#include <elxConjugateGradientFRPR.h>
The ConjugateGradientFRPR class.
This component is based on the itkFRPROptimizer. This is a Fletcher-Reeves conjugate gradient optimizer, in combination with an exact (dBrent) line search, based on the description in Numerical Recipes in C++
This optimizer support the NewSamplesEveryIteration option. It requests new samples upon every derivative evaluation, but actually this makes no sense for a conjugate gradient optimizer. So, think twice before using it.
Optimizer: Select this optimizer as follows:
(Optimizer "ConjugateGradientFRPR")
MaximumNumberOfIterations: The maximum number of iterations in each resolution.
example: (MaximumNumberOfIterations 100 100 50)
Default value: 100.
MaximumNumberOfLineSearchIterations: The maximum number of iterations in each resolution.
example: (MaximumNumberOfIterations 10 10 5)
Default value: 10.
StepLength: Set the length of the initial step tried by the line seach, used to bracket the minimum.
example: (StepLength 2.0 1.0 0.5)
Default value: 1.0.
ValueTolerance: Convergence is declared if:
example: (ValueTolerance 0.001 0.00001 0.000001)
Default value: 0.00001.
where:
= current mininum of the gain
= current brackets around the minimum
example: (LineSearchStepTolerance 0.001 0.00001 0.000001)
Default value: 0.00001.
Definition at line 74 of file elxConjugateGradientFRPR.h.
Public Types | |
using | ConstPointer = itk::SmartPointer<const Self> |
using | DerivativeType = SingleValuedNonLinearOptimizer::DerivativeType |
using | ITKBaseType = typename Superclass2::ITKBaseType |
using | Pointer = itk::SmartPointer<Self> |
using | Self = ConjugateGradientFRPR |
using | Superclass1 = itk::FRPROptimizer |
using | Superclass2 = OptimizerBase<TElastix> |
Public Types inherited from elastix::OptimizerBase< TElastix > | |
using | ElastixType |
using | ITKBaseType = itk::Optimizer |
using | ParametersType = typename ITKBaseType::ParametersType |
using | RegistrationType |
using | Self = OptimizerBase |
using | Superclass = BaseComponentSE<TElastix> |
Public Types inherited from elastix::BaseComponentSE< TElastix > | |
using | ElastixType = TElastix |
using | RegistrationType = typename ElastixType::RegistrationBaseType |
using | Self = BaseComponentSE |
using | Superclass = BaseComponent |
Static Public Member Functions | |
static Pointer | New () |
Static Public Member Functions inherited from elastix::BaseComponent | |
template<typename TBaseComponent > | |
static auto | AsITKBaseType (TBaseComponent *const baseComponent) -> decltype(baseComponent->GetAsITKBaseType()) |
static void | InitializeElastixExecutable () |
static bool | IsElastixLibrary () |
Protected Member Functions | |
void | BracketedLineOptimize (double ax, double bx, double cx, double fa, double fb, double fc, double *extX, double *extVal, ParametersType &tempCoord) override |
ConjugateGradientFRPR () | |
void | GetValueAndDerivative (ParametersType &p, double *val, ParametersType *xi) override |
void | LineBracket (double *ax, double *bx, double *cx, double *fa, double *fb, double *fc, ParametersType &tempCoord) override |
void | LineOptimize (ParametersType *p, ParametersType &xi, double *val, ParametersType &tempCoord) override |
virtual void | SetLineBracketing (bool _arg) |
virtual void | SetLineOptimizing (bool _arg) |
~ConjugateGradientFRPR () override=default | |
Protected Member Functions inherited from elastix::OptimizerBase< TElastix > | |
virtual bool | GetNewSamplesEveryIteration () const |
OptimizerBase ()=default | |
virtual void | SelectNewSamples () |
~OptimizerBase () override=default | |
Protected Member Functions inherited from elastix::BaseComponentSE< TElastix > | |
BaseComponentSE ()=default | |
~BaseComponentSE () override=default | |
Protected Member Functions inherited from elastix::BaseComponent | |
BaseComponent ()=default | |
virtual | ~BaseComponent ()=default |
Protected Attributes | |
double | m_CurrentDerivativeMagnitude |
double | m_CurrentSearchDirectionMagnitude |
double | m_CurrentStepLength |
Protected Attributes inherited from elastix::BaseComponentSE< TElastix > | |
Configuration::ConstPointer | m_Configuration {} |
itk::WeakPointer< TElastix > | m_Elastix {} |
RegistrationType * | m_Registration {} |
Private Member Functions | |
const char * | DeterminePhase () const |
Private Attributes | |
elxOverrideGetSelfMacro | |
bool | m_LineBracketing |
bool | m_LineOptimizing |
Additional Inherited Members | |
Protected Types inherited from elastix::OptimizerBase< TElastix > | |
using | SettingsVectorType = typename std::vector<SettingsType> |
Static Protected Member Functions inherited from elastix::OptimizerBase< TElastix > | |
static void | PrintSettingsVector (const SettingsVectorType &settings) |
using elastix::ConjugateGradientFRPR< TElastix >::ConstPointer = itk::SmartPointer<const Self> |
Definition at line 86 of file elxConjugateGradientFRPR.h.
using elastix::ConjugateGradientFRPR< TElastix >::DerivativeType = SingleValuedNonLinearOptimizer::DerivativeType |
Definition at line 103 of file elxConjugateGradientFRPR.h.
using elastix::ConjugateGradientFRPR< TElastix >::ITKBaseType = typename Superclass2::ITKBaseType |
Definition at line 108 of file elxConjugateGradientFRPR.h.
using elastix::ConjugateGradientFRPR< TElastix >::Pointer = itk::SmartPointer<Self> |
Definition at line 85 of file elxConjugateGradientFRPR.h.
using elastix::ConjugateGradientFRPR< TElastix >::Self = ConjugateGradientFRPR |
Standard ITK.
Definition at line 82 of file elxConjugateGradientFRPR.h.
using elastix::ConjugateGradientFRPR< TElastix >::Superclass1 = itk::FRPROptimizer |
Definition at line 83 of file elxConjugateGradientFRPR.h.
using elastix::ConjugateGradientFRPR< TElastix >::Superclass2 = OptimizerBase<TElastix> |
Definition at line 84 of file elxConjugateGradientFRPR.h.
|
protected |
|
overrideprotecteddefault |
|
overridevirtual |
Reimplemented from elastix::BaseComponent.
|
overridevirtual |
Reimplemented from elastix::BaseComponent.
|
overridevirtual |
Reimplemented from elastix::BaseComponent.
|
overridevirtual |
Reimplemented from elastix::BaseComponent.
|
overridevirtual |
Methods to set parameters and print output at different stages in the registration process.
Reimplemented from elastix::BaseComponent.
|
overrideprotected |
Given a bracketing triple of points and their function values, returns a bounded extreme. These values are in parameter space, along the current line and wrt the current origin set via SetLine. Optimization terminates based on MaximumIteration, StepTolerance, or ValueTolerance. Implemented as Brent line optimers from NRC.
This implementation sets the LineOptimizing flag to 'true', calls the the superclass's implementation, stores extX as the current step length, and sets the LineOptimizing flag to 'false' again.
|
private |
elastix::ConjugateGradientFRPR< TElastix >::elxClassNameMacro | ( | "ConjugateGradientFRPR< TElastix >" | ) |
Name of this class.
|
virtual |
Run-time type information (and related methods).
Reimplemented from elastix::OptimizerBase< TElastix >.
|
virtual |
Return the magnitude of the cached derivative
|
virtual |
Get the magnitude of the line search direction
|
virtual |
Get the current gain
|
virtual |
|
virtual |
Check if the optimizer is currently Bracketing the minimum, or is optimizing along a line
|
overrideprotected |
Get the value of the n-dimensional cost function at this scalar step distance along the current line direction from the current line origin. Line origin and distances are set via SetLine.
This implementation calls the Superclass' implementation and caches the computed derivative's magnitude. Besides, it invokes the SelectNewSamples method.
elastix::ConjugateGradientFRPR< TElastix >::ITK_DISALLOW_COPY_AND_MOVE | ( | ConjugateGradientFRPR< TElastix > | ) |
|
overrideprotected |
The LineBracket routine from NRC. Uses current origin and line direction (from SetLine) to find a triple of points (ax, bx, cx) that bracket the extreme "near" the origin. Search first considers the point StepLength distance from ax. IMPORTANT: The value of ax and the value of the function at ax (i.e., fa), must both be provided to this function.
This implementation sets the LineBracketing flag to 'true', calls the superclass' implementation, stores bx as the current step length, invokes an iteration event, and sets the LineBracketing flag to 'false'
|
overrideprotected |
store the line search direction's (xi) magnitude and call the superclass' implementation.
|
static |
Method for creation through the object factory.
|
override |
Override the SetInitialPosition.
|
protectedvirtual |
|
protectedvirtual |
Set if the optimizer is currently bracketing the minimum, or is optimizing along a line
|
private |
Definition at line 215 of file elxConjugateGradientFRPR.h.
|
protected |
To store the latest computed derivative's magnitude
Definition at line 150 of file elxConjugateGradientFRPR.h.
|
protected |
Variable to store the line search direction magnitude
Definition at line 153 of file elxConjugateGradientFRPR.h.
|
protected |
the current gain
Definition at line 156 of file elxConjugateGradientFRPR.h.
|
private |
Definition at line 218 of file elxConjugateGradientFRPR.h.
|
private |
Definition at line 217 of file elxConjugateGradientFRPR.h.
Generated on 2024-07-17 for elastix by 1.11.0 (9b424b03c9833626cd435af22a444888fbbb192d) |