otbScalarImageToTexturesFilter.hxx

Go to the documentation of this file.

/*
 * Copyright (C) 2005-2022 Centre National d'Etudes Spatiales (CNES)
 *
 * This file is part of Orfeo Toolbox
 *
 *     https://www.orfeo-toolbox.org/
 *
 * 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
 *
 * 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 otbScalarImageToTexturesFilter_hxx
#define otbScalarImageToTexturesFilter_hxx
 
#include "otbScalarImageToTexturesFilter.h"
#include "itkImageRegionIteratorWithIndex.h"
#include "itkConstNeighborhoodIterator.h"
#include "itkImageRegionIterator.h"
#include "itkProgressReporter.h"
#include "itkNumericTraits.h"
#include <vector>
#include <cmath>
 
namespace otb
{
template <class TInputImage, class TOutputImage>
ScalarImageToTexturesFilter<TInputImage, TOutputImage>::ScalarImageToTexturesFilter()
  : m_Radius(),
    m_Offset(),
    m_NeighborhoodRadius(),
    m_NumberOfBinsPerAxis(8),
    m_InputImageMinimum(0),
    m_InputImageMaximum(255),
    m_SubsampleFactor(),
    m_SubsampleOffset()
{
  // There are 8 outputs corresponding to the 8 textures indices
  this->SetNumberOfRequiredOutputs(8);
 
  // Create the 8 outputs
  this->SetNthOutput(0, OutputImageType::New());
  this->SetNthOutput(1, OutputImageType::New());
  this->SetNthOutput(2, OutputImageType::New());
  this->SetNthOutput(3, OutputImageType::New());
  this->SetNthOutput(4, OutputImageType::New());
  this->SetNthOutput(5, OutputImageType::New());
  this->SetNthOutput(6, OutputImageType::New());
  this->SetNthOutput(7, OutputImageType::New());
 
  this->m_SubsampleFactor.Fill(1);
  this->m_SubsampleOffset.Fill(0);
}
 
template <class TInputImage, class TOutputImage>
ScalarImageToTexturesFilter<TInputImage, TOutputImage>::~ScalarImageToTexturesFilter()
{
}
 
template <class TInputImage, class TOutputImage>
typename ScalarImageToTexturesFilter<TInputImage, TOutputImage>::OutputImageType* ScalarImageToTexturesFilter<TInputImage, TOutputImage>::GetEnergyOutput()
{
  if (this->GetNumberOfOutputs() < 1)
  {
    return nullptr;
  }
  return static_cast<OutputImageType*>(this->GetOutput(0));
}
 
template <class TInputImage, class TOutputImage>
typename ScalarImageToTexturesFilter<TInputImage, TOutputImage>::OutputImageType* ScalarImageToTexturesFilter<TInputImage, TOutputImage>::GetEntropyOutput()
{
  if (this->GetNumberOfOutputs() < 2)
  {
    return nullptr;
  }
  return static_cast<OutputImageType*>(this->GetOutput(1));
}
 
template <class TInputImage, class TOutputImage>
typename ScalarImageToTexturesFilter<TInputImage, TOutputImage>::OutputImageType* ScalarImageToTexturesFilter<TInputImage, TOutputImage>::GetCorrelationOutput()
{
  if (this->GetNumberOfOutputs() < 3)
  {
    return nullptr;
  }
  return static_cast<OutputImageType*>(this->GetOutput(2));
}
 
template <class TInputImage, class TOutputImage>
typename ScalarImageToTexturesFilter<TInputImage, TOutputImage>::OutputImageType*
ScalarImageToTexturesFilter<TInputImage, TOutputImage>::GetInverseDifferenceMomentOutput()
{
  if (this->GetNumberOfOutputs() < 4)
  {
    return nullptr;
  }
  return static_cast<OutputImageType*>(this->GetOutput(3));
}
 
template <class TInputImage, class TOutputImage>
typename ScalarImageToTexturesFilter<TInputImage, TOutputImage>::OutputImageType* ScalarImageToTexturesFilter<TInputImage, TOutputImage>::GetInertiaOutput()
{
  if (this->GetNumberOfOutputs() < 5)
  {
    return nullptr;
  }
  return static_cast<OutputImageType*>(this->GetOutput(4));
}
 
template <class TInputImage, class TOutputImage>
typename ScalarImageToTexturesFilter<TInputImage, TOutputImage>::OutputImageType*
ScalarImageToTexturesFilter<TInputImage, TOutputImage>::GetClusterShadeOutput()
{
  if (this->GetNumberOfOutputs() < 6)
  {
    return nullptr;
  }
  return static_cast<OutputImageType*>(this->GetOutput(5));
}
 
template <class TInputImage, class TOutputImage>
typename ScalarImageToTexturesFilter<TInputImage, TOutputImage>::OutputImageType*
ScalarImageToTexturesFilter<TInputImage, TOutputImage>::GetClusterProminenceOutput()
{
  if (this->GetNumberOfOutputs() < 7)
  {
    return nullptr;
  }
  return static_cast<OutputImageType*>(this->GetOutput(6));
}
 
template <class TInputImage, class TOutputImage>
typename ScalarImageToTexturesFilter<TInputImage, TOutputImage>::OutputImageType*
ScalarImageToTexturesFilter<TInputImage, TOutputImage>::GetHaralickCorrelationOutput()
{
  if (this->GetNumberOfOutputs() < 8)
  {
    return nullptr;
  }
  return static_cast<OutputImageType*>(this->GetOutput(7));
}
 
template <class TInputImage, class TOutputImage>
void ScalarImageToTexturesFilter<TInputImage, TOutputImage>::GenerateOutputInformation()
{
  // First, call superclass implementation
  Superclass::GenerateOutputInformation();
 
  // Compute output size, origin & spacing
  InputRegionType  inputRegion = this->GetInput()->GetLargestPossibleRegion();
  OutputRegionType outputRegion;
  outputRegion.SetIndex(0, 0);
  outputRegion.SetIndex(1, 0);
  outputRegion.SetSize(0, 1 + (inputRegion.GetSize(0) - 1 - m_SubsampleOffset[0]) / m_SubsampleFactor[0]);
  outputRegion.SetSize(1, 1 + (inputRegion.GetSize(1) - 1 - m_SubsampleOffset[1]) / m_SubsampleFactor[1]);
 
  typename OutputImageType::SpacingType outSpacing = this->GetInput()->GetSignedSpacing();
  outSpacing[0] *= m_SubsampleFactor[0];
  outSpacing[1] *= m_SubsampleFactor[1];
 
  typename OutputImageType::PointType outOrigin;
  this->GetInput()->TransformIndexToPhysicalPoint(inputRegion.GetIndex() + m_SubsampleOffset, outOrigin);
 
  for (unsigned int i = 0; i < this->GetNumberOfOutputs(); i++)
  {
    OutputImagePointerType outputPtr = this->GetOutput(i);
    outputPtr->SetLargestPossibleRegion(outputRegion);
    outputPtr->SetOrigin(outOrigin);
    outputPtr->SetSignedSpacing(outSpacing);
  }
}
 
template <class TInputImage, class TOutputImage>
void ScalarImageToTexturesFilter<TInputImage, TOutputImage>::GenerateInputRequestedRegion()
{
  // First, call superclass implementation
  Superclass::GenerateInputRequestedRegion();
 
  // Retrieve the input and output pointers
  InputImagePointerType  inputPtr  = const_cast<InputImageType*>(this->GetInput());
  OutputImagePointerType outputPtr = this->GetOutput();
 
  if (!inputPtr || !outputPtr)
  {
    return;
  }
 
  // Retrieve the output requested region
  // We use only the first output since requested regions for all outputs are enforced to be equal
  // by the default GenerateOutputRequestedRegiont() implementation
  OutputRegionType outputRequestedRegion = outputPtr->GetRequestedRegion();
 
  typename OutputRegionType::IndexType outputIndex = outputRequestedRegion.GetIndex();
  typename OutputRegionType::SizeType  outputSize  = outputRequestedRegion.GetSize();
  typename InputRegionType::IndexType  inputIndex;
  typename InputRegionType::SizeType   inputSize;
  InputRegionType                      inputLargest = inputPtr->GetLargestPossibleRegion();
 
  // Convert index and size to full grid
  outputIndex[0] = outputIndex[0] * m_SubsampleFactor[0] + m_SubsampleOffset[0] + inputLargest.GetIndex(0);
  outputIndex[1] = outputIndex[1] * m_SubsampleFactor[1] + m_SubsampleOffset[1] + inputLargest.GetIndex(1);
  outputSize[0]  = 1 + (outputSize[0] - 1) * m_SubsampleFactor[0];
  outputSize[1]  = 1 + (outputSize[1] - 1) * m_SubsampleFactor[1];
 
  // First, apply offset
  for (unsigned int dim = 0; dim < InputImageType::ImageDimension; ++dim)
  {
    inputIndex[dim] = std::min(outputIndex[dim], outputIndex[dim] + m_Offset[dim]);
    inputSize[dim]  = std::max(outputIndex[dim] + outputSize[dim], outputIndex[dim] + outputSize[dim] + m_Offset[dim]) - inputIndex[dim];
  }
 
  // Build the input requested region
  InputRegionType inputRequestedRegion;
  inputRequestedRegion.SetIndex(inputIndex);
  inputRequestedRegion.SetSize(inputSize);
 
  // Apply the radius
  inputRequestedRegion.PadByRadius(m_Radius);
 
  // Try to apply the requested region to the input image
  if (inputRequestedRegion.Crop(inputPtr->GetLargestPossibleRegion()))
  {
    inputPtr->SetRequestedRegion(inputRequestedRegion);
  }
  else
  {
    // Build an exception
    itk::InvalidRequestedRegionError e(__FILE__, __LINE__);
    e.SetLocation(ITK_LOCATION);
    e.SetDescription("Requested region is (at least partially) outside the largest possible region.");
    e.SetDataObject(inputPtr);
    throw e;
  }
}
 
template <class TInputImage, class TOutputImage>
void ScalarImageToTexturesFilter<TInputImage, TOutputImage>::BeforeThreadedGenerateData()
{
  unsigned int minRadius = 0;
  for (unsigned int i = 0; i < m_Offset.GetOffsetDimension(); i++)
  {
    unsigned int distance = std::abs(m_Offset[i]);
    if (distance > minRadius)
    {
      minRadius = distance;
    }
  }
  m_NeighborhoodRadius.Fill(minRadius);
}
 
template <class TInputImage, class TOutputImage>
void ScalarImageToTexturesFilter<TInputImage, TOutputImage>::ThreadedGenerateData(const OutputRegionType& outputRegionForThread, itk::ThreadIdType threadId)
{
  // Retrieve the input and output pointers
  InputImagePointerType  inputPtr             = const_cast<InputImageType*>(this->GetInput());
  OutputImagePointerType energyPtr            = this->GetEnergyOutput();
  OutputImagePointerType entropyPtr           = this->GetEntropyOutput();
  OutputImagePointerType correlationPtr       = this->GetCorrelationOutput();
  OutputImagePointerType invDiffMomentPtr     = this->GetInverseDifferenceMomentOutput();
  OutputImagePointerType inertiaPtr           = this->GetInertiaOutput();
  OutputImagePointerType clusterShadePtr      = this->GetClusterShadeOutput();
  OutputImagePointerType clusterProminencePtr = this->GetClusterProminenceOutput();
  OutputImagePointerType haralickCorPtr       = this->GetHaralickCorrelationOutput();
 
  // Build output iterators
  itk::ImageRegionIteratorWithIndex<OutputImageType> energyIt(energyPtr, outputRegionForThread);
  itk::ImageRegionIterator<OutputImageType>          entropyIt(entropyPtr, outputRegionForThread);
  itk::ImageRegionIterator<OutputImageType>          correlationIt(correlationPtr, outputRegionForThread);
  itk::ImageRegionIterator<OutputImageType>          invDiffMomentIt(invDiffMomentPtr, outputRegionForThread);
  itk::ImageRegionIterator<OutputImageType>          inertiaIt(inertiaPtr, outputRegionForThread);
  itk::ImageRegionIterator<OutputImageType>          clusterShadeIt(clusterShadePtr, outputRegionForThread);
  itk::ImageRegionIterator<OutputImageType>          clusterProminenceIt(clusterProminencePtr, outputRegionForThread);
  itk::ImageRegionIterator<OutputImageType>          haralickCorIt(haralickCorPtr, outputRegionForThread);
 
  // Go to begin
  energyIt.GoToBegin();
  entropyIt.GoToBegin();
  correlationIt.GoToBegin();
  invDiffMomentIt.GoToBegin();
  inertiaIt.GoToBegin();
  clusterShadeIt.GoToBegin();
  clusterProminenceIt.GoToBegin();
  haralickCorIt.GoToBegin();
 
  const double log2 = std::log(2.0);
 
  InputRegionType inputLargest = inputPtr->GetLargestPossibleRegion();
 
  // Set-up progress reporting
  itk::ProgressReporter progress(this, threadId, outputRegionForThread.GetNumberOfPixels());
 
  // Iterate on outputs to compute textures
  while (!energyIt.IsAtEnd() && !entropyIt.IsAtEnd() && !correlationIt.IsAtEnd() && !invDiffMomentIt.IsAtEnd() && !inertiaIt.IsAtEnd() &&
         !clusterShadeIt.IsAtEnd() && !clusterProminenceIt.IsAtEnd() && !haralickCorIt.IsAtEnd())
  {
    // Compute the region on which co-occurence will be estimated
    typename InputRegionType::IndexType inputIndex;
    typename InputRegionType::SizeType  inputSize;
 
    // Convert index to full grid
    typename OutputImageType::IndexType outIndex;
 
    // First, create an window for neighborhood iterator based on m_Radius
    // For example, if xradius and yradius is 2. window size is 5x5 (2 * radius + 1).
    for (unsigned int dim = 0; dim < InputImageType::ImageDimension; ++dim)
    {
      outIndex[dim]   = energyIt.GetIndex()[dim] * m_SubsampleFactor[dim] + m_SubsampleOffset[dim] + inputLargest.GetIndex(dim);
      inputIndex[dim] = outIndex[dim] - m_Radius[dim];
      inputSize[dim]  = 2 * m_Radius[dim] + 1;
    }
 
    // Build the input  region
    InputRegionType inputRegion;
    inputRegion.SetIndex(inputIndex);
    inputRegion.SetSize(inputSize);
    inputRegion.Crop(inputPtr->GetRequestedRegion());
 
    CooccurrenceIndexedListPointerType GLCIList = CooccurrenceIndexedListType::New();
    GLCIList->Initialize(m_NumberOfBinsPerAxis, m_InputImageMinimum, m_InputImageMaximum);
 
    typedef itk::ConstNeighborhoodIterator<InputImageType> NeighborhoodIteratorType;
    NeighborhoodIteratorType                               neighborIt;
    neighborIt = NeighborhoodIteratorType(m_NeighborhoodRadius, inputPtr, inputRegion);
    for (neighborIt.GoToBegin(); !neighborIt.IsAtEnd(); ++neighborIt)
    {
      const InputPixelType centerPixelIntensity = neighborIt.GetCenterPixel();
      bool                 pixelInBounds;
      const InputPixelType pixelIntensity = neighborIt.GetPixel(m_Offset, pixelInBounds);
      if (!pixelInBounds)
      {
        continue; // don't put a pixel in the co-occurrence list if the value is
                  // out of bounds
      }
      GLCIList->AddPixelPair(centerPixelIntensity, pixelIntensity);
    }
 
    double pixelMean = 0.;
    double marginalMean;
    double marginalDevSquared = 0.;
    double pixelVariance      = 0.;
 
    // Create and Initialize marginalSums
    std::vector<double> marginalSums(m_NumberOfBinsPerAxis, 0);
 
    // get co-occurrence vector and totalfrequency
    VectorType glcVector      = GLCIList->GetVector();
    double     totalFrequency = static_cast<double>(GLCIList->GetTotalFrequency());
 
    // Normalize the co-occurrence indexed list and compute mean, marginalSum
    typename VectorType::iterator it = glcVector.begin();
    while (it != glcVector.end())
    {
      double                frequency = (*it).second / totalFrequency;
      CooccurrenceIndexType index     = (*it).first;
      pixelMean += index[0] * frequency;
      marginalSums[index[0]] += frequency;
      ++it;
    }
 
    /* Now get the mean and deviaton of the marginal sums.
       Compute incremental mean and SD, a la Knuth, "The  Art of Computer
       Programming, Volume 2: Seminumerical Algorithms",  section 4.2.2.
       Compute mean and standard deviation using the recurrence relation:
       M(1) = x(1), M(k) = M(k-1) + (x(k) - M(k-1) ) / k
       S(1) = 0, S(k) = S(k-1) + (x(k) - M(k-1)) * (x(k) - M(k))
       for 2 <= k <= n, then
       sigma = std::sqrt(S(n) / n) (or divide by n-1 for sample SD instead of
       population SD).
     */
    std::vector<double>::const_iterator msIt = marginalSums.begin();
    marginalMean                             = *msIt;
    // Increment iterator to start with index 1
    ++msIt;
    for (int k = 2; msIt != marginalSums.end(); ++k, ++msIt)
    {
      double M_k_minus_1 = marginalMean;
      double S_k_minus_1 = marginalDevSquared;
      double x_k         = *msIt;
      double M_k         = M_k_minus_1 + (x_k - M_k_minus_1) / k;
      double S_k         = S_k_minus_1 + (x_k - M_k_minus_1) * (x_k - M_k);
      marginalMean       = M_k;
      marginalDevSquared = S_k;
    }
    marginalDevSquared = marginalDevSquared / m_NumberOfBinsPerAxis;
 
    VectorConstIteratorType constVectorIt;
    constVectorIt = glcVector.begin();
    while (constVectorIt != glcVector.end())
    {
      RelativeFrequencyType frequency = (*constVectorIt).second / totalFrequency;
      CooccurrenceIndexType index     = (*constVectorIt).first;
      pixelVariance += (index[0] - pixelMean) * (index[0] - pixelMean) * frequency;
      ++constVectorIt;
    }
 
    double pixelVarianceSquared = pixelVariance * pixelVariance;
    // Variance is only used in correlation. If variance is 0, then (index[0] - pixelMean) * (index[1] - pixelMean)
    // should be zero as well. In this case, set the variance to 1. in order to
    // avoid NaN correlation.
    if (pixelVarianceSquared < GetPixelValueTolerance())
    {
      pixelVarianceSquared = 1.;
    }
 
    // Initialize texture variables;
    PixelValueType energy                  = itk::NumericTraits<PixelValueType>::Zero;
    PixelValueType entropy                 = itk::NumericTraits<PixelValueType>::Zero;
    PixelValueType correlation             = itk::NumericTraits<PixelValueType>::Zero;
    PixelValueType inverseDifferenceMoment = itk::NumericTraits<PixelValueType>::Zero;
    PixelValueType inertia                 = itk::NumericTraits<PixelValueType>::Zero;
    PixelValueType clusterShade            = itk::NumericTraits<PixelValueType>::Zero;
    PixelValueType clusterProminence       = itk::NumericTraits<PixelValueType>::Zero;
    PixelValueType haralickCorrelation     = itk::NumericTraits<PixelValueType>::Zero;
 
    // Compute textures
    constVectorIt = glcVector.begin();
    while (constVectorIt != glcVector.end())
    {
      CooccurrenceIndexType index     = (*constVectorIt).first;
      RelativeFrequencyType frequency = (*constVectorIt).second / totalFrequency;
      energy += frequency * frequency;
      entropy -= (frequency > GetPixelValueTolerance()) ? frequency * std::log(frequency) / log2 : 0;
      correlation += ((index[0] - pixelMean) * (index[1] - pixelMean) * frequency) / pixelVarianceSquared;
      inverseDifferenceMoment += frequency / (1.0 + (index[0] - index[1]) * (index[0] - index[1]));
      inertia += (index[0] - index[1]) * (index[0] - index[1]) * frequency;
      clusterShade += std::pow((index[0] - pixelMean) + (index[1] - pixelMean), 3) * frequency;
      clusterProminence += std::pow((index[0] - pixelMean) + (index[1] - pixelMean), 4) * frequency;
      haralickCorrelation += index[0] * index[1] * frequency;
      ++constVectorIt;
    }
 
    haralickCorrelation = (fabs(marginalDevSquared) > 1E-8) ? (haralickCorrelation - marginalMean * marginalMean) / marginalDevSquared : 0;
 
    // Fill outputs
    energyIt.Set(energy);
    entropyIt.Set(entropy);
    correlationIt.Set(correlation);
    invDiffMomentIt.Set(inverseDifferenceMoment);
    inertiaIt.Set(inertia);
    clusterShadeIt.Set(clusterShade);
    clusterProminenceIt.Set(clusterProminence);
    haralickCorIt.Set(haralickCorrelation);
 
    // Update progress
    progress.CompletedPixel();
 
    // Increment iterators
    ++energyIt;
    ++entropyIt;
    ++correlationIt;
    ++invDiffMomentIt;
    ++inertiaIt;
    ++clusterShadeIt;
    ++clusterProminenceIt;
    ++haralickCorIt;
  }
}
 
} // End namespace otb
 
#endif