/*========================================================================= * * Copyright NumFOCUS * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * https://www.apache.org/licenses/LICENSE-2.0.txt * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. * *=========================================================================*/ #ifndef itkSingleValuedNonLinearVnlOptimizer_h #define itkSingleValuedNonLinearVnlOptimizer_h #include "itkSingleValuedNonLinearOptimizer.h" #include "itkSingleValuedVnlCostFunctionAdaptor.h" #include "itkCommand.h" #include "ITKOptimizersExport.h" namespace itk { /** \class SingleValuedNonLinearVnlOptimizer * \brief This class is a base for the Optimization methods that * optimize a single valued function. * * It is an Adaptor class for optimizers provided by the vnl library * * \ingroup Numerics Optimizers * \ingroup ITKOptimizers */ class ITKOptimizers_EXPORT SingleValuedNonLinearVnlOptimizer : public SingleValuedNonLinearOptimizer { public: ITK_DISALLOW_COPY_AND_MOVE(SingleValuedNonLinearVnlOptimizer); /** Standard class type aliases. */ using Self = SingleValuedNonLinearVnlOptimizer; using Superclass = SingleValuedNonLinearOptimizer; using Pointer = SmartPointer; using ConstPointer = SmartPointer; /** \see LightObject::GetNameOfClass() */ itkOverrideGetNameOfClassMacro(SingleValuedNonLinearVnlOptimizer); /** Command observer that will interact with the ITKVNL cost-function * adaptor in order to generate iteration events. This will allow to overcome * the limitation of VNL optimizers not offering callbacks for every * iteration */ using CommandType = ReceptorMemberCommand; /** Set the cost Function. This method has to be overloaded * by derived classes because the CostFunctionAdaptor requires * to know the number of parameters at construction time. This * number of parameters is obtained at run-time from the itkCostFunction. * As a consequence each derived optimizer should construct its own * CostFunctionAdaptor when overloading this method */ void SetCostFunction(SingleValuedCostFunction * costFunction) override = 0; /** Methods to define whether the cost function will be maximized or * minimized. By default the VNL amoeba optimizer is only a minimizer. * Maximization is implemented here by notifying the CostFunctionAdaptor * which in its turn will multiply the function values and its derivative by * -1.0. */ itkGetConstReferenceMacro(Maximize, bool); itkSetMacro(Maximize, bool); itkBooleanMacro(Maximize); bool GetMinimize() const { return !m_Maximize; } void SetMinimize(bool v) { this->SetMaximize(!v); } void MinimizeOn() { this->MaximizeOff(); } void MinimizeOff() { this->MaximizeOn(); } /** Return Cached Values. These method have the advantage of not triggering a * recomputation of the metric value, but it has the disadvantage of returning * a value that may not be the one corresponding to the current parameters. For * GUI update purposes, this method is a good option, for mathematical * validation you should rather call GetValue(). */ itkGetConstReferenceMacro(CachedValue, MeasureType); itkGetConstReferenceMacro(CachedDerivative, DerivativeType); itkGetConstReferenceMacro(CachedCurrentPosition, ParametersType); /** Returns true if derived optimizer supports using scales. * For optimizers that do not support scaling, this * default function is overridden to return false.*/ virtual bool CanUseScales() const { return true; } protected: SingleValuedNonLinearVnlOptimizer(); ~SingleValuedNonLinearVnlOptimizer() override; using CostFunctionAdaptorType = SingleValuedVnlCostFunctionAdaptor; void SetCostFunctionAdaptor(CostFunctionAdaptorType * adaptor); const CostFunctionAdaptorType * GetCostFunctionAdaptor() const; CostFunctionAdaptorType * GetCostFunctionAdaptor(); /** The purpose of this method is to get around the lack of * const-correctness in VNL cost-functions and optimizers */ CostFunctionAdaptorType * GetNonConstCostFunctionAdaptor() const; /** Print out internal state */ void PrintSelf(std::ostream & os, Indent indent) const override; private: /** Callback function for the Command Observer */ void IterationReport(const EventObject & event); CostFunctionAdaptorType * m_CostFunctionAdaptor{}; bool m_Maximize{}; CommandType::Pointer m_Command{}; mutable ParametersType m_CachedCurrentPosition{}; mutable MeasureType m_CachedValue{}; mutable DerivativeType m_CachedDerivative{}; }; } // end namespace itk #endif