TransformMatrixType GetVoxelSpaceToRASPhysicalSpaceMatrix(typename ImageType::Pointer image)
{
// Generate intermediate terms
vnl_matrix<double> m_dir, m_ras_matrix;
vnl_diag_matrix<double> m_scale, m_lps_to_ras;
vnl_vector<double> v_origin, v_ras_offset;
// Compute the matrix
m_dir = image->GetDirection().GetVnlMatrix();
m_scale.set(image->GetSpacing().GetVnlVector());
m_lps_to_ras.set(vnl_vector<double>(ImageType::ImageDimension, 1.0));
m_lps_to_ras[0] = -1;
m_lps_to_ras[1] = -1;
m_ras_matrix = m_lps_to_ras * m_dir * m_scale;
// Compute the vector
v_origin = image->GetOrigin().GetVnlVector();
v_ras_offset = m_lps_to_ras * v_origin;
// Create the larger matrix
TransformMatrixType mat;
vnl_vector<double> vcol(ImageType::ImageDimension+1, 1.0);
vcol.update(v_ras_offset);
mat.SetIdentity();
mat.GetVnlMatrix().update(m_ras_matrix);
mat.GetVnlMatrix().set_column(ImageType::ImageDimension, vcol);
return mat;
}
void operator()(Parameters& params)
{
typedef typename ::itk::Image< PIXELTYPE, 3 > ImageType;
const typename ImageType::Pointer itkImage = ::fwItkIO::itkImageFactory< ImageType >(params.i_image);
typename ::itk::ResampleImageFilter<ImageType, ImageType>::Pointer resampler =
::itk::ResampleImageFilter<ImageType, ImageType>::New();
typename ::itk::MinimumMaximumImageCalculator< ImageType >::Pointer minCalculator =
::itk::MinimumMaximumImageCalculator< ImageType >::New();
minCalculator->SetImage(itkImage);
minCalculator->ComputeMinimum();
resampler->SetDefaultPixelValue(minCalculator->GetMinimum());
resampler->SetTransform(params.i_trf.GetPointer());
resampler->SetInput(itkImage);
typename ImageType::SizeType size = itkImage->GetLargestPossibleRegion().GetSize();
typename ImageType::PointType origin = itkImage->GetOrigin();
typename ImageType::SpacingType spacing = itkImage->GetSpacing();
typename ImageType::DirectionType direction = itkImage->GetDirection();
SLM_ASSERT("Input spacing can't be null along any axis", spacing[0] > 0 && spacing[1] > 0 && spacing[2] > 0);
if(params.i_targetImage)
{
for(std::uint8_t i = 0; i < 3; ++i)
{
// ITK uses unsigned long to store sizes.
size[i] = static_cast<typename ImageType::SizeType::SizeValueType>(params.i_targetImage->getSize()[i]);
origin[i] = params.i_targetImage->getOrigin()[i];
spacing[i] = params.i_targetImage->getSpacing()[i];
SLM_ASSERT("Output spacing can't be null along any axis.", spacing[i] > 0);
}
}
resampler->SetSize(size);
resampler->SetOutputOrigin(origin);
resampler->SetOutputDirection(direction);
resampler->SetOutputSpacing(spacing);
resampler->Update();
typename ImageType::Pointer outputImage = resampler->GetOutput();
::fwItkIO::itkImageToFwDataImage(outputImage, params.o_image);
}
template <class ImagePixelType, class MaskPixelType> int update()
{
typedef itk::Image<ImagePixelType, 3> ImageType;
typedef itk::Image<MaskPixelType, 3> MaskType;
if ( !input ||!input->data() || !mask ||!mask->data())
return EXIT_FAILURE;
typedef itk::MaskImageFilter< ImageType, MaskType> MaskFilterType;
typename MaskFilterType::Pointer maskFilter = MaskFilterType::New();
typename ImageType::Pointer imgInput = dynamic_cast<ImageType *> ( ( itk::Object* ) ( input->data() )) ;
typename MaskType::Pointer maskInput = dynamic_cast<MaskType *> ( ( itk::Object* ) ( mask->data() )) ;
try
{
maskInput->SetOrigin(imgInput->GetOrigin());
maskInput->SetSpacing(imgInput->GetSpacing());
maskFilter->SetInput(imgInput);
maskFilter->SetMaskingValue(maskBackgroundValue);
maskFilter->SetMaskImage(maskInput);
//Outside values set to the lowest reachable value
typedef itk::MinimumMaximumImageCalculator <ImageType> ImageCalculatorFilterType;
typename ImageCalculatorFilterType::Pointer imageCalculatorFilter
= ImageCalculatorFilterType::New ();
imageCalculatorFilter->SetImage(imgInput);
imageCalculatorFilter->ComputeMinimum();
maskFilter->SetOutsideValue(std::min(double(imageCalculatorFilter->GetMinimum()), 0.0));
maskFilter->Update();
output->setData(maskFilter->GetOutput());
}
catch( itk::ExceptionObject & err )
{
std::cerr << "ExceptionObject caught in medMaskApplication!" << std::endl;
std::cerr << err << std::endl;
return EXIT_FAILURE;
}
medUtilities::setDerivedMetaData(output, input, "masked");
return EXIT_SUCCESS;
}
template <class inputType, unsigned int Dimension> medAbstractJob::medJobExitStatus medItkBiasCorrectionProcess::N4BiasCorrectionCore()
{
medJobExitStatus eRes = medAbstractJob::MED_JOB_EXIT_SUCCESS;
typedef itk::Image<inputType, Dimension > ImageType;
typedef itk::Image <float, Dimension> OutputImageType;
typedef itk::Image<unsigned char, Dimension> MaskImageType;
typedef itk::N4BiasFieldCorrectionImageFilter<OutputImageType, MaskImageType, OutputImageType> BiasFilter;
typedef itk::ConstantPadImageFilter<OutputImageType, OutputImageType> PadderType;
typedef itk::ConstantPadImageFilter<MaskImageType, MaskImageType> MaskPadderType;
typedef itk::ShrinkImageFilter<OutputImageType, OutputImageType> ShrinkerType;
typedef itk::ShrinkImageFilter<MaskImageType, MaskImageType> MaskShrinkerType;
typedef itk::BSplineControlPointImageFilter<typename BiasFilter::BiasFieldControlPointLatticeType, typename BiasFilter::ScalarImageType> BSplinerType;
typedef itk::ExpImageFilter<OutputImageType, OutputImageType> ExpFilterType;
typedef itk::DivideImageFilter<OutputImageType, OutputImageType, OutputImageType> DividerType;
typedef itk::ExtractImageFilter<OutputImageType, OutputImageType> CropperType;
unsigned int uiThreadNb = static_cast<unsigned int>(m_poUIThreadNb->value());
unsigned int uiShrinkFactors = static_cast<unsigned int>(m_poUIShrinkFactors->value());
unsigned int uiSplineOrder = static_cast<unsigned int>(m_poUISplineOrder->value());
float fWienerFilterNoise = static_cast<float>(m_poFWienerFilterNoise->value());
float fbfFWHM = static_cast<float>(m_poFbfFWHM->value());
float fConvergenceThreshold = static_cast<float>(m_poFConvergenceThreshold->value());
float fSplineDistance = static_cast<float>(m_poFSplineDistance->value());
float fProgression = 0;
QStringList oListValue = m_poSMaxIterations->value().split("x");
std::vector<unsigned int> oMaxNumbersIterationsVector(oListValue.size());
std::vector<float> oInitialMeshResolutionVect(Dimension);
for (int i=0; i<oMaxNumbersIterationsVector.size(); ++i)
{
oMaxNumbersIterationsVector[i] = (unsigned int)oListValue[i].toInt();
}
oInitialMeshResolutionVect[0] = static_cast<float>(m_poFInitialMeshResolutionVect1->value());
oInitialMeshResolutionVect[1] = static_cast<float>(m_poFInitialMeshResolutionVect2->value());
oInitialMeshResolutionVect[2] = static_cast<float>(m_poFInitialMeshResolutionVect3->value());
typename ImageType::Pointer image = dynamic_cast<ImageType *>((itk::Object*)(this->input()->data()));
typedef itk::CastImageFilter <ImageType, OutputImageType> CastFilterType;
typename CastFilterType::Pointer castFilter = CastFilterType::New();
castFilter->SetInput(image);
/********************************************************************************/
/***************************** PREPARING STARTING *******************************/
/********************************************************************************/
/*** 0 ******************* Create filter and accessories ******************/
ABORT_CHECKING(m_bAborting);
typename BiasFilter::Pointer filter = BiasFilter::New();
typename BiasFilter::ArrayType oNumberOfControlPointsArray;
m_filter = filter;
/*** 1 ******************* Read input image *******************************/
ABORT_CHECKING(m_bAborting);
fProgression = 1;
updateProgression(fProgression);
/*** 2 ******************* Creating Otsu mask *****************************/
ABORT_CHECKING(m_bAborting);
itk::TimeProbe timer;
timer.Start();
typename MaskImageType::Pointer maskImage = ITK_NULLPTR;
typedef itk::OtsuThresholdImageFilter<OutputImageType, MaskImageType> ThresholderType;
typename ThresholderType::Pointer otsu = ThresholderType::New();
m_filter = otsu;
otsu->SetInput(castFilter->GetOutput());
otsu->SetNumberOfHistogramBins(200);
otsu->SetInsideValue(0);
otsu->SetOutsideValue(1);
otsu->SetNumberOfThreads(uiThreadNb);
otsu->Update();
updateProgression(fProgression);
maskImage = otsu->GetOutput();
/*** 3A *************** Set Maximum number of Iterations for the filter ***/
ABORT_CHECKING(m_bAborting);
typename BiasFilter::VariableSizeArrayType itkTabMaximumIterations;
itkTabMaximumIterations.SetSize(oMaxNumbersIterationsVector.size());
for (int i = 0; i < oMaxNumbersIterationsVector.size(); ++i)
{
itkTabMaximumIterations[i] = oMaxNumbersIterationsVector[i];
}
filter->SetMaximumNumberOfIterations(itkTabMaximumIterations);
/*** 3B *************** Set Fitting Levels for the filter *****************/
typename BiasFilter::ArrayType oFittingLevelsTab;
oFittingLevelsTab.Fill(oMaxNumbersIterationsVector.size());
filter->SetNumberOfFittingLevels(oFittingLevelsTab);
updateProgression(fProgression);
/*** 4 ******************* Save image's index, size, origine **************/
ABORT_CHECKING(m_bAborting);
typename ImageType::IndexType oImageIndex = image->GetLargestPossibleRegion().GetIndex();
typename ImageType::SizeType oImageSize = image->GetLargestPossibleRegion().GetSize();
typename ImageType::PointType newOrigin = image->GetOrigin();
typename OutputImageType::Pointer outImage = castFilter->GetOutput();
if (fSplineDistance > 0)
{
/*** 5 ******************* Compute number of control points **************/
ABORT_CHECKING(m_bAborting);
itk::SizeValueType lowerBound[3];
itk::SizeValueType upperBound[3];
for (unsigned int i = 0; i < 3; i++)
{
float domain = static_cast<float>(image->GetLargestPossibleRegion().GetSize()[i] - 1) * image->GetSpacing()[i];
unsigned int numberOfSpans = static_cast<unsigned int>(std::ceil(domain / fSplineDistance));
unsigned long extraPadding = static_cast<unsigned long>((numberOfSpans * fSplineDistance - domain) / image->GetSpacing()[i] + 0.5);
lowerBound[i] = static_cast<unsigned long>(0.5 * extraPadding);
upperBound[i] = extraPadding - lowerBound[i];
newOrigin[i] -= (static_cast<float>(lowerBound[i]) * image->GetSpacing()[i]);
oNumberOfControlPointsArray[i] = numberOfSpans + filter->GetSplineOrder();
}
updateProgression(fProgression);
/*** 6 ******************* Padder ****************************************/
ABORT_CHECKING(m_bAborting);
typename PadderType::Pointer imagePadder = PadderType::New();
m_filter = imagePadder;
imagePadder->SetInput(castFilter->GetOutput());
imagePadder->SetPadLowerBound(lowerBound);
imagePadder->SetPadUpperBound(upperBound);
imagePadder->SetConstant(0);
imagePadder->SetNumberOfThreads(uiThreadNb);
imagePadder->Update();
updateProgression(fProgression);
outImage = imagePadder->GetOutput();
/*** 7 ******************** Handle the mask image *************************/
ABORT_CHECKING(m_bAborting);
typename MaskPadderType::Pointer maskPadder = MaskPadderType::New();
m_filter = maskPadder;
maskPadder->SetInput(maskImage);
maskPadder->SetPadLowerBound(lowerBound);
maskPadder->SetPadUpperBound(upperBound);
maskPadder->SetConstant(0);
maskPadder->SetNumberOfThreads(uiThreadNb);
maskPadder->Update();
updateProgression(fProgression);
maskImage = maskPadder->GetOutput();
/*** 8 ******************** SetNumber Of Control Points *******************/
ABORT_CHECKING(m_bAborting);
filter->SetNumberOfControlPoints(oNumberOfControlPointsArray);
}
else if (oInitialMeshResolutionVect.size() == 3)
{
/*** 9 ******************** SetNumber Of Control Points alternative *******/
ABORT_CHECKING(m_bAborting);
for (unsigned i = 0; i < 3; i++)
{
oNumberOfControlPointsArray[i] = static_cast<unsigned int>(oInitialMeshResolutionVect[i]) + filter->GetSplineOrder();
}
filter->SetNumberOfControlPoints(oNumberOfControlPointsArray);
updateProgression(fProgression, 3);
}
else
{
fProgression = 0;
updateProgression(fProgression);
std::cout << "No BSpline distance and Mesh Resolution is ignored because not 3 dimensions" << std::endl;
}
/*** 10 ******************* Shrinker image ********************************/
ABORT_CHECKING(m_bAborting);
typename ShrinkerType::Pointer imageShrinker = ShrinkerType::New();
m_filter = imageShrinker;
imageShrinker->SetInput(outImage);
/*** 11 ******************* Shrinker mask *********************************/
ABORT_CHECKING(m_bAborting);
typename MaskShrinkerType::Pointer maskShrinker = MaskShrinkerType::New();
m_filter = maskShrinker;
maskShrinker->SetInput(maskImage);
/*** 12 ******************* Shrink mask and image *************************/
ABORT_CHECKING(m_bAborting);
imageShrinker->SetShrinkFactors(uiShrinkFactors);
maskShrinker->SetShrinkFactors(uiShrinkFactors);
imageShrinker->SetNumberOfThreads(uiThreadNb);
maskShrinker->SetNumberOfThreads(uiThreadNb);
imageShrinker->Update();
updateProgression(fProgression);
maskShrinker->Update();
updateProgression(fProgression);
/*** 13 ******************* Filter setings ********************************/
ABORT_CHECKING(m_bAborting);
filter->SetSplineOrder(uiSplineOrder);
filter->SetWienerFilterNoise(fWienerFilterNoise);
filter->SetBiasFieldFullWidthAtHalfMaximum(fbfFWHM);
filter->SetConvergenceThreshold(fConvergenceThreshold);
filter->SetInput(imageShrinker->GetOutput());
filter->SetMaskImage(maskShrinker->GetOutput());
/*** 14 ******************* Apply filter **********************************/
ABORT_CHECKING(m_bAborting);
try
{
filter->SetNumberOfThreads(uiThreadNb);
filter->Update();
updateProgression(fProgression, 5);
}
catch (itk::ExceptionObject & err)
{
std::cerr << "ExceptionObject caught !" << std::endl;
std::cerr << err << std::endl;
eRes = medAbstractJob::MED_JOB_EXIT_FAILURE;
return eRes;
}
/**
* Reconstruct the bias field at full image resolution. Divide
* the original input image by the bias field to get the final
* corrected image.
*/
ABORT_CHECKING(m_bAborting);
typename BSplinerType::Pointer bspliner = BSplinerType::New();
m_filter = bspliner;
bspliner->SetInput(filter->GetLogBiasFieldControlPointLattice());
bspliner->SetSplineOrder(filter->GetSplineOrder());
bspliner->SetSize(image->GetLargestPossibleRegion().GetSize());
bspliner->SetOrigin(newOrigin);
bspliner->SetDirection(image->GetDirection());
bspliner->SetSpacing(image->GetSpacing());
bspliner->SetNumberOfThreads(uiThreadNb);
bspliner->Update();
updateProgression(fProgression);
/*********************** Logarithm phase ***************************/
ABORT_CHECKING(m_bAborting);
typename OutputImageType::Pointer logField = OutputImageType::New();
logField->SetOrigin(image->GetOrigin());
logField->SetSpacing(image->GetSpacing());
logField->SetRegions(image->GetLargestPossibleRegion());
logField->SetDirection(image->GetDirection());
logField->Allocate();
itk::ImageRegionIterator<typename BiasFilter::ScalarImageType> IB(bspliner->GetOutput(), bspliner->GetOutput()->GetLargestPossibleRegion());
itk::ImageRegionIterator<OutputImageType> IF(logField, logField->GetLargestPossibleRegion());
for (IB.GoToBegin(), IF.GoToBegin(); !IB.IsAtEnd(); ++IB, ++IF)
{
IF.Set(IB.Get()[0]);
}
/*********************** Exponential phase *************************/
ABORT_CHECKING(m_bAborting);
typename ExpFilterType::Pointer expFilter = ExpFilterType::New();
m_filter = expFilter;
expFilter->SetInput(logField);
expFilter->SetNumberOfThreads(uiThreadNb);
expFilter->Update();
updateProgression(fProgression);
/************************ Dividing phase ***************************/
ABORT_CHECKING(m_bAborting);
typename DividerType::Pointer divider = DividerType::New();
m_filter = divider;
divider->SetInput1(castFilter->GetOutput());
divider->SetInput2(expFilter->GetOutput());
divider->SetNumberOfThreads(uiThreadNb);
divider->Update();
updateProgression(fProgression);
/******************** Prepare cropping phase ***********************/
ABORT_CHECKING(m_bAborting);
typename ImageType::RegionType inputRegion;
inputRegion.SetIndex(oImageIndex);
inputRegion.SetSize(oImageSize);
/************************ Cropping phase ***************************/
ABORT_CHECKING(m_bAborting);
typename CropperType::Pointer cropper = CropperType::New();
m_filter = cropper;
cropper->SetInput(divider->GetOutput());
cropper->SetExtractionRegion(inputRegion);
cropper->SetDirectionCollapseToSubmatrix();
cropper->SetNumberOfThreads(uiThreadNb);
cropper->Update();
updateProgression(fProgression);
/********************** Write output image *************************/
ABORT_CHECKING(m_bAborting);
medAbstractImageData *out = qobject_cast<medAbstractImageData *>(medAbstractDataFactory::instance()->create("itkDataImageFloat3"));
out->setData(cropper->GetOutput());
this->setOutput(out);
m_filter = 0;
return eRes;
}