScalesEstimator()

◆ GetLearningRate()

template<typename TInternalComputationValueType>

virtual const TInternalComputationValueType& itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::GetLearningRate ( ) const

virtual

Set/Get the learning rate to apply. It is overridden by automatic learning rate estimation if enabled. See main documentation.

◆ GetMaximumStepSizeInPhysicalUnits()

template<typename TInternalComputationValueType>

virtual const TInternalComputationValueType& itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::GetMaximumStepSizeInPhysicalUnits ( ) const

virtual

Set/Get the maximum step size, in physical space units.

 Only relevant when m_ScalesEstimator is set by user,
 and automatic learning rate estimation is enabled.
 See main documentation.

◆ GetNameOfClass()

template<typename TInternalComputationValueType>

const char* itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::GetNameOfClass ( ) const

overridevirtual

See also: LightObject::GetNameOfClass()

Reimplemented from itk::GradientDescentOptimizerBasev4Template< TInternalComputationValueType >.

Reimplemented in itk::LBFGS2Optimizerv4Template< TInternalComputationValueType >, itk::QuasiNewtonOptimizerv4Template< TInternalComputationValueType >, itk::MultiGradientOptimizerv4Template< TInternalComputationValueType >, and itk::RegularStepGradientDescentOptimizerv4< TInternalComputationValueType >.

◆ GetReturnBestParametersAndValue()

template<typename TInternalComputationValueType>

virtual const bool& itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::GetReturnBestParametersAndValue ( ) const

virtual

Flag. Set to have the optimizer track and return the best best metric value and corresponding best parameters that were calculated during the optimization. This captures the best solution when the optimizer oversteps or oscillates near the end of an optimization. Results are stored in m_CurrentMetricValue and in the assigned metric's parameters, retrievable via optimizer->GetCurrentPosition(). This option requires additional memory to store the best parameters, which can be large when working with high-dimensional transforms such as DisplacementFieldTransform.

◆ ModifyGradientByLearningRateOverSubRange()

template<typename TInternalComputationValueType>

void itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::ModifyGradientByLearningRateOverSubRange ( const IndexRangeType & subrange )

overrideprotectedvirtual

Modify the gradient by learning rate over a given index range.

Implements itk::GradientDescentOptimizerBasev4Template< TInternalComputationValueType >.

Reimplemented in itk::RegularStepGradientDescentOptimizerv4< TInternalComputationValueType >.

◆ ModifyGradientByScalesOverSubRange()

template<typename TInternalComputationValueType>

void itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::ModifyGradientByScalesOverSubRange ( const IndexRangeType & subrange )

overrideprotectedvirtual

Modify the gradient by scales and weights over a given index range.

Implements itk::GradientDescentOptimizerBasev4Template< TInternalComputationValueType >.

Reimplemented in itk::RegularStepGradientDescentOptimizerv4< TInternalComputationValueType >.

◆ New()

template<typename TInternalComputationValueType>

static Pointer itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::New ( )

static

New macro for creation of through a Smart Pointer

◆ PrintSelf()

template<typename TInternalComputationValueType>

void itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::PrintSelf	(	std::ostream &	os,
		Indent	indent
	)		const

overrideprotectedvirtual

Methods invoked by Print() to print information about the object including superclasses. Typically not called by the user (use Print() instead) but used in the hierarchical print process to combine the output of several classes.

Reimplemented from itk::GradientDescentOptimizerBasev4Template< TInternalComputationValueType >.

Reimplemented in itk::LBFGS2Optimizerv4Template< TInternalComputationValueType >, itk::QuasiNewtonOptimizerv4Template< TInternalComputationValueType >, itk::RegularStepGradientDescentOptimizerv4< TInternalComputationValueType >, and itk::MultiGradientOptimizerv4Template< TInternalComputationValueType >.

◆ ResumeOptimization()

template<typename TInternalComputationValueType>

void itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::ResumeOptimization ( )

overridevirtual

Resume the optimization.

Implements itk::GradientDescentOptimizerBasev4Template< TInternalComputationValueType >.

Reimplemented in itk::LBFGS2Optimizerv4Template< TInternalComputationValueType >, and itk::MultiGradientOptimizerv4Template< TInternalComputationValueType >.

◆ ReturnBestParametersAndValueOn()

template<typename TInternalComputationValueType>

virtual void itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::ReturnBestParametersAndValueOn ( )

virtual

Flag. Set to have the optimizer track and return the best best metric value and corresponding best parameters that were calculated during the optimization. This captures the best solution when the optimizer oversteps or oscillates near the end of an optimization. Results are stored in m_CurrentMetricValue and in the assigned metric's parameters, retrievable via optimizer->GetCurrentPosition(). This option requires additional memory to store the best parameters, which can be large when working with high-dimensional transforms such as DisplacementFieldTransform.

◆ SetConvergenceWindowSize()

template<typename TInternalComputationValueType>

virtual void itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::SetConvergenceWindowSize ( SizeValueType _arg )

virtual

Window size for the convergence checker. The convergence checker calculates convergence value by fitting to a window of the energy (metric value) profile.

The default m_ConvergenceWindowSize is set to 50 to pass all tests. It is suggested to use 10 for less stringent convergence checking.

Reimplemented in itk::LBFGS2Optimizerv4Template< TInternalComputationValueType >.

◆ SetDoEstimateLearningRateAtEachIteration()

template<typename TInternalComputationValueType>

virtual void itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::SetDoEstimateLearningRateAtEachIteration ( bool _arg )

virtual

Option to use ScalesEstimator for learning rate estimation at each iteration. The estimation overrides the learning rate set by SetLearningRate(). Default is false.

See also: SetDoEstimateLearningRateOnce(); SetScalesEstimator()

◆ SetDoEstimateLearningRateOnce()

template<typename TInternalComputationValueType>

virtual void itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::SetDoEstimateLearningRateOnce ( bool _arg )

virtual

Option to use ScalesEstimator for learning rate estimation only once, during first iteration. The estimation overrides the learning rate set by SetLearningRate(). Default is true.

See also: SetDoEstimateLearningRateAtEachIteration(); SetScalesEstimator()

◆ SetLearningRate()

template<typename TInternalComputationValueType>

virtual void itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::SetLearningRate ( TInternalComputationValueType _arg )

virtual

Set/Get the learning rate to apply. It is overridden by automatic learning rate estimation if enabled. See main documentation.

◆ SetMaximumStepSizeInPhysicalUnits()

template<typename TInternalComputationValueType>

virtual void itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::SetMaximumStepSizeInPhysicalUnits ( TInternalComputationValueType _arg )

virtual

Set/Get the maximum step size, in physical space units.

 Only relevant when m_ScalesEstimator is set by user,
 and automatic learning rate estimation is enabled.
 See main documentation.

◆ SetMinimumConvergenceValue()

template<typename TInternalComputationValueType>

virtual void itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::SetMinimumConvergenceValue ( TInternalComputationValueType _arg )

virtual

Minimum convergence value for convergence checking. The convergence checker calculates convergence value by fitting to a window of the energy profile. When the convergence value reaches a small value, it would be treated as converged.

The default m_MinimumConvergenceValue is set to 1e-8 to pass all tests. It is suggested to use 1e-6 for less stringent convergence checking.

◆ SetReturnBestParametersAndValue()

template<typename TInternalComputationValueType>

virtual void itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::SetReturnBestParametersAndValue ( bool _arg )

virtual

Flag. Set to have the optimizer track and return the best best metric value and corresponding best parameters that were calculated during the optimization. This captures the best solution when the optimizer oversteps or oscillates near the end of an optimization. Results are stored in m_CurrentMetricValue and in the assigned metric's parameters, retrievable via optimizer->GetCurrentPosition(). This option requires additional memory to store the best parameters, which can be large when working with high-dimensional transforms such as DisplacementFieldTransform.

◆ StartOptimization()

template<typename TInternalComputationValueType>

void itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::StartOptimization ( bool doOnlyInitialization = false )

overridevirtual

Start and run the optimization.

Reimplemented from itk::GradientDescentOptimizerBasev4Template< TInternalComputationValueType >.

Reimplemented in itk::LBFGS2Optimizerv4Template< TInternalComputationValueType >, itk::RegularStepGradientDescentOptimizerv4< TInternalComputationValueType >, itk::MultiGradientOptimizerv4Template< TInternalComputationValueType >, and itk::QuasiNewtonOptimizerv4Template< TInternalComputationValueType >.

◆ StopOptimization()

template<typename TInternalComputationValueType>

void itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::StopOptimization ( )

overridevirtual

Stop the optimization.

Reimplemented from itk::GradientDescentOptimizerBasev4Template< TInternalComputationValueType >.

Reimplemented in itk::MultiGradientOptimizerv4Template< TInternalComputationValueType >.

Member Data Documentation

◆ m_BestParameters

template<typename TInternalComputationValueType>

ParametersType itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::m_BestParameters {}

protected

Definition at line 240 of file itkGradientDescentOptimizerv4.h.

◆ m_ConvergenceValue

template<typename TInternalComputationValueType>

TInternalComputationValueType itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::m_ConvergenceValue {}

protected

Definition at line 236 of file itkGradientDescentOptimizerv4.h.

◆ m_CurrentBestValue

template<typename TInternalComputationValueType>

MeasureType itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::m_CurrentBestValue {}

protected

Store the best value and related parameters.

Definition at line 239 of file itkGradientDescentOptimizerv4.h.

◆ m_LearningRate

template<typename TInternalComputationValueType>

TInternalComputationValueType itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::m_LearningRate {}

protected

Definition at line 234 of file itkGradientDescentOptimizerv4.h.

◆ m_MinimumConvergenceValue

template<typename TInternalComputationValueType>

TInternalComputationValueType itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::m_MinimumConvergenceValue {}

protected

Definition at line 235 of file itkGradientDescentOptimizerv4.h.

◆ m_PreviousGradient

template<typename TInternalComputationValueType>

DerivativeType itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::m_PreviousGradient {}

protected

Store the previous gradient value at each iteration, so we can detect the changes in gradient direction. This is needed by the regular step gradient descent and Quasi Newton optimizers.

Definition at line 249 of file itkGradientDescentOptimizerv4.h.

◆ m_ReturnBestParametersAndValue

template<typename TInternalComputationValueType>

bool itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >::m_ReturnBestParametersAndValue { false }

protected

Definition at line 242 of file itkGradientDescentOptimizerv4.h.

The documentation for this class was generated from the following file:

itkGradientDescentOptimizerv4.h

Detailed Description

template<typename TInternalComputationValueType> class itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >

Public Types

Public Member Functions

Static Public Member Functions

Protected Member Functions

Protected Attributes

Member Typedef Documentation

◆ ConstPointer

◆ InternalComputationValueType

◆ Pointer

◆ Self

◆ Superclass

Constructor & Destructor Documentation

◆ GradientDescentOptimizerv4Template()

◆ ~GradientDescentOptimizerv4Template()

Member Function Documentation

◆ AdvanceOneStep()

◆ DoEstimateLearningRateAtEachIterationOn()

◆ DoEstimateLearningRateOnceOn()

◆ EstimateLearningRate()

◆ GetConvergenceValue()

◆ GetDoEstimateLearningRateAtEachIteration()

◆ GetDoEstimateLearningRateOnce()

◆ GetLearningRate()

◆ GetMaximumStepSizeInPhysicalUnits()

◆ GetNameOfClass()

◆ GetReturnBestParametersAndValue()

◆ ModifyGradientByLearningRateOverSubRange()

◆ ModifyGradientByScalesOverSubRange()

◆ New()

◆ PrintSelf()

◆ ResumeOptimization()

◆ ReturnBestParametersAndValueOn()

◆ SetConvergenceWindowSize()

◆ SetDoEstimateLearningRateAtEachIteration()

◆ SetDoEstimateLearningRateOnce()

◆ SetLearningRate()

◆ SetMaximumStepSizeInPhysicalUnits()

◆ SetMinimumConvergenceValue()

◆ SetReturnBestParametersAndValue()

◆ StartOptimization()

◆ StopOptimization()

Member Data Documentation

◆ m_BestParameters

◆ m_ConvergenceValue

◆ m_CurrentBestValue

◆ m_LearningRate

◆ m_MinimumConvergenceValue

◆ m_PreviousGradient

◆ m_ReturnBestParametersAndValue

template<typename TInternalComputationValueType>
class itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >