Data Structures
struct	tensor_t
	Structure handling anisotropic gradient parameters. More...
class	ImageTensors
	Class used to render tensors. More...
Functions
template<typename S_VIEW , typename D_VIEW >
void	deriche (const S_VIEW &src, D_VIEW &dst, const float sigma, const int order=0, const char axe= 'x', const bool cond=true)
	The Canny-Deriche filter is a recursive algorithm allowing to compute blurred derivatives of order 0,1 or 2 of an image.
template<typename S_VIEW , typename D_VIEW >
void	dericheFilter (const S_VIEW &src, D_VIEW &dst, const float sigma, int order=0, float threshold=0.0f)
	Use deriche filter amon x and y direction to apply blur.
template<typename S_VIEW , typename D_VIEW >
D_VIEW &	simple_structure_tensor (const S_VIEW &src, D_VIEW &final_dst, float threshold)
	Apply precise forward/backward finite differences to get a gradient image.
template<typename S_VIEW , typename D_VIEW >
void	harris (const S_VIEW &src, D_VIEW &dst, float threshold)
template<typename ViewIt >
void	symmetric_eigen (ViewIt &tensorXY, std::vector< double > &val, std::vector< double > &vec)
	Compute the eigenvalues and eigenvectors of a symmetric matrix.
const float &	pix2d (const float *W, const int x, const int y, const int c, const int w, const int h, const int nc)
	Return right pixel from unsafe coordinates.
template<typename DstView >
const float	pix2d (DstView &myView, const int x, const int y, const int c, const int w, const int h)
	Return right pixel from unsafe coordinates.
float	linear_pix2d (const float *W, const float fx, const float fy, const int c, const int w, const int h, const int nc)
	Return interpolated pixel.
template<typename DstView >
float	linear_pix2d (DstView &myView, const float fx, const float fy, const int c, const int w, const int h)
	Return interpolated pixel.
template<class S_VIEW , class D_VIEW >
D_VIEW &	pseudo_residual (S_VIEW &src, D_VIEW &dst)
	Compute somme pseudo-residuals The pseudo residual r_i of the image Y_i are so thar E[r_i^2] = E[Y_i^2]. This is the 2D pseudo-implementation.
template<class S_VIEW , typename t >
double	variance_mean (S_VIEW &src, t &mean)
	Return the variance and the mean of the image Least Mean of Square.
template<class S_VIEW >
double	noise_variance (S_VIEW &src)
	Robustly estimatate the variance of a the noise using the pseudo-residuals.

Function Documentation

template<typename S_VIEW , typename D_VIEW >

void tuttle::imageUtils::deriche	(	const S_VIEW &	src,
		D_VIEW &	dst,
		const float	sigma,
		const int	order = `0`,
		const char	axe = `'x'`,
		const bool	cond = `true`
	)

The Canny-Deriche filter is a recursive algorithm allowing to compute blurred derivatives of order 0,1 or 2 of an image.

Parameters:

[in,out]	src	Source view
[out]	dst	Destination view
[in]	sigma	Standard variation of the gaussian distribution
[in]	order	Derivative order (0 = kind of blur, 1 = kind of blurred edge detection, 2 = kind of very blurred edge detection)
[in]	axe	Direction of the blur
[in]	cond	Miscallenous

Returns:: Return the result of the Deriche filter

See also:: blur

Definition at line 37 of file blurFilters.hpp.

References tuttle::plugin::merge::copy_pixels().

Referenced by dericheFilter().

Here is the call graph for this function:

template<typename S_VIEW , typename D_VIEW >

void tuttle::imageUtils::dericheFilter	(	const S_VIEW &	src,
		D_VIEW &	dst,
		const float	sigma,
		int	order = `0`,
		float	threshold = `0.0f`
	)

Use deriche filter amon x and y direction to apply blur.

Parameters:

[in,out]	src	Source view
[out]	dst	Destination view
[in]	sigma	Standard variation of the gaussian distribution

Definition at line 226 of file blurFilters.hpp.

References deriche(), and boost::gil::view().

Referenced by tuttle::imageUtils::ImageTensors< View >::anisotropic_gradient().

Here is the call graph for this function:

template<typename S_VIEW , typename D_VIEW >

void tuttle::imageUtils::harris	(	const S_VIEW &	src,
		D_VIEW &	dst,
		float	threshold
	)

Definition at line 125 of file edgeDetect.hpp.

Referenced by tuttle::imageUtils::ImageTensors< View >::anisotropic_gradient().

float tuttle::imageUtils::linear_pix2d	(	const float *	W,
		const float	fx,
		const float	fy,
		const int	c,
		const int	w,
		const int	h,
		const int	nc
	)		`[inline]`

Return interpolated pixel.

Parameters:

[in]	W	input image (linear buffer, cache optimisation)
[in]	fx	float abscisse
[in]	fy	float ordinate
[in]	c	channel index
[in]	w	image width
[in]	h	image height
[in]	nc	channels number

Returns:: right pixel

Definition at line 67 of file interpolation.hpp.

References pix2d().

Here is the call graph for this function:

template<typename DstView >

float tuttle::imageUtils::linear_pix2d	(	DstView &	myView,
		const float	fx,
		const float	fy,
		const int	c,
		const int	w,
		const int	h
	)

Return interpolated pixel.

Parameters:

[in]	myView	input view
[in]	fx	float abscisse
[in]	fy	float ordinate
[in]	c	channel index
[in]	w	image width
[in]	h	image height
[in]	nc	channels number

Returns:: right pixel

Definition at line 97 of file interpolation.hpp.

References pix2d().

Here is the call graph for this function:

template<class S_VIEW >

double tuttle::imageUtils::noise_variance ( S_VIEW & src )

Robustly estimatate the variance of a the noise using the pseudo-residuals.

See also:: variance_estimation()

Definition at line 116 of file noiseAnalysis.hpp.

References pseudo_residual(), variance_mean(), and boost::gil::view().

Referenced by tuttle::plugin::nlmDenoiser::NLMDenoiserProcess< View >::computeWeights().

Here is the call graph for this function:

const float& tuttle::imageUtils::pix2d	(	const float *	W,
		const int	x,
		const int	y,
		const int	c,
		const int	w,
		const int	h,
		const int	nc
	)		`[inline]`

Return right pixel from unsafe coordinates.

Parameters:

[in]	W	input image (linear buffer, cache optimisation)
[in]	x	abscisse
[in]	y	ordinate
[in]	c	channel index
[in]	w	image width
[in]	h	image height
[in]	nc	channels number

Returns:: right pixel

Definition at line 28 of file interpolation.hpp.

Referenced by linear_pix2d().

template<typename DstView >

const float tuttle::imageUtils::pix2d	(	DstView &	myView,
		const int	x,
		const int	y,
		const int	c,
		const int	w,
		const int	h
	)

Return right pixel from unsafe coordinates.

Parameters:

[in]	myView	input view
[in]	x	abscisse
[in]	y	ordinate
[in]	c	channel index
[in]	w	image width
[in]	h	image height
[in]	nc	channels number

Returns:: right pixel

Definition at line 48 of file interpolation.hpp.

template<class S_VIEW , class D_VIEW >

D_VIEW& tuttle::imageUtils::pseudo_residual	(	S_VIEW &	src,
		D_VIEW &	dst
	)

Compute somme pseudo-residuals The pseudo residual r_i of the image Y_i are so thar E[r_i^2] = E[Y_i^2]. This is the 2D pseudo-implementation.

Definition at line 22 of file noiseAnalysis.hpp.

Referenced by noise_variance().

template<typename S_VIEW , typename D_VIEW >

D_VIEW& tuttle::imageUtils::simple_structure_tensor	(	const S_VIEW &	src,
		D_VIEW &	final_dst,
		float	threshold
	)

Apply precise forward/backward finite differences to get a gradient image.

Parameters:

[in]	src	Source view
[out]	final_dst	Destination view

Returns:: outline image (destination)

Warning:: Destination MUST BE different from source.

Definition at line 32 of file edgeDetect.hpp.

Referenced by tuttle::imageUtils::ImageTensors< View >::anisotropic_gradient().

template<typename ViewIt >

void tuttle::imageUtils::symmetric_eigen	(	ViewIt &	tensorXY,
		std::vector< double > &	val,
		std::vector< double > &	vec
	)

Compute the eigenvalues and eigenvectors of a symmetric matrix.

Parameters:

[in]	tensorXY	tensor (view iterator) at a certain position
[out]	val	eigenvalue
[out]	val	eigenvector

Warning:: vec MUST BE a linear float vector of dimension 4.

Definition at line 247 of file edgeDetect.hpp.

Referenced by tuttle::imageUtils::ImageTensors< View >::anisotropic_gradient().

template<class S_VIEW , typename t >

double tuttle::imageUtils::variance_mean	(	S_VIEW &	src,
		t &	mean
	)

Return the variance and the mean of the image Least Mean of Square.

Definition at line 78 of file noiseAnalysis.hpp.

Referenced by noise_variance().

Data Structures

Functions

Function Documentation