Convolves image with Gaussian PSF. More...

#include <GaussConvolve.h>

Public Types
enum	FilterType { FFT_BASED, RECURSIVE_BASED }
	identifier which type of gausssian filter this is

enum	DerivativeType { DERIVATIVE_NONE, DERIVATIVE_X, DERIVATIVE_Y, DERIVATIVE_XX, DERIVATIVE_YY, DERIVATIVE_XY }
	identifier which type of gausssian filter this is

Public Member Functions
	GaussConvolve ()
	constructor

	~GaussConvolve ()
	destructor

Static Public Member Functions
static ArrayGrid< double > *	CreateBlurredBars (int width, int height, int barWidth, double sigmaBlurMax)
	generate test pattern of horizontal bars that become gradually more and more blurred More...

static ArrayGrid< double > *	Convolve (ArrayGrid< double > *pImage, double sigma, FilterType myType=FFT_BASED)
	Applies Gaussian filter to grid Generates new grid, which the caller is responsible for. More...

static Image *	Convolve (Image *pImage, double sigma, FilterType myType=FFT_BASED)
	Applies Gaussian filter to image Generates new image, which the caller is responsible for. More...

static ArrayGrid< double > *	DerivativeConvolveFFT (ArrayGrid< double > *pGridIn, double sigmaX, double sigmaY, DerivativeType myType)
	Applies Gaussian filter or its derivatives (applied in Fourier domain) to a grid Generates new grid with filter response, which the caller is responsible for. More...

static ArrayGrid< double > *	DerivativeConvolveSeparable (ArrayGrid< double > *pGridIn, double sigma, DerivativeType myType, GridExtender< double >::ExtensionType myExtensionType=GridExtender< double >::EXTEND_MIRROR)
	Applies Gaussian filter or its derivatives (applied in spatial domain by a separable filter in x and y) to grid Generates new grid with filter response, which the caller is responsible for. More...

static ArrayGrid< double > *	ConvolveSeparable (ArrayGrid< double > *pGridIn, double sigma)
	Simple interface to straightforward Gaussian filter (applied in spatial domain by a separable filter in x and y) to grid Generates new grid with filter response, which the caller is responsible for. More...

static Image *	DerivativeConvolveFFT (Image *pImageIn, double sigma, DerivativeType myType, GridExtender< double >::ExtensionType myExtensionType=GridExtender< double >::EXTEND_MIRROR)
	Applies Gaussian filter or its derivatives (applied in Fourier domain) to image Generates new image with filter response, which the caller is responsible for. More...

static Image *	DerivativeConvolveSeparable (Image *pImageIn, double sigma, DerivativeType myType, GridExtender< double >::ExtensionType myExtensionType=GridExtender< double >::EXTEND_MIRROR)
	Applies Gaussian filter or its derivatives (applied in spatial domain by a separable filter in x and y) to image Generates new image with filter response, which the caller is responsible for. More...

static Image *	ConvolveSeparable (Image *pImageIn, double sigma)
	Simple interface to straightforward Gaussian filter (applied in spatial domain by a separable filter in x and y) to grid Generates new image with filter response, which the caller is responsible for. More...

static ArrayGrid< double > *	GenerateGaussianPSF (int width, int height, double sigmaX, double sigmaY)
	Generates a Gaussian filter kernel Generates new image, which the caller is responsible for. More...

static ArrayGrid< double > *	GenerateGaussianPSF (int width, int height, double sigma1, double sigma2, double rho)
	Generates a Gaussian filter kernel Generates new image, which the caller is responsible for. More...

static ArrayGrid< double > *	UpsampleGaussianInterpolated (ArrayGrid< double > *pGridIn, int upscalefactor)
	Generates an upscaled grid, where all pixels are interpolated with a Gaussian function Generates new grid with filter response, which the caller is responsible for. More...

Detailed Description

Convolves image with Gaussian PSF.

Recursive implementation based on:

article{ young95recursive, author = {Young, I. T. and {van~Vliet}, L. J.}, title = {Recursive implementation of the Gaussian filter}, volume = {44}, number = {2}, pages = {139-151}, year = {1995}, url = {http://citeseer.ist.psu.edu/young95recursive.html} }

Member Function Documentation

ArrayGrid< double > * stira::filter::GaussConvolve::Convolve	(	ArrayGrid< double > *	pImage,
		double	sigma,
		FilterType	myType = `FFT_BASED`
	)

static

Applies Gaussian filter to grid Generates new grid, which the caller is responsible for.

Parameters

pImage	input grid
sigma	sigma for the gaussian filter
myType	how to compute the filtering: recursively or in Fourier domain

Referenced by stira::degradationestimation::BlurEstimatorHu::Run(), and stira::contrastenhance::Retinex::~Retinex().

Image * stira::filter::GaussConvolve::Convolve	(	Image *	pImage,
		double	sigma,
		FilterType	myType = `FFT_BASED`
	)

static

Applies Gaussian filter to image Generates new image, which the caller is responsible for.

Parameters

pImage	input image
sigma	sigma for the gaussian filter
myType	how to compute the filtering: recursively or in Fourier domain

ArrayGrid< double > * stira::filter::GaussConvolve::ConvolveSeparable	(	ArrayGrid< double > *	pGridIn,
		double	sigma
	)

static

Simple interface to straightforward Gaussian filter (applied in spatial domain by a separable filter in x and y) to grid Generates new grid with filter response, which the caller is responsible for.

Parameters

pGridIn	input grid
sigma	sigma for the gaussian filter

Referenced by stira::deconvolve::WienerDeconvolve::~WienerDeconvolve().

Image * stira::filter::GaussConvolve::ConvolveSeparable	(	Image *	pImageIn,
		double	sigma
	)

static

Simple interface to straightforward Gaussian filter (applied in spatial domain by a separable filter in x and y) to grid Generates new image with filter response, which the caller is responsible for.

Parameters

pImageIn	input image
sigma	sigma for the gaussian filter

ArrayGrid< double > * stira::filter::GaussConvolve::CreateBlurredBars	(	int	width,
		int	height,
		int	barWidth,
		double	sigmaBlurMax
	)

static

generate test pattern of horizontal bars that become gradually more and more blurred

Parameters

width	width of the test pattern
height	height of the test pattern
barWidth	width of a bar in the image
sigmaBlurMax	the maximal blur applied to the bars (at the right of the image)

References stira::filter::SeparableFilter::BasicFilter1D(), stira::common::MathUtils::Gaussian(), stira::common::MathUtils::GaussianFirstDerivativeX1D(), stira::common::MathUtils::GaussianSecondDerivateX1D(), stira::imagetools::GridGenerator::GenerateBars(), stira::imagedata::ArrayGrid< T >::GetValue(), and stira::imagedata::ArrayGrid< T >::SetValue().

ArrayGrid< double > * stira::filter::GaussConvolve::DerivativeConvolveFFT	(	ArrayGrid< double > *	pGridIn,
		double	sigmaX,
		double	sigmaY,
		DerivativeType	myType
	)

static

Applies Gaussian filter or its derivatives (applied in Fourier domain) to a grid Generates new grid with filter response, which the caller is responsible for.

Parameters

pGridIn	input grid
sigma	sigma for the gaussian filter
myType	which derivative kernel to use for the filtering

Referenced by stira::imageanalysis::StegerLineDetector::Run(), and stira::filter::UnsharpMask::~UnsharpMask().

Image * stira::filter::GaussConvolve::DerivativeConvolveFFT	(	Image *	pImageIn,
		double	sigma,
		DerivativeType	myType,
		GridExtender< double >::ExtensionType	myExtensionType = `GridExtender<double>::EXTEND_MIRROR`
	)

static

Applies Gaussian filter or its derivatives (applied in Fourier domain) to image Generates new image with filter response, which the caller is responsible for.

Parameters

pImageIn	input grid
sigma	sigma for the gaussian filter
myType	which derivative kernel to use for the filtering
myExtensionType	identifies how borders should be extended (mirror, tiled periodically, zero's)

References stira::imagedata::Image::AddBand(), stira::imagetools::ImageTools::CropBorder(), stira::imagedata::Image::GetBands(), stira::imagedata::Image::GetHeight(), stira::imagedata::Image::GetImageName(), stira::imagedata::Image::GetNumberOfBands(), stira::imagedata::Image::GetWidth(), stira::imagetools::ImageTools::MirrorBorder(), stira::imagetools::ImageTools::PaddBorder(), and stira::imagedata::Image::SetImageName().

ArrayGrid< double > * stira::filter::GaussConvolve::DerivativeConvolveSeparable	(	ArrayGrid< double > *	pGridIn,
		double	sigma,
		DerivativeType	myType,
		GridExtender< double >::ExtensionType	myExtensionType = `GridExtender<double>::EXTEND_MIRROR`
	)

static

Applies Gaussian filter or its derivatives (applied in spatial domain by a separable filter in x and y) to grid Generates new grid with filter response, which the caller is responsible for.

Parameters

pGridIn	input grid
sigma	sigma for the gaussian filter
myType	which derivative kernel to use for the filtering
myExtensionType	identifies how borders should be extended (mirror, tiled periodically, zero's)

References stira::filter::SeparableFilter::RunRowColumn().

Image * stira::filter::GaussConvolve::DerivativeConvolveSeparable	(	Image *	pImageIn,
		double	sigma,
		DerivativeType	myType,
		GridExtender< double >::ExtensionType	myExtensionType = `GridExtender<double>::EXTEND_MIRROR`
	)

static

Applies Gaussian filter or its derivatives (applied in spatial domain by a separable filter in x and y) to image Generates new image with filter response, which the caller is responsible for.

Parameters

pImageIn	input image
sigma	sigma for the gaussian filter
myType	which derivative kernel to use for the filtering
myExtensionType	identifies how borders should be extended (mirror, tile, black)

References stira::imagedata::Image::AddBand(), stira::imagedata::Image::GetBands(), stira::imagedata::Image::GetHeight(), stira::imagedata::Image::GetImageName(), stira::imagedata::Image::GetNumberOfBands(), stira::imagedata::Image::GetWidth(), and stira::imagedata::Image::SetImageName().

static ArrayGrid<double>* stira::filter::GaussConvolve::GenerateGaussianPSF	(	int	width,
		int	height,
		double	sigmaX,
		double	sigmaY
	)

static

Generates a Gaussian filter kernel Generates new image, which the caller is responsible for.

Parameters

width	width of image to generate
height	height of image to generate
sigmaX	sigma in X direction for the gaussian filter
sigmaY	sigma in Y direction for the gaussian filter

static ArrayGrid<double>* stira::filter::GaussConvolve::GenerateGaussianPSF	(	int	width,
		int	height,
		double	sigma1,
		double	sigma2,
		double	rho
	)

static

Generates a Gaussian filter kernel Generates new image, which the caller is responsible for.

Parameters

width	width of image to generate
height	height of image to generate
sigma1	sigma of x for the gaussian filter
sigma2	sigma of y for the gaussian filter
rho	correlation coefficient, must be between -1 and 1

ArrayGrid< double > * stira::filter::GaussConvolve::UpsampleGaussianInterpolated	(	ArrayGrid< double > *	pGridIn,
		int	upscalefactor
	)

static

Generates an upscaled grid, where all pixels are interpolated with a Gaussian function Generates new grid with filter response, which the caller is responsible for.

Parameters

pGridIn	input grid
upscalefactor	factor with which grid needs to be upscaled

Referenced by stira::pyramidapplications::PyramidKeyPointDetector::~PyramidKeyPointDetector().

The documentation for this class was generated from the following files:

filter/filter/GaussConvolve.h
filter/filter/GaussConvolve.cpp

Public Types

Public Member Functions

Static Public Member Functions

Detailed Description

Member Function Documentation