Convert Image to Another Type#

Synopsis#

Converts an image to another type of image.

Results#

Code#

C++#

#include "itkRGBPixel.h"
#include "itkRGBAPixel.h"
#include "itkImage.h"
#include "itkDefaultConvertPixelTraits.h"
#include "itkConvertPixelBuffer.h"
#include "itkImageRandomConstIteratorWithIndex.h"

int
main()
{
  constexpr int xDimension = 200;
  constexpr int yDimension = 100;

  using ComponentType = unsigned char;
  using RGBPixelType = itk::RGBPixel<ComponentType>;
  using RGBAPixelType = itk::RGBAPixel<ComponentType>;
  using RGBImageType = itk::Image<RGBPixelType, 2>;
  using RGBAImageType = itk::Image<RGBAPixelType, 2>;
  using TraitsType = itk::DefaultConvertPixelTraits<RGBPixelType>;
  using RGBAConverterType = itk::ConvertPixelBuffer<ComponentType, RGBPixelType, TraitsType>;

  auto rgbImg = RGBImageType::New();
  auto rgbaImg = RGBAImageType::New();

  // Create the two images
  // RGBAImage
  RGBAImageType::IndexType rgbaStart;
  rgbaStart[0] = 0;
  rgbaStart[1] = 0;

  RGBAImageType::SizeType rgbaSize;
  rgbaSize[0] = xDimension;
  rgbaSize[1] = yDimension;

  itk::ImageRegion<2> rgbaRegion(rgbaStart, rgbaSize);
  rgbaImg->SetRegions(rgbaRegion);
  rgbaImg->Allocate();

  RGBAPixelType rgbaDefault;
  rgbaDefault[0] = 127;
  rgbaDefault[1] = 100;
  rgbaDefault[2] = 230;
  rgbaDefault[3] = 255;
  rgbaImg->FillBuffer(rgbaDefault);

  // RGBImage
  itk::ImageRegion<2> rgbRegion = rgbaRegion;
  rgbImg->SetRegions(rgbRegion);
  rgbImg->Allocate();

  size_t numberOfPixels = rgbImg->GetLargestPossibleRegion().GetNumberOfPixels();

  // Convert a raw buffer to a buffer of pixel types
  RGBAConverterType::Convert(rgbaImg->GetBufferPointer()->GetDataPointer(),
                             rgbaImg->GetNumberOfComponentsPerPixel(),
                             rgbImg->GetBufferPointer(),
                             numberOfPixels);

  // Check a few random values
  itk::ImageRandomConstIteratorWithIndex<RGBImageType> rgbIterator(rgbImg, rgbImg->GetLargestPossibleRegion());
  rgbIterator.SetNumberOfSamples(numberOfPixels / 10);
  rgbIterator.GoToBegin();

  while (!rgbIterator.IsAtEnd())
  {
    if (rgbImg->GetPixel(rgbIterator.GetIndex())[0] != rgbaImg->GetPixel(rgbIterator.GetIndex())[0] ||

        rgbImg->GetPixel(rgbIterator.GetIndex())[1] != rgbaImg->GetPixel(rgbIterator.GetIndex())[1] ||

        rgbImg->GetPixel(rgbIterator.GetIndex())[2] != rgbaImg->GetPixel(rgbIterator.GetIndex())[2])
    {
      std::cout << "Copy failed for index " << rgbIterator.GetIndex() << " got "
                << rgbImg->GetPixel(rgbIterator.GetIndex()) << " but expected "
                << rgbaImg->GetPixel(rgbIterator.GetIndex()) << std::endl;
    }
    ++rgbIterator;
  }
  return EXIT_SUCCESS;
}

Classes demonstrated#

template<typename InputPixelType, typename OutputPixelType, typename OutputConvertTraits> class ConvertPixelBuffer

Class to convert blocks of data from one type to another.

ConvertPixelBuffer has a static method Convert(). It is used by itkImageFileReader to convert from an unknown type at run-time to the compile-time template pixel type in itkImageFileReader. To work with complex pixel types like RGB and RGBA a traits class is used. OutputConvertTraits() is the traits class. The default one used is DefaultConvertPixelTraits.

Derived from

ConvertPixelBuffer has a static method Convert(). It is used to work with pixel type as VariableLengthVector type.

ITK Sphinx Examples:

All ITK Sphinx Examples

Convert Image To Another Type

Derived from

ConvertPixelBuffer has a static method Convert(). It is used to work with pixel type as Array type.

See itk::ConvertPixelBuffer for additional documentation.