itkObjectToObjectMultiMetricv4Test.cxx

/*=========================================================================
 *
 *  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
 *
 *         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.
 *
 *=========================================================================*/

#include "itkObjectToObjectMultiMetricv4.h"
#include "itkMeanSquaresImageToImageMetricv4.h"
#include "itkMattesMutualInformationImageToImageMetricv4.h"
#include "itkJointHistogramMutualInformationImageToImageMetricv4.h"
#include "itkANTSNeighborhoodCorrelationImageToImageMetricv4.h"
#include "itkTranslationTransform.h"
#include "itkLinearInterpolateImageFunction.h"
#include "itkImage.h"
#include "itkGaussianImageSource.h"
#include "itkShiftScaleImageFilter.h"
#include "itkTestingMacros.h"
#include "itkCompositeTransform.h"
#include "itkEuclideanDistancePointSetToPointSetMetricv4.h"
#include "itkExpectationBasedPointSetToPointSetMetricv4.h"
#include "itkRegistrationParameterScalesFromPhysicalShift.h"


/** This test illustrates the use of the MultivariateImageToImageMetric class, which
    takes N metrics and assigns a weight to each metric's result.
 */

constexpr unsigned int ObjectToObjectMultiMetricv4TestDimension = 2;
using ObjectToObjectMultiMetricv4TestMultiMetricType =
  itk::ObjectToObjectMultiMetricv4<ObjectToObjectMultiMetricv4TestDimension, ObjectToObjectMultiMetricv4TestDimension>;

//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////

int
itkObjectToObjectMultiMetricv4TestEvaluate(ObjectToObjectMultiMetricv4TestMultiMetricType::Pointer & multiVariateMetric,
                                           bool useDisplacementTransform)
{
  int testStatus = EXIT_SUCCESS;
  using MultiMetricType = ObjectToObjectMultiMetricv4TestMultiMetricType;

  // Setup weights
  MultiMetricType::WeightsArrayType origMetricWeights(multiVariateMetric->GetNumberOfMetrics());
  MultiMetricType::WeightValueType  weightSum = 0;
  for (itk::SizeValueType n = 0; n < multiVariateMetric->GetNumberOfMetrics(); ++n)
  {
    origMetricWeights[n] = static_cast<MultiMetricType::WeightValueType>(n + 1);
    weightSum += origMetricWeights[n];
  }
  multiVariateMetric->SetMetricWeights(origMetricWeights);

  // Initialize. This initializes all the component metrics.
  std::cout << "Initialize" << std::endl;
  multiVariateMetric->Initialize();

  // Print out metric value and derivative.
  using MeasureType = MultiMetricType::MeasureType;
  MeasureType                     measure = 0;
  MultiMetricType::DerivativeType DerivResultOfGetValueAndDerivative;
  std::cout << "GetValueAndDerivative" << std::endl;
  try
  {
    multiVariateMetric->GetValueAndDerivative(measure, DerivResultOfGetValueAndDerivative);
  }
  catch (const itk::ExceptionObject & exp)
  {
    std::cerr << "Exception caught during call to GetValueAndDerivative:" << std::endl;
    std::cerr << exp << std::endl;
    testStatus = EXIT_FAILURE;
  }
  std::cout << "Multivariate measure: " << measure << std::endl;
  if (!useDisplacementTransform)
  {
    std::cout << "  Derivative : " << DerivResultOfGetValueAndDerivative << std::endl << std::endl;
  }

  // Test GetDerivative
  MultiMetricType::DerivativeType ResultOfGetDerivative;
  multiVariateMetric->GetDerivative(ResultOfGetDerivative);
  for (MultiMetricType::NumberOfParametersType p = 0; p < multiVariateMetric->GetNumberOfParameters(); ++p)
  {
    // When accumulation is done accross multiple threads, the accumulations can be done
    // in different orders resulting in slightly different numerical results.
    // The FloatAlmostEqual is used to address the multi-threaded accumulation differences
    if (!itk::Math::FloatAlmostEqual(ResultOfGetDerivative[p], DerivResultOfGetValueAndDerivative[p], 8, 1e-15))
    {
      std::cerr << "Results do not match between GetValueAndDerivative and GetDerivative." << std::endl;
      std::cout << ResultOfGetDerivative << " != " << DerivResultOfGetValueAndDerivative << std::endl;
      std::cout << "DIFF: " << ResultOfGetDerivative - DerivResultOfGetValueAndDerivative << std::endl;
      testStatus = EXIT_FAILURE;
    }
  }

  // Test GetValue method
  MeasureType measure2 = 0;
  std::cout << "GetValue" << std::endl;
  try
  {
    measure2 = multiVariateMetric->GetValue();
  }
  catch (const itk::ExceptionObject & exp)
  {
    std::cerr << "Exception caught during call to GetValue:" << std::endl;
    std::cerr << exp << std::endl;
    testStatus = EXIT_FAILURE;
  }
  if (!itk::Math::FloatAlmostEqual(measure2, measure))
  {
    std::cerr << "measure does not match between calls to GetValue and GetValueAndDerivative: "
              << "measure: " << measure << " measure2: " << measure2 << std::endl;
    testStatus = EXIT_FAILURE;
  }

  // Evaluate individually
  MeasureType                     metricValue = itk::NumericTraits<MeasureType>::ZeroValue();
  MeasureType                     weightedMetricValue = itk::NumericTraits<MeasureType>::ZeroValue();
  MultiMetricType::DerivativeType metricDerivative;
  MultiMetricType::DerivativeType DerivResultOfGetValueAndDerivativeTruth(multiVariateMetric->GetNumberOfParameters());
  DerivResultOfGetValueAndDerivativeTruth.Fill(itk::NumericTraits<MultiMetricType::DerivativeValueType>::ZeroValue());
  MultiMetricType::DerivativeValueType totalMagnitude =
    itk::NumericTraits<MultiMetricType::DerivativeValueType>::ZeroValue();

  for (itk::SizeValueType i = 0; i < multiVariateMetric->GetNumberOfMetrics(); ++i)
  {
    std::cout << "GetValueAndDerivative on component metrics" << std::endl;
    multiVariateMetric->GetMetricQueue()[i]->GetValueAndDerivative(metricValue, metricDerivative);
    std::cout << " Metric " << i << " value : " << metricValue << std::endl;
    if (!useDisplacementTransform)
    {
      std::cout << " Metric " << i << " derivative : " << metricDerivative << std::endl << std::endl;
    }
    if (!itk::Math::FloatAlmostEqual(metricValue, multiVariateMetric->GetValueArray()[i]))
    {
      std::cerr << "Individual metric value " << metricValue
                << " does not match that returned from multi-variate metric: " << multiVariateMetric->GetValueArray()[i]
                << std::endl;
      testStatus = EXIT_FAILURE;
    }
    weightedMetricValue += metricValue * origMetricWeights[i] / weightSum;
    for (MultiMetricType::NumberOfParametersType p = 0; p < multiVariateMetric->GetNumberOfParameters(); ++p)
    {
      DerivResultOfGetValueAndDerivativeTruth[p] +=
        metricDerivative[p] * (origMetricWeights[i] / weightSum) / metricDerivative.magnitude();
    }
    totalMagnitude += metricDerivative.magnitude();
  }
  totalMagnitude /= multiVariateMetric->GetNumberOfMetrics();
  for (MultiMetricType::NumberOfParametersType p = 0; p < multiVariateMetric->GetNumberOfParameters(); ++p)
  {
    DerivResultOfGetValueAndDerivativeTruth[p] *= totalMagnitude;
  }

  if (itk::Math::abs(weightedMetricValue - multiVariateMetric->GetWeightedValue()) > 1e-6)
  {
    std::cerr << "Computed weighted metric value " << weightedMetricValue << " does match returned value "
              << multiVariateMetric->GetWeightedValue() << std::endl;
    testStatus = EXIT_FAILURE;
  }

  for (MultiMetricType::NumberOfParametersType p = 0; p < multiVariateMetric->GetNumberOfParameters(); ++p)
  {
    auto tolerance = static_cast<MultiMetricType::DerivativeValueType>(1e-6);
    if (itk::Math::abs(DerivResultOfGetValueAndDerivativeTruth[p] - DerivResultOfGetValueAndDerivative[p]) > tolerance)
    {
      std::cerr << "Error: DerivResultOfGetValueAndDerivative does not match expected result." << std::endl;
      if (useDisplacementTransform)
      {
        std::cerr << "  DerivResultOfGetValueAndDerivative[" << p << "]: " << DerivResultOfGetValueAndDerivative[p]
                  << std::endl
                  << "  DerivResultOfGetValueAndDerivativeTruth[" << p
                  << "]: " << DerivResultOfGetValueAndDerivativeTruth[p] << std::endl;
      }
      else
      {
        std::cerr << "  DerivResultOfGetValueAndDerivative: " << DerivResultOfGetValueAndDerivative << std::endl
                  << "  DerivResultOfGetValueAndDerivativeTruth: " << DerivResultOfGetValueAndDerivativeTruth
                  << std::endl;
      }
      testStatus = EXIT_FAILURE;
    }
  }

  return testStatus;
}

////////////////////////////////////////////////////////////

int
itkObjectToObjectMultiMetricv4TestRun(bool useDisplacementTransform)
{
  // Create two simple images
  const unsigned int Dimension = ObjectToObjectMultiMetricv4TestDimension;
  using PixelType = double;
  using CoordinateRepresentationType = double;

  // Allocate Images
  using FixedImageType = itk::Image<PixelType, Dimension>;
  using MovingImageType = itk::Image<PixelType, Dimension>;

  // Declare Gaussian Sources
  using FixedImageSourceType = itk::GaussianImageSource<FixedImageType>;

  // Note: the following declarations are classical arrays
  FixedImageType::SizeValueType    fixedImageSize[] = { 100, 100 };
  FixedImageType::SpacingValueType fixedImageSpacing[] = { 1.0f, 1.0f };
  FixedImageType::PointValueType   fixedImageOrigin[] = { 0.0f, 0.0f };
  auto                             fixedImageSource = FixedImageSourceType::New();

  fixedImageSource->SetSize(fixedImageSize);
  fixedImageSource->SetOrigin(fixedImageOrigin);
  fixedImageSource->SetSpacing(fixedImageSpacing);
  fixedImageSource->SetNormalized(false);
  fixedImageSource->SetScale(1.0f);
  fixedImageSource->Update(); // Force the filter to run
  FixedImageType::Pointer fixedImage = fixedImageSource->GetOutput();

  using ShiftScaleFilterType = itk::ShiftScaleImageFilter<FixedImageType, MovingImageType>;
  auto shiftFilter = ShiftScaleFilterType::New();
  shiftFilter->SetInput(fixedImage);
  shiftFilter->SetShift(2.0);
  shiftFilter->Update();
  MovingImageType::Pointer movingImage = shiftFilter->GetOutput();

  // Set up the metric.
  using MultiMetricType = ObjectToObjectMultiMetricv4TestMultiMetricType;
  auto multiVariateMetric = MultiMetricType::New();

  // Instantiate and Add metrics to the queue
  using JointHistorgramMetrictype =
    itk::JointHistogramMutualInformationImageToImageMetricv4<FixedImageType, MovingImageType>;
  using MeanSquaresMetricType = itk::MeanSquaresImageToImageMetricv4<FixedImageType, MovingImageType>;
  using MattesMutualInformationMetricType =
    itk::MattesMutualInformationImageToImageMetricv4<FixedImageType, MovingImageType>;
  using ANTSNCMetricType = itk::ANTSNeighborhoodCorrelationImageToImageMetricv4<FixedImageType, MovingImageType>;

  auto m1 = MeanSquaresMetricType::New();
  auto m2 = MattesMutualInformationMetricType::New();
  auto m3 = JointHistorgramMetrictype::New();
  auto m4 = ANTSNCMetricType::New();

  // Set up a transform
  using TransformType = itk::Transform<CoordinateRepresentationType, Dimension, Dimension>;
  using DisplacementTransformType = itk::DisplacementFieldTransform<double, Dimension>;
  using TranslationTransformType = itk::TranslationTransform<CoordinateRepresentationType, Dimension>;
  TransformType::Pointer transform;

  if (useDisplacementTransform)
  {
    using FieldType = DisplacementTransformType::DisplacementFieldType;
    using VectorType = itk::Vector<double, Dimension>;

    VectorType zero;
    zero.Fill(0.0);

    auto field = FieldType::New();
    field->SetRegions(fixedImage->GetBufferedRegion());
    field->SetSpacing(fixedImage->GetSpacing());
    field->SetOrigin(fixedImage->GetOrigin());
    field->Allocate();
    field->FillBuffer(zero);

    auto displacementTransform = DisplacementTransformType::New();
    displacementTransform->SetDisplacementField(field);
    transform = displacementTransform;
  }
  else
  {
    auto translationTransform = TranslationTransformType::New();
    translationTransform->SetIdentity();
    transform = translationTransform;
  }

  // Plug the images and transform into the metrics
  std::cout << "Setup metrics" << std::endl;
  m1->SetFixedImage(fixedImage);
  m1->SetMovingImage(movingImage);
  m1->SetMovingTransform(transform);
  m2->SetFixedImage(fixedImage);
  m2->SetMovingImage(movingImage);
  m2->SetMovingTransform(transform);
  m3->SetFixedImage(fixedImage);
  m3->SetMovingImage(movingImage);
  m3->SetMovingTransform(transform);
  m4->SetFixedImage(fixedImage);
  m4->SetMovingImage(movingImage);
  m4->SetMovingTransform(transform);

  // Add the component metrics
  std::cout << "Add component metrics" << std::endl;
  multiVariateMetric->AddMetric(m1);
  multiVariateMetric->AddMetric(m2);
  multiVariateMetric->AddMetric(m3);
  multiVariateMetric->AddMetric(m4);

  if (multiVariateMetric->GetMetricQueue()[0] != m1 || multiVariateMetric->GetMetricQueue()[3] != m4)
  {
    std::cerr << "AddMetric or GetMetricQueue failed." << std::endl;
    return EXIT_FAILURE;
  }

  // Expect return true because all image metrics
  if (multiVariateMetric->SupportsArbitraryVirtualDomainSamples() == false)
  {
    std::cerr << "Expected SupportsArbitraryVirtualDomainSamples() to return false, but got true. " << std::endl;
    return EXIT_FAILURE;
  }

  // Test Set/Get Transform mechanics
  multiVariateMetric->Initialize();
  if (multiVariateMetric->GetMovingTransform() != transform.GetPointer())
  {
    std::cerr << "Automatic transform assignment failed. transform: " << transform.GetPointer()
              << " GetMovingTranform: " << multiVariateMetric->GetMovingTransform() << std::endl;
    return EXIT_FAILURE;
  }
  multiVariateMetric->SetMovingTransform(nullptr);
  for (itk::SizeValueType n = 0; n < multiVariateMetric->GetNumberOfMetrics(); ++n)
  {
    if (multiVariateMetric->GetMovingTransform() != nullptr ||
        multiVariateMetric->GetMetricQueue()[n]->GetMovingTransform() != nullptr)
    {
      std::cerr << "Assignment of null transform failed. multiVariateMetric->GetMovingTransform(): "
                << multiVariateMetric->GetMovingTransform() << " multiVariateMetric->GetMetricQueue()[" << n
                << "]->GetMovingTransform(): " << multiVariateMetric->GetMetricQueue()[n]->GetMovingTransform()
                << std::endl;
      return EXIT_FAILURE;
    }
  }
  multiVariateMetric->SetMovingTransform(transform);
  for (itk::SizeValueType n = 0; n < multiVariateMetric->GetNumberOfMetrics(); ++n)
  {
    if (multiVariateMetric->GetMovingTransform() != transform.GetPointer() ||
        multiVariateMetric->GetMetricQueue()[0]->GetMovingTransform() != transform.GetPointer())
    {
      std::cerr << "Assignment of transform failed." << std::endl;
      return EXIT_FAILURE;
    }
  }
  if (multiVariateMetric->GetMovingTransform() != transform.GetPointer())
  {
    std::cerr << "Retrieval of transform failed." << std::endl;
  }

  // Test with images
  std::cout << "*** Test image metrics *** " << std::endl;
  if (itkObjectToObjectMultiMetricv4TestEvaluate(multiVariateMetric, useDisplacementTransform) != EXIT_SUCCESS)
  {
    return EXIT_FAILURE;
  }

  std::cout << "*** Test with mismatched transforms *** " << std::endl;
  auto transform2 = TranslationTransformType::New();
  m4->SetMovingTransform(transform2);
  ITK_TRY_EXPECT_EXCEPTION(multiVariateMetric->Initialize());
  m4->SetMovingTransform(transform);

  std::cout << "*** Test with proper CompositeTransform ***" << std::endl;
  using CompositeTransformType = itk::CompositeTransform<CoordinateRepresentationType, Dimension>;
  auto compositeTransform = CompositeTransformType::New();
  compositeTransform->AddTransform(transform2);
  compositeTransform->AddTransform(transform);
  compositeTransform->SetOnlyMostRecentTransformToOptimizeOn();
  m4->SetMovingTransform(compositeTransform);
  if (itkObjectToObjectMultiMetricv4TestEvaluate(multiVariateMetric, useDisplacementTransform) != EXIT_SUCCESS)
  {
    std::cerr << "Failed with proper CompositeTransform." << std::endl;
    return EXIT_FAILURE;
  }

  std::cout << "*** Test with CompositeTransform - too many active transforms ***" << std::endl;
  compositeTransform->SetAllTransformsToOptimizeOn();
  ITK_TRY_EXPECT_EXCEPTION(multiVariateMetric->Initialize());

  std::cout << "*** Test with CompositeTransform - one active transform, but wrong one ***" << std::endl;
  compositeTransform->SetAllTransformsToOptimizeOff();
  compositeTransform->SetNthTransformToOptimizeOn(0);
  ITK_TRY_EXPECT_EXCEPTION(multiVariateMetric->Initialize());

  // Reset transform
  m4->SetMovingTransform(transform);

  //
  // Test with adding point set metrics
  //
  using PointSetType = itk::PointSet<float, Dimension>;
  auto fixedPoints = PointSetType::New();
  auto movingPoints = PointSetType::New();
  fixedPoints->Initialize();
  movingPoints->Initialize();

  PointSetType::PointType point;
  for (itk::SizeValueType n = 0; n < 100; ++n)
  {
    point[0] = n * 1.0;
    point[1] = n * 2.0;
    fixedPoints->SetPoint(n, point);
    point[0] += 0.5;
    point[1] += 0.5;
    movingPoints->SetPoint(n, point);
  }

  using ExpectationPointSetMetricType = itk::ExpectationBasedPointSetToPointSetMetricv4<PointSetType>;
  using EuclideanPointSetMetricType = itk::EuclideanDistancePointSetToPointSetMetricv4<PointSetType>;
  auto expectationPointSetMetric = ExpectationPointSetMetricType::New();
  auto euclideanPointSetMetric = EuclideanPointSetMetricType::New();

  expectationPointSetMetric->SetFixedPointSet(fixedPoints);
  expectationPointSetMetric->SetMovingPointSet(movingPoints);
  expectationPointSetMetric->SetMovingTransform(transform);
  euclideanPointSetMetric->SetFixedPointSet(fixedPoints);
  euclideanPointSetMetric->SetMovingPointSet(movingPoints);
  euclideanPointSetMetric->SetMovingTransform(transform);

  multiVariateMetric->AddMetric(expectationPointSetMetric);
  multiVariateMetric->AddMetric(euclideanPointSetMetric);


  // Expect return false because of point set metrics
  if (multiVariateMetric->SupportsArbitraryVirtualDomainSamples() == true)
  {
    std::cerr << "Expected SupportsArbitraryVirtualDomainSamples() to return true, but got false. " << std::endl;
    return EXIT_FAILURE;
  }

  // Test
  std::cout << "*** Test with PointSet metrics and Image metrics *** " << std::endl;
  if (itkObjectToObjectMultiMetricv4TestEvaluate(multiVariateMetric, useDisplacementTransform) != EXIT_SUCCESS)
  {
    return EXIT_FAILURE;
  }

  //
  // Exercise basic operation with a scales estimator
  //
  using ScalesEstimatorMultiType = itk::RegistrationParameterScalesFromPhysicalShift<MultiMetricType>;
  auto shiftScaleEstimator = ScalesEstimatorMultiType::New();
  shiftScaleEstimator->SetMetric(multiVariateMetric);
  // Have to assign virtual domain sampling points when using a point set with scales estimator
  shiftScaleEstimator->SetVirtualDomainPointSet(expectationPointSetMetric->GetVirtualTransformedPointSet());

  ScalesEstimatorMultiType::ScalesType scales;
  shiftScaleEstimator->EstimateScales(scales);
  std::cout << "Estimated scales: " << scales << std::endl;

  ScalesEstimatorMultiType::FloatType      stepScale;
  ScalesEstimatorMultiType::ParametersType step;
  step.SetSize(multiVariateMetric->GetNumberOfParameters());
  step.Fill(itk::NumericTraits<ScalesEstimatorMultiType::ParametersType::ValueType>::OneValue());
  stepScale = shiftScaleEstimator->EstimateStepScale(step);
  std::cout << "Estimated stepScale: " << stepScale << std::endl;

  //
  // Test that we get the same scales/step estimation
  // with a single metric and the same metric twice in a multimetric
  //
  ScalesEstimatorMultiType::ScalesType singleScales, multiSingleScales, multiDoubleScales;
  ScalesEstimatorMultiType::FloatType  singleStep, multiSingleStep, multiDoubleStep;
  step.SetSize(m1->GetNumberOfParameters());
  step.Fill(itk::NumericTraits<ScalesEstimatorMultiType::ParametersType::ValueType>::OneValue());

  using ScalesEstimatorMeanSquaresType = itk::RegistrationParameterScalesFromPhysicalShift<MeanSquaresMetricType>;
  auto singleShiftScaleEstimator = ScalesEstimatorMeanSquaresType::New();
  singleShiftScaleEstimator->SetMetric(m1);
  m1->Initialize();
  singleShiftScaleEstimator->EstimateScales(singleScales);
  std::cout << "Single metric estimated scales: " << singleScales << std::endl;
  singleStep = singleShiftScaleEstimator->EstimateStepScale(step);
  std::cout << "Single metric estimated stepScale: " << singleStep << std::endl;

  auto multiSingleMetric = MultiMetricType::New();
  multiSingleMetric->AddMetric(m1);
  multiSingleMetric->Initialize();
  shiftScaleEstimator->SetMetric(multiSingleMetric);
  shiftScaleEstimator->EstimateScales(multiSingleScales);
  std::cout << "multi-single estimated scales: " << multiSingleScales << std::endl;
  multiSingleStep = shiftScaleEstimator->EstimateStepScale(step);
  std::cout << "multi-single estimated stepScale: " << multiSingleStep << std::endl;

  auto multiDoubleMetric = MultiMetricType::New();
  multiDoubleMetric->AddMetric(m1);
  multiDoubleMetric->AddMetric(m1);
  multiDoubleMetric->Initialize();
  shiftScaleEstimator->SetMetric(multiDoubleMetric);
  shiftScaleEstimator->EstimateScales(multiDoubleScales);
  std::cout << "multi-double estimated scales: " << multiDoubleScales << std::endl;
  multiDoubleStep = shiftScaleEstimator->EstimateStepScale(step);
  std::cout << "multi-double estimated stepScale: " << multiDoubleStep << std::endl;

  // Check that results are the same for all three estimations
  bool passedEstimation = true;
  auto tolerance = static_cast<ScalesEstimatorMultiType::FloatType>(1e-6);
  if (itk::Math::abs(singleStep - multiSingleStep) > tolerance ||
      itk::Math::abs(singleStep - multiDoubleStep) > tolerance)
  {
    std::cerr << "Steps do not match as expected between estimation on same metric." << std::endl;
    passedEstimation = false;
  }
  if (itk::Math::abs(singleScales[0] - multiSingleScales[0]) > tolerance ||
      itk::Math::abs(singleScales[1] - multiSingleScales[1]) > tolerance ||
      itk::Math::abs(singleScales[0] - multiDoubleScales[0]) > tolerance ||
      itk::Math::abs(singleScales[1] - multiDoubleScales[1]) > tolerance)
  {
    std::cerr << "Scales do not match as expected between estimation on same metric." << std::endl;
    passedEstimation = false;
  }
  if (!passedEstimation)
  {
    return EXIT_FAILURE;
  }

  if (!useDisplacementTransform)
  {
    // Exercising the Print function
    std::cout << "Print: " << std::endl;
    multiVariateMetric->Print(std::cout);

    // Test ClearMetricQueue
    multiVariateMetric->ClearMetricQueue();
    if (multiVariateMetric->GetNumberOfMetrics() != 0)
    {
      std::cerr << "ClearMetricQueue() failed. Number of metrics is not zero." << std::endl;
      return EXIT_FAILURE;
    }
  }

  return EXIT_SUCCESS;
}

int
itkObjectToObjectMultiMetricv4Test(int, char *[])
{
  std::cout << "XXX Test with TranslationTransform XXX" << std::endl << std::endl;
  int result = itkObjectToObjectMultiMetricv4TestRun(false);
  if (result == EXIT_FAILURE)
  {
    std::cerr << "Failed test with translation transform. See message above." << std::endl;
    return EXIT_FAILURE;
  }

  std::cout << std::endl << std::endl << "XXX Test with DisplacementFieldTransform XXX" << std::endl << std::endl;
  result = itkObjectToObjectMultiMetricv4TestRun(true);
  if (result == EXIT_FAILURE)
  {
    std::cerr << "Failed test with displacement field transform. See message above." << std::endl;
    return EXIT_FAILURE;
  }

  return EXIT_SUCCESS;
}