Functors to Transform Images

Classes
class	BrightnessContrastFunctor< PixelType >
	Adjust brightness and contrast of an image. More...
class	GammaFunctor< PixelType >
	Perform gamma correction of an image. More...
class	Threshold< SrcValueType, DestValueType >
	Threshold an image. More...
class	VectorNormFunctor< ValueType >
	A functor for computing the vector norm. More...
class	VectorNormSqFunctor< ValueType >
	A functor for computing the squared vector norm. More...

Functions
template<class Multiplier , class DestValueType >
LinearIntensityTransform < DestValueType, Multiplier >	linearIntensityTransform (Multiplier scale, DestValueType offset)
	Apply a linear transform to the source pixel values. More...
template<class SrcValueType , class DestValueType >
LinearIntensityTransform < DestValueType, typename NumericTraits< DestValueType > ::RealPromote >	linearRangeMapping (SrcValueType src_min, SrcValueType src_max, DestValueType dest_min, DestValueType dest_max)
	Map a source intensity range linearly to a destination range. More...

Detailed Description

Note that the unary functors of the STL can also be used in connection with transformImage().

Function Documentation

LinearIntensityTransform<DestValueType, Multiplier> vigra::linearIntensityTransform	(	Multiplier	scale,
		DestValueType	offset
	)

Apply a linear transform to the source pixel values.

Factory function for a functor that linearly transforms the source pixel values. The functor applies the transform 'destvalue = scale * (srcvalue + offset)' to every pixel. This can, for example, be used to transform images into the visible range 0...255 or to invert an image.

If you leave out the second parameter / offset, you will get an optimized version of the functor which only scales by the given factor, however you have to make the template parameter (pixel type) explicit then.

Traits defined:

FunctorTraits::isUnaryFunctor is true (VigraTrueType)

Declaration:

namespace vigra {
    template <class Multiplier, class DestValueType>
    LinearIntensityTransform<DestValueType, Multiplier>
    linearIntensityTransform(Multiplier scale, DestValueType offset);

    template <class DestValueType, class Multiplier>
    ScalarIntensityTransform<DestValueType, Multiplier>
    linearIntensityTransform(Multiplier scale);
}

Usage:

#include <vigra/transformimage.hxx>
Namespace: vigra

vigra::IImage src(width, height);
vigra::BImage dest(width, height);
...
vigra::FindMinMax<IImage::PixelType> minmax;   // functor to find range

vigra::inspectImage(srcImageRange(src), minmax); // find original range

// transform to range 0...255
vigra::transformImage(srcImageRange(src), destImage(dest),
                      linearIntensityTransform(
                        255.0 / (minmax.max - minmax.min), // scaling
                      - minmax.min));                    // offset

The one-parameter version can be used like this:

// scale from 0..255 to 0..1.0
FImage dest(src.size());

vigra::transformImage(srcImageRange(src), destImage(dest),
                      linearIntensityTransform<float>(1.0 / 255));

Required Interface:

The source and destination value types must be models of LinearSpace in both cases.

Examples:

invert.cxx.

LinearIntensityTransform<DestValueType, typename NumericTraits<DestValueType>::RealPromote> vigra::linearRangeMapping	(	SrcValueType	src_min,
		SrcValueType	src_max,
		DestValueType	dest_min,
		DestValueType	dest_max
	)

Map a source intensity range linearly to a destination range.

Factory function for a functor that linearly transforms the source pixel values. The functor applies the transform 'destvalue = scale * (srcvalue + offset)' to every pixel, where scale = (dest_max - dest_min) / (src_max - src_min) and offset = dest_min / scale - src_min. As a result, the pixel values src_max, src_min in the source image are mapped onto dest_max, dest_min respectively. This works for scalar as well as vector pixel types. Instead of src_min and src_max, you may also pass a functor FindMinMax.

Declaration:

namespace vigra {
    template <class SrcValueType, class DestValueType>
    LinearIntensityTransform<DestValueType, typename NumericTraits<DestValueType>::RealPromote>
    linearRangeMapping(SrcValueType src_min, SrcValueType src_max,
                       DestValueType dest_min, DestValueType dest_max );

                       template <class SrcValueType, class DestValueType>
    LinearIntensityTransform<DestValueType, typename NumericTraits<DestValueType>::RealPromote>
    linearRangeMapping(SrcValueType src_min, SrcValueType src_max,
                       DestValueType dest_min, DestValueType dest_max );
}

Usage:

#include <vigra/transformimage.hxx>
Namespace: vigra

vigra::IImage src(width, height);
vigra::BImage dest(width, height);
...
vigra::FindMinMax<IImage::PixelType> minmax;   // functor to find range

vigra::inspectImage(srcImageRange(src), minmax); // find original range

// transform to range 0...255
vigra::transformImage(srcImageRange(src), destImage(dest),
                      linearRangeMapping(
                        minmax.min, minmax.max,  // src range
                        0, 255)                  // dest range
                      );

// equivalent, but shorter
vigra::transformImage(srcImageRange(src), destImage(dest),
                      linearRangeMapping(
                        minmax,                 // src range
                        0, 255)                 // dest range
                      );

Required Interface:

The source and destination value types must be models of LinearSpace in both cases.