elxAdaGrad.h

Go to the documentation of this file.

/*=========================================================================
 *
 *  Copyright UMC Utrecht and contributors
 *
 *  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
 *
 *        http://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 __elxAdaGrad_h
#define __elxAdaGrad_h
 
#include "elxIncludes.h" // include first to avoid MSVS warning
#include "itkAdaptiveStepsizeOptimizer.h"
 
#include "itkComputePreconditionerUsingDisplacementDistribution.h"
#include "itkComputeDisplacementDistribution.h" // For fast step size estimation
 
#include "elxProgressCommand.h"
#include "itkAdvancedTransform.h"
#include "itkMersenneTwisterRandomVariateGenerator.h"
#include "itkAdvancedBSplineDeformableTransformBase.h"
#include "itkImageRandomSampler.h"
 
 
namespace elastix
{
template< class TElastix >
class AdaGrad :
  public itk::AdaptiveStepsizeOptimizer,
  public OptimizerBase< TElastix >
{
public:
 
  typedef AdaGrad Self;
  typedef AdaptiveStepsizeOptimizer             Superclass1;
  typedef OptimizerBase< TElastix >               Superclass2;
  typedef itk::SmartPointer< Self >               Pointer;
  typedef itk::SmartPointer< const Self >         ConstPointer;
 
  itkNewMacro( Self );
 
  itkTypeMacro( AdaGrad,
    VoxelWiseASGDOptimizer );
 
  elxClassNameMacro( "AdaGrad" );
 
  typedef Superclass1::CostFunctionType    CostFunctionType;
  typedef Superclass1::CostFunctionPointer CostFunctionPointer;
  typedef Superclass1::StopConditionType   StopConditionType;
 
  typedef typename Superclass2::ElastixType          ElastixType;
  typedef typename Superclass2::ElastixPointer       ElastixPointer;
  typedef typename Superclass2::ConfigurationType    ConfigurationType;
  typedef typename Superclass2::ConfigurationPointer ConfigurationPointer;
  typedef typename Superclass2::RegistrationType     RegistrationType;
  typedef typename Superclass2::RegistrationPointer  RegistrationPointer;
  typedef typename Superclass2::ITKBaseType          ITKBaseType;
  typedef itk::SizeValueType                         SizeValueType;
 
  typedef typename Superclass1::ParametersType ParametersType;
 
  void BeforeRegistration( void ) override;
 
  void BeforeEachResolution( void ) override;
 
  void AfterEachResolution( void ) override;
 
  void AfterEachIteration( void ) override;
 
  void AfterRegistration( void ) override;
 
  void StartOptimization( void ) override;
 
  void AdvanceOneStep( void ) override;
 
  void ResumeOptimization( void ) override;
 
  void MetricErrorResponse( itk::ExceptionObject & err ) override;
 
  itkSetMacro( AutomaticParameterEstimation, bool );
  itkGetConstMacro( AutomaticParameterEstimation, bool );
 
  itkSetMacro( MaximumStepLength, double );
  itkGetConstReferenceMacro( MaximumStepLength, double );
 
  itkSetClampMacro( RegularizationKappa, double, 0.0, 1.0 );
  itkGetConstReferenceMacro( RegularizationKappa, double );
 
  itkSetMacro( MaximumNumberOfSamplingAttempts, SizeValueType );
  itkGetConstReferenceMacro( MaximumNumberOfSamplingAttempts, SizeValueType );
 
protected :
 
  AdaGrad();
  ~AdaGrad() override {}
 
  typedef typename RegistrationType::FixedImageType  FixedImageType;
  typedef typename RegistrationType::MovingImageType MovingImageType;
 
  typedef typename FixedImageType::RegionType         FixedImageRegionType;
  typedef typename FixedImageType::IndexType          FixedImageIndexType;
  typedef typename FixedImageType::PointType          FixedImagePointType;
  typedef typename RegistrationType::ITKBaseType      itkRegistrationType;
  typedef typename itkRegistrationType::TransformType TransformType;
  typedef typename TransformType::JacobianType        JacobianType;
  typedef typename JacobianType::ValueType JacobianValueType;
  struct SettingsType { double a, A, alpha, fmax, fmin, omega; };
  typedef typename std::vector< SettingsType > SettingsVectorType;
  typedef typename ElastixType::FixedImageType OutputImageType;
 
  typedef itk::ComputePreconditionerUsingDisplacementDistribution<
    FixedImageType, TransformType >                     PreconditionerEstimationType;
  typedef typename PreconditionerEstimationType::Pointer PreconditionerEstimationPointer;
 
  typedef itk::ComputeDisplacementDistribution<
    FixedImageType, TransformType >                    ComputeDisplacementDistributionType;
 
 
  typedef itk::ImageSamplerBase< FixedImageType >       ImageSamplerBaseType;
  typedef typename ImageSamplerBaseType::Pointer        ImageSamplerBasePointer;
  typedef itk::ImageRandomSamplerBase< FixedImageType > ImageRandomSamplerBaseType;
  typedef typename
    ImageRandomSamplerBaseType::Pointer ImageRandomSamplerBasePointer;
  typedef
    itk::ImageRandomCoordinateSampler< FixedImageType > ImageRandomCoordinateSamplerType;
  typedef typename
    ImageRandomCoordinateSamplerType::Pointer ImageRandomCoordinateSamplerPointer;
  typedef itk::ImageRandomSampler< FixedImageType > ImageRandomSamplerType;
  typedef typename ImageRandomSamplerType::Pointer ImageRandomSamplerPointer;
  typedef itk::ImageGridSampler< FixedImageType > ImageGridSamplerType;
  typedef typename ImageGridSamplerType::Pointer  ImageGridSamplerPointer;
  typedef typename
    ImageGridSamplerType::ImageSampleContainerType ImageSampleContainerType;
  typedef typename ImageSampleContainerType::Pointer ImageSampleContainerPointer;
 
  typedef itk::Statistics::MersenneTwisterRandomVariateGenerator RandomGeneratorType;
  typedef typename RandomGeneratorType::Pointer                  RandomGeneratorPointer;
  typedef ProgressCommand                                        ProgressCommandType;
  typedef typename ProgressCommand::Pointer                      ProgressCommandPointer;
 
  typedef JacobianType TransformJacobianType;
  itkStaticConstMacro( FixedImageDimension, unsigned int, FixedImageType::ImageDimension );
  itkStaticConstMacro( MovingImageDimension, unsigned int, MovingImageType::ImageDimension );
  typedef typename TransformType::ScalarType CoordinateRepresentationType;
 
  typedef itk::AdvancedTransform<
    CoordinateRepresentationType,
    itkGetStaticConstMacro( FixedImageDimension ),
    itkGetStaticConstMacro( MovingImageDimension ) >            AdvancedTransformType;
  typedef typename AdvancedTransformType::Pointer               AdvancedTransformPointer;
  typedef typename
    AdvancedTransformType::NonZeroJacobianIndicesType NonZeroJacobianIndicesType;
 
  typedef itk::AdvancedBSplineDeformableTransformBase<
    CoordinateRepresentationType,
    itkGetStaticConstMacro( FixedImageDimension )>              AdvancedBSplineDeformableTransformType;
  typedef typename AdvancedBSplineDeformableTransformType::Pointer BSplineTransformBasePointer;
 
  SettingsVectorType m_SettingsVector;
 
  SizeValueType m_NumberOfGradientMeasurements;
  SizeValueType m_NumberOfJacobianMeasurements;
  SizeValueType m_NumberOfSamplesForNoiseCompensationFactor;
  SizeValueType m_NumberOfSamplesForPrecondition;
  SizeValueType m_NumberOfSpatialSamples;
 
  AdvancedTransformPointer m_AdvancedTransform;
 
  RandomGeneratorPointer m_RandomGenerator;
 
  double                    m_SigmoidScaleFactor;
  double                    m_NoiseFactor;
  double                    m_GlobalStepSize;
  double                    m_RegularizationKappa;
  double                    m_ConditionNumber;
 
  virtual void CheckForAdvancedTransform( void );
 
  virtual void PrintSettingsVector( const SettingsVectorType & settings ) const;
 
  virtual void AutomaticPreconditionerEstimation( void );
 
  virtual void SampleGradients( const ParametersType & mu0,
    double perturbationSigma, double & gg, double & ee );
 
  virtual void GetScaledDerivativeWithExceptionHandling(
    const ParametersType & parameters, DerivativeType & derivative );
 
  virtual void AddRandomPerturbation( ParametersType & parameters, double sigma );
 
private:
 
  AdaGrad( const Self & );  // purposely not implemented
  void operator=( const Self & ); // purposely not implemented
 
  bool   m_AutomaticParameterEstimation;
  double m_MaximumStepLength;
  double m_MaximumStepLengthRatio;
 
  SizeValueType m_MaximumNumberOfSamplingAttempts;
  SizeValueType m_CurrentNumberOfSamplingAttempts;
  SizeValueType m_PreviousErrorAtIteration;
  bool          m_AutomaticParameterEstimationDone;
 
  SizeValueType m_MaxBandCovSize;
  SizeValueType m_NumberOfBandStructureSamples;
 
  bool m_UseNoiseCompensation;
  bool m_OriginalButSigmoidToDefault;
 
};
 
} // end namespace elastix
 
#ifndef ITK_MANUAL_INSTANTIATION
#include "elxAdaGrad.hxx"
#endif
 
#endif // end #ifndef __elxAdaGrad_h