lti::featureSaliencyIK Class Reference

Feature Saliency for color images. More...

#include <ltiFeatureSaliencyIK.h>

Inheritance diagram for lti::featureSaliencyIK:

[legend]

Collaboration diagram for lti::featureSaliencyIK:

[legend]

List of all members.

Classes
class	parameters
	the parameters for the class featureSaliencyIK More...
Public Member Functions
	featureSaliencyIK ()
	featureSaliencyIK (const featureSaliencyIK &other)
virtual	~featureSaliencyIK ()
virtual const char *	getTypeName () const
bool	apply (const image &src, channel &dest) const
bool	apply (const image &src, channel &conspIntensity, channel &conspColor, channel &conspOrientation) const
featureSaliencyIK &	copy (const featureSaliencyIK &other)
virtual functor *	clone () const
const parameters &	getParameters () const
Protected Member Functions
void	getColorChannels (const image &img, channel &RG, channel &BY, channel &I) const
int	getNumberPyramids () const
int	getIndex (const int &c, const int &s) const
void	getLimits (const int &idx, int &from, int &to) const
void	getUpscaledPyramids (const gaussianPyramid< channel > &src, std::vector< gaussianPyramid< channel > > &usp) const
void	subtAndNorm (const channel &a, const channel &b, channel &result) const
void	addAndNorm (const channel &a, const channel &b, channel &result) const
void	normAndAcc (const channel &c, channel &acc) const
void	normalize (channel &chnl) const
void	conspicuityI (const gaussianPyramid< channel > &pyr, channel &mapI) const
void	conspicuityC (const gaussianPyramid< channel > &pRG, const gaussianPyramid< channel > &pBY, channel &mapI) const
void	conspicuityO (const gaussianPyramid< channel > &p00, const gaussianPyramid< channel > &p45, const gaussianPyramid< channel > &p90, const gaussianPyramid< channel > &p135, channel &mapI) const
void	add (const gaussianPyramid< channel > &pyr, const int &pyrIndex, const int &scale, channel &result) const
void	itNorm (const int &steps, channel &chnl) const

Detailed Description

Feature Saliency for color images.

This class allows the generation of a Saliency Map from a color image based on the description in L. Itti et.at. "A Model of Saliency- Based Visual Attention for Rapid Scene Analysis", IEEE Trans. on Pattern Analysis and Machine Intelligence, Vol. 20, No. 11, November 1998

Constructor & Destructor Documentation

lti::featureSaliencyIK::featureSaliencyIK ( )

default constructor

lti::featureSaliencyIK::featureSaliencyIK ( const featureSaliencyIK & other )

copy constructor

Parameters:

other

the object to be copied

virtual lti::featureSaliencyIK::~featureSaliencyIK ( ) [virtual]

destructor

Member Function Documentation

void lti::featureSaliencyIK::add	(	const gaussianPyramid< channel > &	pyr,
		const int &	pyrIndex,
		const int &	scale,
		channel &	result
	)			const `[protected]`

add the scales in the given pyramid at the given scale, and leave the result in the given channel.

The size of the channel must be properly set before calling this function!

void lti::featureSaliencyIK::addAndNorm	(	const channel &	a,
		const channel &	b,
		channel &	result
	)			const `[protected]`

modified addition (+) of two inter-resolutional channels.

.. This does an element-wise subtraction and absolute value: result = | a + b | The member assumes "connected" channels (not lined!)

bool lti::featureSaliencyIK::apply	(	const image &	src,
		channel &	conspIntensity,
		channel &	conspColor,
		channel &	conspOrientation
	)			const

extract the conspicuity maps of intensity, color and orientation from the given color image

Parameters:

	src	image with the source data.
	conspIntensity	conspicuity map for intensity
	conspColor	conspicuity map for color
	conspOrientation	conspicuity map for orientation

Returns:: true if successful, false otherwise.

bool lti::featureSaliencyIK::apply	(	const image &	src,
		channel &	dest
	)			const

extract the saliency map of the image and leave it on dest

Parameters:

	src	image with the source data.
	dest	saliency map will be left.

Returns:: true if successful, false otherwise.

virtual functor* lti::featureSaliencyIK::clone ( ) const [virtual]

returns a pointer to a clone of this functor.

Implements lti::functor.

void lti::featureSaliencyIK::conspicuityC	(	const gaussianPyramid< channel > &	pRG,
		const gaussianPyramid< channel > &	pBY,
		channel &	mapI
	)			const `[protected]`

opponent colors conspicuity C(c,s)=N(RG(c,s))+N(BY(c,s)) the given pyramid must contain the subtraction of two color channels, i.e.

R-G or B-Y

void lti::featureSaliencyIK::conspicuityI	(	const gaussianPyramid< channel > &	pyr,
		channel &	mapI
	)			const `[protected]`

contrast channel is define as I(c,s)=|I(c) (-) I(s)|

void lti::featureSaliencyIK::conspicuityO	(	const gaussianPyramid< channel > &	p00,
		const gaussianPyramid< channel > &	p45,
		const gaussianPyramid< channel > &	p90,
		const gaussianPyramid< channel > &	p135,
		channel &	mapI
	)			const `[protected]`

orientation conspicuity.

The gaussian pyramids must contain gabor pyramids for the angles 0, 45, 90 and 135. (Note that gabor pyramids are a sort gaussian pyramids for the upsampling reconstruction)

featureSaliencyIK& lti::featureSaliencyIK::copy ( const featureSaliencyIK & other )

copy data of "other" functor.

Parameters:

other

the functor to be copied

Returns:: a reference to this functor object

Reimplemented from lti::functor.

void lti::featureSaliencyIK::getColorChannels	(	const image &	img,
		channel &	RG,
		channel &	BY,
		channel &	I
	)			const `[protected]`

calculate the after Itti & Koch required color channels:

With r', g' and b' the original color channels I = (r'+g'+b')/3 r = r'/I g = g'/I b = b'/I

the red, green, blue and yellow broadly tuned color channels are defined as:

R = r-(g+b)/2; // negative values are set to zero! G = g-(r+b)/2; // negative values are set to zero! B = b-(r+g)/2; // negative values are set to zero! Y = (r+g)/2-|r-g|/2-b; // negative values are set to zero!

int lti::featureSaliencyIK::getIndex	(	const int &	c,
		const int &	s
	)			const `[protected]`

return the index of the pyramid that contains the scaling of s to c

void lti::featureSaliencyIK::getLimits	(	const int &	idx,
		int &	from,
		int &	to
	)			const `[protected]`

return the used scales of the pyramid with index idx

int lti::featureSaliencyIK::getNumberPyramids ( ) const [protected]

returns the number of pyramids required to represent all reconstructed scales needed by the conspicuity maps

const parameters& lti::featureSaliencyIK::getParameters ( ) const

returns used parameters

Reimplemented from lti::modifier.

virtual const char* lti::featureSaliencyIK::getTypeName ( ) const [virtual]

returns the name of this type ("featureSaliencyIK")

Reimplemented from lti::modifier.

void lti::featureSaliencyIK::getUpscaledPyramids	(	const gaussianPyramid< channel > &	src,
		std::vector< gaussianPyramid< channel > > &	usp
	)			const `[protected]`

generate the upsampling pyramids required for the given pyramid

void lti::featureSaliencyIK::itNorm	(	const int &	steps,
		channel &	chnl
	)			const `[protected]`

iterative normalization

void lti::featureSaliencyIK::normalize ( channel & chnl ) const [protected]

normalize

void lti::featureSaliencyIK::normAndAcc	(	const channel &	c,
		channel &	acc
	)			const `[protected]`

accumulate the normed channel c in the accumulator-channel acc

void lti::featureSaliencyIK::subtAndNorm	(	const channel &	a,
		const channel &	b,
		channel &	result
	)			const `[protected]`

modified subtration (-) of two inter-resolutional channels.

.. This does an element-wise subtraction and absolute value: result = | a - b | The member assumes "connected" channels (not lined!)

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

ltiFeatureSaliencyIK.h

lti::featureSaliencyIK Class Reference

Classes

Public Member Functions

Protected Member Functions

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation