Iterative Hole Filling#

Synopsis#

Fill hole or cavity in one itk::Image

Results#

Code#

C++#

#include "itkImageFileReader.h"
#include "itkImageFileWriter.h"
#include "itkVotingBinaryIterativeHoleFillingImageFilter.h"

int
main(int argc, char * argv[])
{
  if (argc != 6)
  {
    std::cerr << "Usage: " << std::endl;
    std::cerr << argv[0];
    std::cerr << " <InputFileName> <OutputFileName> <radius> <majority threshold> <number of iterations>";
    std::cerr << std::endl;
    return EXIT_FAILURE;
  }

  const char * inputFileName = argv[1];
  const char * outputFileName = argv[2];

  int          r = std::stoi(argv[3]);
  int          majorityThreshold = std::stoi(argv[4]);
  unsigned int numberOfIterations = std::stoi(argv[5]);

  constexpr unsigned int Dimension = 2;

  using PixelType = unsigned char;
  using ImageType = itk::Image<PixelType, Dimension>;

  const auto input = itk::ReadImage<ImageType>(inputFileName);

  using FilterType = itk::VotingBinaryIterativeHoleFillingImageFilter<ImageType>;
  FilterType::InputSizeType radius;
  radius.Fill(r);

  auto filter = FilterType::New();
  filter->SetInput(input);
  filter->SetRadius(radius);
  filter->SetMajorityThreshold(majorityThreshold);
  filter->SetBackgroundValue(itk::NumericTraits<PixelType>::Zero);
  filter->SetForegroundValue(itk::NumericTraits<PixelType>::max());
  filter->SetMaximumNumberOfIterations(numberOfIterations);

  try
  {
    itk::WriteImage(filter->GetOutput(), outputFileName);
  }
  catch (const itk::ExceptionObject & error)
  {
    std::cerr << "Error: " << error << std::endl;
    return EXIT_FAILURE;
  }

  return EXIT_SUCCESS;
}

Classes demonstrated#

template<typename TImage> class VotingBinaryIterativeHoleFillingImageFilter : public itk::ImageToImageFilter<TImage, TImage>

Fills in holes and cavities by iteratively applying a voting operation.

This filter uses internally the VotingBinaryHoleFillingImageFilter, and runs it iteratively until no pixels are being changed or until it reaches the maximum number of iterations. The purpose of the filter is to fill in holes of medium size (tens of pixels in radius). In principle the number of iterations is related to the size of the holes to be filled in. The larger the holes, the more iteration must be run with this filter in order to fill in the full hole. The size of the neighborhood is also related to the curvature of the hole borders and therefore the hole size. Note that as a collateral effect this filter may also fill in cavities in the external side of structures.

This filter is templated over a single image type because the output image type must be the same as the input image type. This is required in order to make the iterations possible, since the output image of one iteration is taken as the input image for the next iteration.

See: Image
See: VotingBinaryImageFilter
See: VotingBinaryHoleFillingImageFilter
See: Neighborhood
See: NeighborhoodOperator
See: NeighborhoodIterator

See itk::VotingBinaryIterativeHoleFillingImageFilter for additional documentation.