1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350
|
/*=========================================================================
*
* 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.
*
*=========================================================================*/
#include "itkRegistrationParameterScalesFromJacobian.h"
#include "itkImageToImageMetricv4.h"
#include "itkAffineTransform.h"
#include "itkDisplacementFieldTransform.h"
#include "itkMath.h"
/**
* \class RegistrationParameterScalesFromJacobianTestMetric for test.
* Create a simple metric to use for testing here.
*/
template <typename TFixedImage, typename TMovingImage, typename TVirtualImage = TFixedImage>
class RegistrationParameterScalesFromJacobianTestMetric
: public itk::ImageToImageMetricv4<TFixedImage, TMovingImage, TVirtualImage>
{
public:
/** Standard class type aliases. */
using Self = RegistrationParameterScalesFromJacobianTestMetric;
using Superclass = itk::ImageToImageMetricv4<TFixedImage, TMovingImage, TVirtualImage>;
using Pointer = itk::SmartPointer<Self>;
using ConstPointer = itk::SmartPointer<const Self>;
using typename Superclass::MeasureType;
using typename Superclass::DerivativeType;
using typename Superclass::ParametersType;
using typename Superclass::ParametersValueType;
itkOverrideGetNameOfClassMacro(RegistrationParameterScalesFromJacobianTestMetric);
itkNewMacro(Self);
// Pure virtual functions that all Metrics must provide
unsigned int
GetNumberOfParameters() const override
{
return 5;
}
MeasureType
GetValue() const override
{
return 1.0;
}
void
GetValueAndDerivative(MeasureType & value, DerivativeType & derivative) const override
{
value = 1.0;
derivative.Fill(0.0);
}
unsigned int
GetNumberOfLocalParameters() const override
{
return 0;
}
void
UpdateTransformParameters(const DerivativeType &, ParametersValueType) override
{}
const ParametersType &
GetParameters() const override
{
return m_Parameters;
}
void
Initialize() override
{}
ParametersType m_Parameters;
// Image related types
using FixedImageType = TFixedImage;
using MovingImageType = TMovingImage;
using VirtualImageType = TVirtualImage;
using FixedImageConstPointer = typename FixedImageType::ConstPointer;
using MovingImageConstPointer = typename MovingImageType::ConstPointer;
using VirtualImagePointer = typename VirtualImageType::Pointer;
using VirtualRegionType = typename VirtualImageType::RegionType;
/* Image dimension accessors */
static constexpr itk::SizeValueType FixedImageDimension = FixedImageType::ImageDimension;
static constexpr itk::SizeValueType MovingImageDimension = MovingImageType::ImageDimension;
static constexpr itk::SizeValueType VirtualImageDimension = VirtualImageType::ImageDimension;
private:
RegistrationParameterScalesFromJacobianTestMetric() = default;
~RegistrationParameterScalesFromJacobianTestMetric() override = default;
};
/**
*/
int
itkRegistrationParameterScalesFromJacobianTest(int, char *[])
{
// Image begins
constexpr itk::SizeValueType ImageDimension = 2;
using PixelType = double;
using FloatType = double;
// Image Types
using FixedImageType = itk::Image<PixelType, ImageDimension>;
using MovingImageType = itk::Image<PixelType, ImageDimension>;
using VirtualImageType = itk::Image<PixelType, ImageDimension>;
auto fixedImage = FixedImageType::New();
auto movingImage = MovingImageType::New();
VirtualImageType::Pointer virtualImage = fixedImage;
MovingImageType::SizeType size;
size.Fill(100);
movingImage->SetRegions(size);
fixedImage->SetRegions(size);
// Image done
// Transform begins
using MovingTransformType = itk::AffineTransform<double, ImageDimension>;
auto movingTransform = MovingTransformType::New();
movingTransform->SetIdentity();
using FixedTransformType = itk::TranslationTransform<double, ImageDimension>;
auto fixedTransform = FixedTransformType::New();
fixedTransform->SetIdentity();
// Transform done
// Metric begins
using MetricType = RegistrationParameterScalesFromJacobianTestMetric<FixedImageType, MovingImageType>;
auto metric = MetricType::New();
metric->SetVirtualDomainFromImage(virtualImage);
metric->SetFixedImage(fixedImage);
metric->SetMovingImage(movingImage);
metric->SetFixedTransform(fixedTransform);
metric->SetMovingTransform(movingTransform);
// Metric done
// Scales for the affine transform from transform jacobians
using RegistrationParameterScalesFromJacobianType = itk::RegistrationParameterScalesFromJacobian<MetricType>;
RegistrationParameterScalesFromJacobianType::Pointer jacobianScaleEstimator =
RegistrationParameterScalesFromJacobianType::New();
jacobianScaleEstimator->SetMetric(metric);
jacobianScaleEstimator->SetTransformForward(true); // by default
jacobianScaleEstimator->Print(std::cout);
std::cout << std::endl;
RegistrationParameterScalesFromJacobianType::ScalesType jacobianScales(movingTransform->GetNumberOfParameters());
jacobianScaleEstimator->EstimateScales(jacobianScales);
std::cout << "Jacobian scales for the affine transform = " << jacobianScales << std::endl;
// Check the correctness
RegistrationParameterScalesFromJacobianType::ScalesType theoreticalJacobianScales(
movingTransform->GetNumberOfParameters());
VirtualImageType::PointType upperPoint;
virtualImage->TransformIndexToPhysicalPoint(virtualImage->GetLargestPossibleRegion().GetUpperIndex(), upperPoint);
itk::SizeValueType param = 0;
for (itk::SizeValueType row = 0; row < ImageDimension; ++row)
{
for (itk::SizeValueType col = 0; col < ImageDimension; ++col)
{
// uses the corners for affine transform
// = (0 + 0 + n*n + n*n)/4 = n*n/2
theoreticalJacobianScales[param++] = upperPoint[col] * upperPoint[col] / 2.0;
}
}
for (itk::SizeValueType row = 0; row < ImageDimension; ++row)
{
theoreticalJacobianScales[param++] = 1;
}
bool jacobianPass = true;
for (itk::SizeValueType p = 0; p < jacobianScales.GetSize(); ++p)
{
if (itk::Math::abs((jacobianScales[p] - theoreticalJacobianScales[p]) / theoreticalJacobianScales[p]) > 0.01)
{
jacobianPass = false;
break;
}
}
if (!jacobianPass)
{
std::cout << "Failed: the jacobian scales for the affine transform are not correct." << std::endl;
}
else
{
std::cout << "Passed: the jacobian scales for the affine transform are correct." << std::endl;
}
bool nonUniformForJacobian = false;
for (itk::SizeValueType p = 1; p < jacobianScales.GetSize(); ++p)
{
if (itk::Math::NotExactlyEquals(jacobianScales[p], jacobianScales[0]))
{
nonUniformForJacobian = true;
break;
}
}
if (!nonUniformForJacobian)
{
std::cout << "Error: the jacobian scales for an affine transform are equal for all parameters." << std::endl;
}
// Testing the step scale for the affine transform
MovingTransformType::ParametersType movingStep(movingTransform->GetNumberOfParameters());
movingStep = movingTransform->GetParameters();
FloatType stepScale = jacobianScaleEstimator->EstimateStepScale(movingStep);
std::cout << "The step scale of Jacobian for the affine transform = " << stepScale << std::endl;
FloatType learningRate = 1.0 / stepScale;
std::cout << "The learning rate of Jacobian for the affine transform = " << learningRate << std::endl;
FloatType theoreticalStepScale = 0.0;
FloatType count = 0.0;
VirtualImageType::PointType lowerPoint;
virtualImage->TransformIndexToPhysicalPoint(virtualImage->GetLargestPossibleRegion().GetIndex(), lowerPoint);
for (FloatType x = lowerPoint[0]; x <= upperPoint[0]; x += upperPoint[0] - lowerPoint[0])
{
for (FloatType y = lowerPoint[1]; y <= upperPoint[1]; y += upperPoint[1] - lowerPoint[1])
{
theoreticalStepScale += std::sqrt(x * x + y * y);
count++;
}
}
theoreticalStepScale /= count;
bool stepScalePass = false;
if (itk::Math::abs((stepScale - theoreticalStepScale) / theoreticalStepScale) < 0.01)
{
stepScalePass = true;
}
if (!stepScalePass)
{
std::cout << "Failed: the step scale for the affine transform is not correct." << std::endl;
}
else
{
std::cout << "Passed: the step scale for the affine transform is correct." << std::endl;
}
// Testing local scales for a transform with local support, ex. DisplacementFieldTransform
using DisplacementTransformType = itk::DisplacementFieldTransform<double, ImageDimension>;
using FieldType = DisplacementTransformType::DisplacementFieldType;
using VectorType = itk::Vector<double, ImageDimension>;
VectorType zero;
zero.Fill(0.0);
auto field = FieldType::New();
field->SetRegions(virtualImage->GetLargestPossibleRegion());
field->SetSpacing(virtualImage->GetSpacing());
field->SetOrigin(virtualImage->GetOrigin());
field->SetDirection(virtualImage->GetDirection());
field->Allocate();
field->FillBuffer(zero);
auto displacementTransform = DisplacementTransformType::New();
displacementTransform->SetDisplacementField(field);
metric->SetMovingTransform(displacementTransform);
jacobianScaleEstimator->SetTransformForward(true);
RegistrationParameterScalesFromJacobianType::ScalesType localScales;
jacobianScaleEstimator->EstimateScales(localScales);
std::cout << "Shift scales for the displacement field transform = " << localScales << std::endl;
// Check the correctness
RegistrationParameterScalesFromJacobianType::ScalesType theoreticalLocalScales(
displacementTransform->GetNumberOfLocalParameters());
theoreticalLocalScales.Fill(1.0);
bool displacementPass = true;
for (itk::SizeValueType p = 0; p < theoreticalLocalScales.GetSize(); ++p)
{
if (itk::Math::abs((localScales[p] - theoreticalLocalScales[p]) / theoreticalLocalScales[p]) > 0.01)
{
displacementPass = false;
break;
}
}
if (!displacementPass)
{
std::cout << "Failed: the shift scales for the displacement field transform are not correct." << std::endl;
}
else
{
std::cout << "Passed: the shift scales for the displacement field transform are correct." << std::endl;
}
// Testing scales with local support done
// Testing the step scale for the displacement field transform
DisplacementTransformType::ParametersType displacementStep(displacementTransform->GetNumberOfParameters());
displacementStep.Fill(1.0);
FloatType localStepScale = jacobianScaleEstimator->EstimateStepScale(displacementStep);
std::cout << "The step scale of Jacobian for the displacement field transform = " << localStepScale << std::endl;
FloatType localLearningRate = 1.0 / localStepScale;
std::cout << "The learning rate of Jacobian for the displacement field transform = " << localLearningRate
<< std::endl;
bool localStepScalePass = false;
FloatType theoreticalLocalStepScale = std::sqrt(2.0);
if (itk::Math::abs((localStepScale - theoreticalLocalStepScale) / theoreticalLocalStepScale) < 0.01)
{
localStepScalePass = true;
}
if (!localStepScalePass)
{
std::cout << "Failed: the step scale for the displacement field transform is not correct." << std::endl;
}
else
{
std::cout << "Passed: the step scale for the displacement field transform is correct." << std::endl;
}
// Testing the step scale with local support done
// Check the correctness of all cases above
std::cout << std::endl;
if (jacobianPass && nonUniformForJacobian && stepScalePass && displacementPass && localStepScalePass)
{
std::cout << "Test passed" << std::endl;
return EXIT_SUCCESS;
}
else
{
std::cout << "Test failed" << std::endl;
return EXIT_FAILURE;
}
}
|