Examples/RegistrationITKv4/ModelToImageRegistration2.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.
 *
 *=========================================================================*/
 
// Software Guide : BeginLatex
//
//  This example illustrates the use of the \doxygen{SpatialObject} as a
//  component of the registration framework in order to perform model based
//  registration. In this case, a SpatialObject is used for generating a
//  \doxygen{PointSet} whose points are located in a narrow band around the
//  edges of the SpatialObject. This PointSet is then used in order to perform
//  PointSet to Image registration.
//
// Software Guide : EndLatex
 
 
//  Software Guide : BeginLatex
//
//  In this example we use the \doxygen{BoxSpatialObject}, that is one of the
//  simplest SpatialObjects in ITK.
//
//  \index{itk::BoxSpatialObject!header}
//
//  Software Guide : EndLatex
 
//  Software Guide : BeginCodeSnippet
#include "itkBoxSpatialObject.h"
//  Software Guide : EndCodeSnippet
 
 
//  Software Guide : BeginLatex
//
//  The generation of the PointSet is done in two stages. First the
//  SpatialObject is rasterized in order to generate an image containing a
//  binary mask that represents the inside and outside of the SpatialObject.
//  Second, this mask is used for computing a distance map, and the points
//  close to the boundary of the mask are taken as elements of the final
//  PointSet. The pixel values associated to the point in the PointSet are the
//  values of distance from each point to the binary mask.  The first stage is
//  performed by the \doxygen{SpatialObjectToImageFilter}, while the second
//  stage is performed witht eh \doxygen{BinaryMaskToNarrowBandPointSetFilter}
//
//  \index{itk::Spatial\-Object\-To\-Image\-Filter!header}
//  \index{itk::Binary\-Mask\-To\-Narrow\-Band\-Point\-Set\-Filter!header}
//
//  Software Guide : EndLatex
 
//  Software Guide : BeginCodeSnippet
#include "itkSpatialObjectToImageFilter.h"
#include "itkBinaryMaskToNarrowBandPointSetFilter.h"
//  Software Guide : EndCodeSnippet
 
#include "itkBinaryMaskToNarrowBandPointSetFilter.h"
 
#include "itkPointSet.h"
#include "itkPointSetToImageRegistrationMethod.h"
#include "itkNormalizedCorrelationPointSetToImageMetric.h"
#include "itkNearestNeighborInterpolateImageFunction.h"
#include "itkRigid2DTransform.h"
#include "itkRegularStepGradientDescentOptimizer.h"
#include "itkResampleImageFilter.h"
 
#include "itkImageFileReader.h"
#include "itkImageFileWriter.h"
 
//
// Observer to the optimizer
//
class CommandIterationUpdate : public itk::Command
{
public:
  using Self = CommandIterationUpdate;
  using Superclass = itk::Command;
  using Pointer = itk::SmartPointer<Self>;
  itkNewMacro(Self);
 
protected:
  CommandIterationUpdate() = default;
 
public:
  using OptimizerType = itk::RegularStepGradientDescentOptimizer;
  using OptimizerPointer = const OptimizerType *;
 
  void
  Execute(itk::Object * caller, const itk::EventObject & event) override
  {
    Execute((const itk::Object *)caller, event);
  }
 
  void
  Execute(const itk::Object * object, const itk::EventObject & event) override
  {
    auto optimizer = static_cast<OptimizerPointer>(object);
    if (typeid(event) != typeid(itk::IterationEvent))
    {
      return;
    }
 
    OptimizerType::DerivativeType gradient = optimizer->GetGradient();
    OptimizerType::ScalesType     scales = optimizer->GetScales();
 
    double magnitude2 = 0.0;
 
    for (unsigned int i = 0; i < gradient.size(); i++)
    {
      const double fc = gradient[i] / scales[i];
      magnitude2 += fc * fc;
    }
 
    const double gradientMagnitude = std::sqrt(magnitude2);
 
    std::cout << optimizer->GetCurrentIteration() << "   ";
    std::cout << optimizer->GetValue() << "   ";
    std::cout << gradientMagnitude << "   ";
    std::cout << optimizer->GetCurrentPosition() << std::endl;
  }
};
 
int
main(int argc, char * argv[])
{
 
  if (argc < 2)
  {
    std::cerr << "Missing argument" << std::endl;
    std::cerr << "Usage: " << std::endl;
    std::cerr << argv[0] << " movingImageFileName [initialX initialY] "
              << std::endl;
    std::cerr << "[rasterizedObjectFileName] [BoxSizeX BoxSizeY]"
              << std::endl;
    return EXIT_FAILURE;
  }
 
  constexpr unsigned int Dimension = 2;
 
  using MaskPixelType = unsigned char;
 
  using MaskImageType = itk::Image<MaskPixelType, Dimension>;
 
 
  using SpatialObjectType = itk::BoxSpatialObject<Dimension>;
 
  using SpatialObjectToImageFilterType =
    itk::SpatialObjectToImageFilter<SpatialObjectType, MaskImageType>;
 
  using FixedPointSetType = itk::PointSet<float, Dimension>;
 
 
  using NarrowBandFilterType =
    itk::BinaryMaskToNarrowBandPointSetFilter<MaskImageType,
                                              FixedPointSetType>;
 
  using PixelType = signed short;
 
  using ImageType = itk::Image<PixelType, Dimension>;
 
  using MaskPixelType = unsigned char;
 
  using MaskImageType = itk::Image<MaskPixelType, Dimension>;
 
 
  using TransformType = itk::Rigid2DTransform<double>;
 
  using ParametersType = TransformType::ParametersType;
 
 
  using OptimizerType = itk::RegularStepGradientDescentOptimizer;
 
  using LinearInterpolatorType =
    itk::LinearInterpolateImageFunction<ImageType, double>;
 
 
  using MetricType =
    itk::NormalizedCorrelationPointSetToImageMetric<FixedPointSetType,
                                                    ImageType>;
 
 
  using OptimizerScalesType = OptimizerType::ScalesType;
 
 
  using RegistrationType =
    itk::PointSetToImageRegistrationMethod<FixedPointSetType, ImageType>;
 
 
  using IterationObserverType = CommandIterationUpdate;
 
  using ImageReaderType = itk::ImageFileReader<ImageType>;
 
  SpatialObjectType::Pointer      spatialObject;
  TransformType::Pointer          transform;
  OptimizerType::Pointer          optimizer;
  IterationObserverType::Pointer  iterationObserver;
  LinearInterpolatorType::Pointer linearInterpolator;
  MetricType::Pointer             metric;
  RegistrationType::Pointer       registrationMethod;
  ImageReaderType::Pointer        movingImageReader;
  FixedPointSetType::Pointer      fixedPointSet;
  ImageType::ConstPointer         movingImage;
 
  SpatialObjectToImageFilterType::Pointer rasterizationFilter;
  NarrowBandFilterType::Pointer           narrowBandPointSetFilter;
 
 
  metric = MetricType::New();
  transform = TransformType::New();
  optimizer = OptimizerType::New();
  linearInterpolator = LinearInterpolatorType::New();
  registrationMethod = RegistrationType::New();
  iterationObserver = IterationObserverType::New();
 
  spatialObject = SpatialObjectType::New();
  rasterizationFilter = SpatialObjectToImageFilterType::New();
  narrowBandPointSetFilter = NarrowBandFilterType::New();
 
  movingImageReader = ImageReaderType::New();
 
  movingImageReader->SetFileName(argv[1]);
 
  try
  {
    movingImageReader->Update();
  }
  catch (const itk::ExceptionObject & excp)
  {
    std::cerr << "Problem reading Moving image from = " << std::endl;
    std::cerr << argv[1] << std::endl;
    std::cerr << excp << std::endl;
    return EXIT_FAILURE;
  }
 
 
  movingImage = movingImageReader->GetOutput();
 
  SpatialObjectType::SizeType boxSize;
  boxSize[0] = 60.0; // mm
  boxSize[1] = 60.0; // mm
 
  if (argc > 6)
  {
    boxSize[0] = std::stod(argv[5]);
    boxSize[1] = std::stod(argv[6]);
  }
 
  //
  // The geometry of the BoxSpatialObject is such that one of
  // its corners is located at the origin of coordinates.
  //
  spatialObject->SetSizeInObjectSpace(boxSize);
 
  ImageType::RegionType region = movingImage->GetLargestPossibleRegion();
 
  ImageType::SizeType imageSize = region.GetSize();
 
  ImageType::SpacingType spacing = movingImage->GetSpacing();
  ImageType::PointType   origin;
  origin[0] = (boxSize[0] - imageSize[0] * spacing[0]) / 2.0;
  origin[1] = (boxSize[1] - imageSize[1] * spacing[1]) / 2.0;
 
  rasterizationFilter->SetInput(spatialObject);
  rasterizationFilter->SetSize(imageSize);
  rasterizationFilter->SetSpacing(spacing);
  rasterizationFilter->SetOrigin(origin);
 
 
  narrowBandPointSetFilter->SetBandWidth(5.0);
 
  narrowBandPointSetFilter->SetInput(rasterizationFilter->GetOutput());
 
  narrowBandPointSetFilter->Update();
 
  if (argc > 4)
  {
    using MaskWriterType = itk::ImageFileWriter<MaskImageType>;
    MaskWriterType::Pointer maskWriter = MaskWriterType::New();
    maskWriter->SetInput(rasterizationFilter->GetOutput());
    maskWriter->SetFileName(argv[4]);
    maskWriter->Update();
  }
 
  fixedPointSet = narrowBandPointSetFilter->GetOutput();
 
  fixedPointSet->Print(std::cout);
 
  registrationMethod->SetOptimizer(optimizer);
  registrationMethod->SetInterpolator(linearInterpolator);
  registrationMethod->SetMetric(metric);
  registrationMethod->SetTransform(transform);
 
  registrationMethod->SetMovingImage(movingImage);
  registrationMethod->SetFixedPointSet(fixedPointSet);
 
 
  optimizer->SetMaximumStepLength(1.00);
  optimizer->SetMinimumStepLength(0.001);
  optimizer->SetNumberOfIterations(300);
  optimizer->SetRelaxationFactor(0.90);
  optimizer->SetGradientMagnitudeTolerance(0.05);
  optimizer->MinimizeOn();
  optimizer->AddObserver(itk::IterationEvent(), iterationObserver);
 
  TransformType::TranslationType initialTranslation;
  initialTranslation[0] = 0.0;
  initialTranslation[1] = 0.0;
 
  if (argc >= 4)
  {
    initialTranslation[0] = std::stod(argv[2]);
    initialTranslation[1] = std::stod(argv[3]);
  }
 
 
  TransformType::OutputPointType rotationCenter;
  rotationCenter[0] = boxSize[0] / 2.0;
  rotationCenter[1] = boxSize[1] / 2.0;
 
  transform->SetIdentity();
  transform->SetCenter(rotationCenter);
  transform->SetTranslation(initialTranslation);
 
  registrationMethod->SetInitialTransformParameters(
    transform->GetParameters());
 
  OptimizerScalesType optimizerScales(transform->GetNumberOfParameters());
 
  const double translationScale = 1.0 / 1000.0;
 
  optimizerScales[0] = 1.0;
  optimizerScales[1] = translationScale;
  optimizerScales[2] = translationScale;
 
  optimizer->SetScales(optimizerScales);
 
  try
  {
    registrationMethod->Update();
    std::cout
      << "Optimizer stop condition: "
      << registrationMethod->GetOptimizer()->GetStopConditionDescription()
      << std::endl;
  }
  catch (const itk::ExceptionObject & excp)
  {
    std::cerr << "Problem found during the registration" << std::endl;
    std::cerr << argv[1] << std::endl;
    std::cerr << excp << std::endl;
    return EXIT_FAILURE;
  }
 
  ParametersType transformParameters =
    registrationMethod->GetLastTransformParameters();
 
 
  TransformType::OutputPointType center = transform->GetCenter();
 
  std::cout << "Registration parameter = " << std::endl;
  std::cout << "Rotation center = " << center << std::endl;
  std::cout << "Parameters = " << transformParameters << std::endl;
 
 
  return EXIT_SUCCESS;
}