ITK
6.0.0
Insight Toolkit
|
#include <itkObjectToObjectOptimizerBase.h>
Abstract base for object-to-object optimizers.
The goal of this optimizer hierarchy is to work with metrics of any type, i.e. working with any kind of object, such as image or point-set.
Transform parameters are not manipulated directly. Instead, the optimizer retrieves the metric derivative from the metric, modifies the derivative as required, then passes it back to the metric as an update. The metric then processes it as appropriate, typically by passing it to its transform that is being optimized.
The user can scale each component of the gradient (derivative) at each iteration in one of two ways:
1) manually, by setting a scaling vector using method SetScales(). SetScales() allows setting of a per-local-parameter scaling array. If unset, the m_Scales
array will be initialized to all 1's. Note that when used with transforms with local support, these scales correspond to each local parameter, and not to each parameter. For example, in a DisplacementFieldTransform of dimensionality N, the Scales is size N, with each element corresponding to a dimension within the transform's displacement field, and is applied to each vector in the displacement field.
or,
2) automatically, by assigning a ScalesEstimator using SetScalesEstimator(). When ScalesEstimator is assigned, the optimizer is enabled by default to estimate scales, and can be changed via SetDoEstimateScales(). The scales are estimated and assigned once, during the call to StartOptimization(). This option will override any manually-assigned scales.
SetWeights() allows setting of a per-local-parameter weighting array. If unset, the weights are treated as identity. Weights are multiplied by the gradient at the same time scaling is applied. Weights are different than scales, and may be used, for example, to easily mask out a particular parameter during optimization to hold it constant. Or they may be used to apply another kind of prior knowledge.
Threading of some optimizer operations may be handled within derived classes, for example in GradientDescentOptimizer.
Definition at line 113 of file itkObjectToObjectOptimizerBase.h.
Public Member Functions | |
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 StopConditionReturnStringType | GetStopConditionDescription () const =0 |
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 | StartOptimization (bool doOnlyInitialization=false) |
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 |
Protected Member Functions | |
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 | |
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 {} |
Additional Inherited Members | |
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 () |
using itk::ObjectToObjectOptimizerBaseTemplate< TInternalComputationValueType >::ConstPointer = SmartPointer<const Self> |
Definition at line 122 of file itkObjectToObjectOptimizerBase.h.
using itk::ObjectToObjectOptimizerBaseTemplate< TInternalComputationValueType >::DerivativeType = typename MetricType::DerivativeType |
Derivative type
Definition at line 139 of file itkObjectToObjectOptimizerBase.h.
using itk::ObjectToObjectOptimizerBaseTemplate< TInternalComputationValueType >::MeasureType = typename MetricType::MeasureType |
Measure type
Definition at line 145 of file itkObjectToObjectOptimizerBase.h.
using itk::ObjectToObjectOptimizerBaseTemplate< TInternalComputationValueType >::MetricType = ObjectToObjectMetricBaseTemplate<TInternalComputationValueType> |
Metric function type
Definition at line 135 of file itkObjectToObjectOptimizerBase.h.
using itk::ObjectToObjectOptimizerBaseTemplate< TInternalComputationValueType >::MetricTypePointer = typename MetricType::Pointer |
Definition at line 136 of file itkObjectToObjectOptimizerBase.h.
using itk::ObjectToObjectOptimizerBaseTemplate< TInternalComputationValueType >::NumberOfParametersType = typename MetricType::NumberOfParametersType |
Number of parameters type
Definition at line 142 of file itkObjectToObjectOptimizerBase.h.
using itk::ObjectToObjectOptimizerBaseTemplate< TInternalComputationValueType >::ParametersType = OptimizerParameters<TInternalComputationValueType> |
Parameters type.
Definition at line 132 of file itkObjectToObjectOptimizerBase.h.
using itk::ObjectToObjectOptimizerBaseTemplate< TInternalComputationValueType >::Pointer = SmartPointer<Self> |
Definition at line 121 of file itkObjectToObjectOptimizerBase.h.
using itk::ObjectToObjectOptimizerBaseTemplate< TInternalComputationValueType >::ScalesEstimatorType = OptimizerParameterScalesEstimatorTemplate<TInternalComputationValueType> |
Definition at line 129 of file itkObjectToObjectOptimizerBase.h.
using itk::ObjectToObjectOptimizerBaseTemplate< TInternalComputationValueType >::ScalesType = OptimizerParameters<TInternalComputationValueType> |
Scale type.
Definition at line 128 of file itkObjectToObjectOptimizerBase.h.
using itk::ObjectToObjectOptimizerBaseTemplate< TInternalComputationValueType >::Self = ObjectToObjectOptimizerBaseTemplate |
Standard class type aliases.
Definition at line 119 of file itkObjectToObjectOptimizerBase.h.
using itk::ObjectToObjectOptimizerBaseTemplate< TInternalComputationValueType >::StopConditionDescriptionType = std::ostringstream |
Stop condition internal string type
Definition at line 151 of file itkObjectToObjectOptimizerBase.h.
using itk::ObjectToObjectOptimizerBaseTemplate< TInternalComputationValueType >::StopConditionReturnStringType = std::string |
Stop condition return string type
Definition at line 148 of file itkObjectToObjectOptimizerBase.h.
using itk::ObjectToObjectOptimizerBaseTemplate< TInternalComputationValueType >::Superclass = Object |
Definition at line 120 of file itkObjectToObjectOptimizerBase.h.
|
protected |
Default constructor
|
overrideprotected |
Default constructor
|
inlinevirtual |
Returns true if derived optimizer supports using scales. For optimizers that do not support scaling, this default function is overridden to return false.
Reimplemented in itk::LBFGSBOptimizerv4.
Definition at line 279 of file itkObjectToObjectOptimizerBase.h.
|
virtual |
Option to use ScalesEstimator for scales estimation. The estimation is performed once at begin of optimization, and overrides any scales set using SetScales(). Default is true.
|
virtual |
Return current number of iterations.
|
virtual |
Accessor for metric value. Returns the value stored in m_CurrentMetricValue from the most recent call to evaluate the metric.
|
virtual |
Get a reference to the current position of the optimization. This returns the parameters from the assigned metric, since the optimizer itself does not store a position.
|
virtual |
Option to use ScalesEstimator for scales estimation. The estimation is performed once at begin of optimization, and overrides any scales set using SetScales(). Default is true.
|
virtual |
Accessors for Metric
|
overridevirtual |
Reimplemented from itk::Object.
Reimplemented in itk::OnePlusOneEvolutionaryOptimizerv4< TInternalComputationValueType >, itk::QuasiNewtonOptimizerv4Template< TInternalComputationValueType >, itk::PowellOptimizerv4< TInternalComputationValueType >, and itk::RegularStepGradientDescentOptimizerv4< TInternalComputationValueType >.
|
virtual |
Get the number of iterations.
Reimplemented in itk::LBFGS2Optimizerv4Template< TInternalComputationValueType >.
|
virtual |
Get the number of work units set to be used.
|
virtual |
Get current parameters scaling.
|
virtual |
Get whether scales are identity. Cannot be set
bool itk::ObjectToObjectOptimizerBaseTemplate< TInternalComputationValueType >::GetScalesInitialized | ( | ) | const |
Get whether the scales have been set. Returns true if m_Scales.Size() > 0
|
pure virtual |
Stop condition return string type
Implemented in itk::SingleValuedNonLinearVnlOptimizerv4, itk::LBFGS2Optimizerv4Template< TInternalComputationValueType >, itk::OnePlusOneEvolutionaryOptimizerv4< TInternalComputationValueType >, itk::AmoebaOptimizerv4, itk::ExhaustiveOptimizerv4< TInternalComputationValueType >, itk::LBFGSOptimizerBasev4< TInternalVnlOptimizerType >, itk::PowellOptimizerv4< TInternalComputationValueType >, itk::LBFGSOptimizerBasev4< vnl_lbfgs >, itk::LBFGSOptimizerBasev4< vnl_lbfgsb >, itk::MultiStartOptimizerv4Template< TInternalComputationValueType >, itk::GradientDescentOptimizerBasev4Template< TInternalComputationValueType >, and itk::MultiGradientOptimizerv4Template< TInternalComputationValueType >.
|
virtual |
Deprecated accessor for currently stored metric value for use by classes that support both v4 and v3 optimizers.
Reimplemented in itk::OnePlusOneEvolutionaryOptimizerv4< TInternalComputationValueType >, and itk::PowellOptimizerv4< TInternalComputationValueType >.
|
virtual |
Get current parameters weights. This will return an empty array when weights have not been set by user.
|
virtual |
Get whether weights are identity. Cannot be set
|
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::Object.
Reimplemented in itk::QuasiNewtonOptimizerv4Template< TInternalComputationValueType >, itk::OnePlusOneEvolutionaryOptimizerv4< TInternalComputationValueType >, itk::PowellOptimizerv4< TInternalComputationValueType >, and itk::RegularStepGradientDescentOptimizerv4< TInternalComputationValueType >.
|
virtual |
Option to use ScalesEstimator for scales estimation. The estimation is performed once at begin of optimization, and overrides any scales set using SetScales(). Default is true.
|
virtual |
Accessors for Metric
|
virtual |
Set the number of iterations.
Reimplemented in itk::LBFGS2Optimizerv4Template< TInternalComputationValueType >.
|
virtual |
Set the number of work units to use when threading. The default is the global default number of work units decided in the constructor of the MultiThreader.
|
inlinevirtual |
Set current parameters scaling.
Definition at line 174 of file itkObjectToObjectOptimizerBase.h.
|
virtual |
Set the scales estimator.
A ScalesEstimator is required for the scales estimation options to work. See the main documentation. Derived classes may also provide learning-rate estimation, in which case a scales estimator is also required.
|
virtual |
Set current parameters weights.
|
virtual |
Run the optimization.
doOnlyInitialization | This is false by default. It should only be set to true for special cases when the class should be initialized to perform optimization, but no optimization should be run. For example, itkMultiGradientOptimizerv4 needs to do this. |
Reimplemented in itk::LBFGS2Optimizerv4Template< TInternalComputationValueType >, itk::GradientDescentOptimizerv4Template< TInternalComputationValueType >, itk::OnePlusOneEvolutionaryOptimizerv4< TInternalComputationValueType >, itk::PowellOptimizerv4< TInternalComputationValueType >, itk::LBFGSBOptimizerv4, itk::MultiStartOptimizerv4Template< TInternalComputationValueType >, itk::ExhaustiveOptimizerv4< TInternalComputationValueType >, itk::RegularStepGradientDescentOptimizerv4< TInternalComputationValueType >, itk::LBFGSOptimizerv4, itk::LBFGSOptimizerBasev4< TInternalVnlOptimizerType >, itk::LBFGSOptimizerBasev4< vnl_lbfgs >, itk::LBFGSOptimizerBasev4< vnl_lbfgsb >, itk::GradientDescentOptimizerBasev4Template< TInternalComputationValueType >, itk::MultiGradientOptimizerv4Template< TInternalComputationValueType >, itk::QuasiNewtonOptimizerv4Template< TInternalComputationValueType >, itk::AmoebaOptimizerv4, itk::ConjugateGradientLineSearchOptimizerv4Template< TInternalComputationValueType >, and itk::SingleValuedNonLinearVnlOptimizerv4.
|
protected |
Definition at line 292 of file itkObjectToObjectOptimizerBase.h.
|
protected |
Metric measure value at a given iteration, as most recently evaluated.
Definition at line 296 of file itkObjectToObjectOptimizerBase.h.
|
protected |
Flag to control use of the ScalesEstimator (if set) for automatic scale estimation during StartOptimization()
Definition at line 320 of file itkObjectToObjectOptimizerBase.h.
|
protected |
Definition at line 290 of file itkObjectToObjectOptimizerBase.h.
|
protected |
Definition at line 293 of file itkObjectToObjectOptimizerBase.h.
|
protected |
Definition at line 291 of file itkObjectToObjectOptimizerBase.h.
|
protected |
Scales. Size is expected to be == metric->GetNumberOfLocalParameters(). See the main documentation for more details.
Definition at line 300 of file itkObjectToObjectOptimizerBase.h.
|
protected |
Flag to avoid unnecessary arithmetic when scales are identity.
Definition at line 309 of file itkObjectToObjectOptimizerBase.h.
|
protected |
Scales estimator. Optionally provided by user.
Definition at line 312 of file itkObjectToObjectOptimizerBase.h.
|
protected |
Parameter weights. These are applied to local parameters, at the same time as scales. See main documentation. If not set by user, the array remains empty and treated as identity to simplify the reuse of an optimizer with transforms with different numbers of parameters.
Definition at line 306 of file itkObjectToObjectOptimizerBase.h.
|
protected |
Flag to avoid unnecessary arithmetic when weights are identity.
Definition at line 315 of file itkObjectToObjectOptimizerBase.h.