Download this example

Label Connect Components in Grayscale Image

Synopsis

Label connected components in a grayscale image.

Results

Output In VTK Window

Code

C++

#include "itkImage.h"
#include "itkImageFileReader.h"
#include "itkLabelToRGBImageFilter.h"
#include "itkRelabelComponentImageFilter.h"
#include "itkLabelStatisticsImageFilter.h"

#include "itkScalarConnectedComponentImageFilter.h"

#include "itksys/SystemTools.hxx"
#include <sstream>

#ifdef ENABLE_QUICKVIEW
#  include "QuickView.h"
#endif

template <typename TImage>
static void
CreateImage(TImage * const image);

template <typename TImage, typename TLabelImage>
static void
SummarizeLabelStatistics(TImage * image, TLabelImage * labelImage);

int
main(int argc, char * argv[])
{
  constexpr unsigned int Dimension = 2;
  using PixelType = short;
  using ImageType = itk::Image<PixelType, Dimension>;

  using RGBPixelType = itk::RGBPixel<unsigned char>;
  using RGBImageType = itk::Image<RGBPixelType, Dimension>;

  using LabelPixelType = unsigned int;
  using LabelImageType = itk::Image<LabelPixelType, Dimension>;

  ImageType::Pointer image;
  PixelType          distanceThreshold = 4;
  if (argc < 2)
  {
    image = ImageType::New();
    CreateImage(image.GetPointer());
  }
  else
  {
    if (argc > 2)
    {
      distanceThreshold = static_cast<PixelType>(atoi(argv[2]));
    }
    image = itk::ReadImage<ImageType>(argv[1]);
  }


  using ConnectedComponentImageFilterType = itk::ScalarConnectedComponentImageFilter<ImageType, LabelImageType>;

  auto connected = ConnectedComponentImageFilterType::New();
  connected->SetInput(image);
  connected->SetDistanceThreshold(distanceThreshold);

  using RelabelFilterType = itk::RelabelComponentImageFilter<LabelImageType, LabelImageType>;
  auto                              relabel = RelabelFilterType::New();
  RelabelFilterType::ObjectSizeType minSize = 20;
  if (argc > 3)
  {
    minSize = std::stoi(argv[3]);
  }
  relabel->SetInput(connected->GetOutput());
  relabel->SetMinimumObjectSize(minSize);
  relabel->Update();

  SummarizeLabelStatistics(image.GetPointer(), relabel->GetOutput());

  using RGBFilterType = itk::LabelToRGBImageFilter<LabelImageType, RGBImageType>;
  auto rgbFilter = RGBFilterType::New();
  rgbFilter->SetInput(relabel->GetOutput());

#ifdef ENABLE_QUICKVIEW
  QuickView viewer;
  viewer.AddImage(
    image.GetPointer(), true, argc > 1 ? itksys::SystemTools::GetFilenameName(argv[1]) : "Generated image");

  std::stringstream desc;
  desc << "Scalar Connected Components:\n# of Objects: " << relabel->GetNumberOfObjects() << " Threshold: "
       << itk::NumericTraits<ConnectedComponentImageFilterType::OutputPixelType>::PrintType(
            connected->GetDistanceThreshold())
       << " Min Size: " << relabel->GetMinimumObjectSize();
  viewer.AddRGBImage(rgbFilter->GetOutput(), true, desc.str());

  viewer.Visualize();
#endif

  return EXIT_SUCCESS;
}

template <typename TImage>
void
CreateImage(TImage * const image)
{
  // Create an image with 2 connected components
  typename TImage::IndexType start = { { 0, 0 } };
  start[0] = 0;
  start[1] = 0;

  typename TImage::SizeType size;
  unsigned int              NumRows = 200;
  unsigned int              NumCols = 300;
  size[0] = NumRows;
  size[1] = NumCols;

  typename TImage::RegionType region(start, size);

  image->SetRegions(region);
  image->Allocate();

  // Make a square
  for (typename TImage::IndexValueType r = 20; r < 80; ++r)
  {
    for (typename TImage::IndexValueType c = 30; c < 100; ++c)
    {
      typename TImage::IndexType pixelIndex = { { r, c } };

      image->SetPixel(pixelIndex, 255);
    }
  }

  // Make another square
  for (typename TImage::IndexValueType r = 100; r < 130; ++r)
  {
    for (typename TImage::IndexValueType c = 115; c < 160; ++c)
    {
      typename TImage::IndexType pixelIndex = { { r, c } };

      image->SetPixel(pixelIndex, 255);
    }
  }
}

template <typename TImage, typename TLabelImage>
void
SummarizeLabelStatistics(TImage * image, TLabelImage * labelImage)
{
  using LabelStatisticsImageFilterType = itk::LabelStatisticsImageFilter<TImage, TLabelImage>;
  auto labelStatisticsImageFilter = LabelStatisticsImageFilterType::New();
  labelStatisticsImageFilter->SetLabelInput(labelImage);
  labelStatisticsImageFilter->SetInput(image);
  labelStatisticsImageFilter->UseHistogramsOn(); // needed to compute median
  labelStatisticsImageFilter->Update();

  std::cout << "Number of labels: " << labelStatisticsImageFilter->GetNumberOfLabels() << std::endl;
  std::cout << std::endl;

  using LabelPixelType = typename LabelStatisticsImageFilterType::LabelPixelType;

  for (auto vIt = labelStatisticsImageFilter->GetValidLabelValues().begin();
       vIt != labelStatisticsImageFilter->GetValidLabelValues().end();
       ++vIt)
  {
    if (labelStatisticsImageFilter->HasLabel(*vIt))
    {
      LabelPixelType labelValue = *vIt;
      std::cout << "Label: " << *vIt << std::endl;
      std::cout << "\tmin: " << labelStatisticsImageFilter->GetMinimum(labelValue) << std::endl;
      std::cout << "\tmax: " << labelStatisticsImageFilter->GetMaximum(labelValue) << std::endl;
      std::cout << "\tmedian: " << labelStatisticsImageFilter->GetMedian(labelValue) << std::endl;
      std::cout << "\tmean: " << labelStatisticsImageFilter->GetMean(labelValue) << std::endl;
      std::cout << "\tsigma: " << labelStatisticsImageFilter->GetSigma(labelValue) << std::endl;
      std::cout << "\tvariance: " << labelStatisticsImageFilter->GetVariance(labelValue) << std::endl;
      std::cout << "\tsum: " << labelStatisticsImageFilter->GetSum(labelValue) << std::endl;
      std::cout << "\tcount: " << labelStatisticsImageFilter->GetCount(labelValue) << std::endl;
      std::cout << "\tregion: " << labelStatisticsImageFilter->GetRegion(labelValue) << std::endl;
    }
  }
}

Classes demonstrated

template<typename TInputImage, typename TOutputImage, typename TMaskImage = TInputImage>
class ScalarConnectedComponentImageFilter : public itk::ConnectedComponentFunctorImageFilter<TInputImage, TOutputImage, Functor::SimilarPixelsFunctor<TInputImage::ValueType>, TMaskImage>

A connected components filter that labels the objects in an arbitrary image. Two pixels are similar if they are within threshold of each other. Uses ConnectedComponentFunctorImageFilter.

ITK Sphinx Examples:

• All ITK Sphinx Examples

• Label Connect Components In Grayscale Image

See itk::ScalarConnectedComponentImageFilter for additional documentation.