Gradient descent optimizer. More...

#include <itkGradientDescentOptimizerv4.h>

Inheritance diagram for itk::GradientDescentOptimizerv4:

Collaboration diagram for itk::GradientDescentOptimizerv4:

Public Types
typedef SmartPointer< const Self >	ConstPointer
typedef Superclass::DerivativeType	DerivativeType
typedef Superclass::InternalComputationValueType	InternalComputationValueType
typedef Superclass::MeasureType	MeasureType
typedef SmartPointer< Self >	Pointer
typedef GradientDescentOptimizerv4	Self
typedef GradientDescentOptimizerBasev4	Superclass
Public Member Functions
virtual ::itk::LightObject::Pointer	CreateAnother (void) const
virtual const InternalComputationValueType &	GetLearningRate ()
virtual const char *	GetNameOfClass () const
virtual void	ResumeOptimization ()
virtual void	SetLearningRate (InternalComputationValueType _arg)
virtual void	SetMaximumStepSizeInPhysicalUnits (InternalComputationValueType _arg)
virtual void	SetScalesEstimator (OptimizerParameterScalesEstimator *_arg)
virtual void	StartOptimization ()
Static Public Member Functions
static Pointer	New ()
Protected Member Functions
virtual void	AdvanceOneStep (void)
virtual void	EstimateLearningRate ()
	GradientDescentOptimizerv4 ()
virtual void	PrintSelf (std::ostream &os, Indent indent) const
virtual	~GradientDescentOptimizerv4 ()

virtual void	ModifyGradientByScalesOverSubRange (const IndexRangeType &subrange)
virtual void	ModifyGradientByLearningRateOverSubRange (const IndexRangeType &subrange)
Protected Attributes
InternalComputationValueType	m_LearningRate
InternalComputationValueType	m_MaximumStepSizeInPhysicalUnits
OptimizerParameterScalesEstimator::Pointer	m_ScalesEstimator
Private Member Functions
	GradientDescentOptimizerv4 (const Self &)
void	operator= (const Self &)

Detailed Description

Gradient descent optimizer.

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

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

The user can scale each component of the df / dp but setting a scaling vector using method SetScales().

The learning rate defaults to 1.0, and can be set via SetLearningRate.

The user may set a member m_ScalesEstimator by calling SetScalesEstimator() before optimization to estimate scales and learning rates automatically.

When m_ScalesEstimator is set, m_MaximumStepSizeInPhysicalUnits may also be set by the user to change the maximum step size at each iteration. Learning rates are automatically restricted such that each step will produce physical impacts on voxels less than m_MaximumStepSizeInPhysicalUnits. m_MaximumStepSizeInPhysicalUnits defaults to the voxel spacing returned by m_ScalesEstimator.

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.

Definition at line 62 of file itkGradientDescentOptimizerv4.h.

Member Typedef Documentation

typedef SmartPointer< const Self > itk::GradientDescentOptimizerv4::ConstPointer

Reimplemented from itk::GradientDescentOptimizerBasev4.

Reimplemented in itk::QuasiNewtonOptimizerv4.

Definition at line 70 of file itkGradientDescentOptimizerv4.h.

typedef Superclass::DerivativeType itk::GradientDescentOptimizerv4::DerivativeType

Derivative type

Reimplemented from itk::GradientDescentOptimizerBasev4.

Definition at line 76 of file itkGradientDescentOptimizerv4.h.

typedef Superclass::InternalComputationValueType itk::GradientDescentOptimizerv4::InternalComputationValueType

Internal computation type, for maintaining a desired precision

Reimplemented from itk::GradientDescentOptimizerBasev4.

Reimplemented in itk::QuasiNewtonOptimizerv4.

Definition at line 83 of file itkGradientDescentOptimizerv4.h.

typedef Superclass::MeasureType itk::GradientDescentOptimizerv4::MeasureType

Metric type over which this class is templated

Reimplemented from itk::GradientDescentOptimizerBasev4.

Definition at line 82 of file itkGradientDescentOptimizerv4.h.

typedef SmartPointer< Self > itk::GradientDescentOptimizerv4::Pointer

Reimplemented from itk::GradientDescentOptimizerBasev4.

Reimplemented in itk::QuasiNewtonOptimizerv4.

Definition at line 69 of file itkGradientDescentOptimizerv4.h.

typedef GradientDescentOptimizerv4 itk::GradientDescentOptimizerv4::Self

Standard class typedefs.

Reimplemented from itk::GradientDescentOptimizerBasev4.

Reimplemented in itk::QuasiNewtonOptimizerv4.

Definition at line 67 of file itkGradientDescentOptimizerv4.h.

typedef GradientDescentOptimizerBasev4 itk::GradientDescentOptimizerv4::Superclass

Reimplemented from itk::GradientDescentOptimizerBasev4.

Reimplemented in itk::QuasiNewtonOptimizerv4.

Definition at line 68 of file itkGradientDescentOptimizerv4.h.

Constructor & Destructor Documentation

itk::GradientDescentOptimizerv4::GradientDescentOptimizerv4 ( ) [protected]

Default constructor

virtual itk::GradientDescentOptimizerv4::~GradientDescentOptimizerv4 ( ) [protected, virtual]

Destructor

itk::GradientDescentOptimizerv4::GradientDescentOptimizerv4 ( const Self & ) [private]

Member Function Documentation

virtual void itk::GradientDescentOptimizerv4::AdvanceOneStep ( void ) [protected, virtual]

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

Reimplemented in itk::QuasiNewtonOptimizerv4.

virtual::itk::LightObject::Pointer itk::GradientDescentOptimizerv4::CreateAnother ( void ) const [virtual]

Create an object from an instance, potentially deferring to a factory. This method allows you to create an instance of an object that is exactly the same type as the referring object. This is useful in cases where an object has been cast back to a base class.

Reimplemented from itk::Object.

Reimplemented in itk::QuasiNewtonOptimizerv4.

virtual void itk::GradientDescentOptimizerv4::EstimateLearningRate ( ) [protected, virtual]

Estimate the learning rate

Implements itk::GradientDescentOptimizerBasev4.

virtual const InternalComputationValueType& itk::GradientDescentOptimizerv4::GetLearningRate ( ) [virtual]

Get the learning rate.

virtual const char* itk::GradientDescentOptimizerv4::GetNameOfClass ( ) const [virtual]

Run-time type information (and related methods).

Reimplemented from itk::GradientDescentOptimizerBasev4.

Reimplemented in itk::QuasiNewtonOptimizerv4.

virtual void itk::GradientDescentOptimizerv4::ModifyGradientByLearningRateOverSubRange ( const IndexRangeType & subrange ) [protected, virtual]

Modify the gradient over a given index range.

Implements itk::GradientDescentOptimizerBasev4.

virtual void itk::GradientDescentOptimizerv4::ModifyGradientByScalesOverSubRange ( const IndexRangeType & subrange ) [protected, virtual]

Modify the gradient over a given index range.

Implements itk::GradientDescentOptimizerBasev4.

static Pointer itk::GradientDescentOptimizerv4::New ( ) [static]

New macro for creation of through a Smart Pointer

Reimplemented from itk::Object.

Reimplemented in itk::QuasiNewtonOptimizerv4.

void itk::GradientDescentOptimizerv4::operator= ( const Self & ) [private]

Mutex lock to protect modification to the reference count

Reimplemented from itk::GradientDescentOptimizerBasev4.

Reimplemented in itk::QuasiNewtonOptimizerv4.

virtual void itk::GradientDescentOptimizerv4::PrintSelf	(	std::ostream &	os,
		Indent	indent
	)		const `[protected, virtual]`

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::GradientDescentOptimizerBasev4.

Reimplemented in itk::QuasiNewtonOptimizerv4.

virtual void itk::GradientDescentOptimizerv4::ResumeOptimization ( ) [virtual]

Resume the optimization. Can be called after StopOptimization to resume. The bulk of the optimization work loop is here.

Implements itk::GradientDescentOptimizerBasev4.

virtual void itk::GradientDescentOptimizerv4::SetLearningRate ( InternalComputationValueType _arg ) [virtual]

Set the learning rate.

virtual void itk::GradientDescentOptimizerv4::SetMaximumStepSizeInPhysicalUnits ( InternalComputationValueType _arg ) [virtual]

Set the maximum step size.

virtual void itk::GradientDescentOptimizerv4::SetScalesEstimator ( OptimizerParameterScalesEstimator * _arg ) [virtual]

Set the scales estimator.

virtual void itk::GradientDescentOptimizerv4::StartOptimization ( ) [virtual]