lti::colorModelSelector Class Reference

This functor selects one color model from a set, which elements were given by addColorModel. More...

#include <ltiColorModelSelector.h>

Inheritance diagram for lti::colorModelSelector:

[legend]

Collaboration diagram for lti::colorModelSelector:

[legend]

List of all members.

Classes
class	namedModel
	Internal data type to facilitate the administration of named color models. More...
class	parameters
	the parameters for the class colorModelSelector More...
Public Member Functions
	colorModelSelector ()
	colorModelSelector (const colorModelSelector &other)
virtual	~colorModelSelector ()
virtual const char *	getTypeName () const
bool	apply (const image &img, probabilityMap &colorMap, principalComponents< float > &pca)
bool	apply (const image &img, const probabilityMap &pColorMap, const principalComponents< float > &pPca)
bool	apply (const image &img, const rectangle &window, probabilityMap &colorMap, principalComponents< float > &pca)
bool	apply (const image &img, const rectangle &window, const probabilityMap &pColorMap, const principalComponents< float > &pPca)
bool	apply (const image &img, probabilityMap &colorMap) const
bool	apply (const image &img, const probabilityMap *&pColorMap) const
bool	apply (const image &img, const probabilityMap *&pColorMap, std::string &modelName) const
bool	apply (const image &img, const rectangle &window, const probabilityMap &pColorMap, const principalComponents< float > &pPca, std::string &modelName)
bool	apply (const image &img, const rectangle &window, probabilityMap &colorMap) const
bool	apply (const image &img, const rectangle &window, const probabilityMap *&pColorMap) const
bool	apply (const image &img, const rectangle &window, const probabilityMap *&pColorMap, std::string &modelName) const
bool	addColorModel (const std::string &name, const std::string &colorHistFileName, const std::string &nonColorHistFileName)
bool	addColorModel (const std::string &name, const thistogram< double > &colorHist, const thistogram< double > &nonColorHist)
bool	addColorModel (const std::string &name, const probabilityMap &colorMap)
bool	updateModels (const probabilityMap::parameters &newAttribs)
bool	initPCA ()
bool	clearColorModels ()
bool	removeColorModel (const std::string &name)
colorModelSelector &	copy (const colorModelSelector &other)
colorModelSelector &	operator= (const colorModelSelector &other)
bool	read (ioHandler &handler, const bool complete=true)
bool	write (ioHandler &handler, const bool complete=true) const
virtual functor *	clone () const
const parameters &	getParameters () const
Protected Attributes
std::list< namedModel >	modelList

Detailed Description

This functor selects one color model from a set, which elements were given by addColorModel.

The selected model (a lti::probabilityMap functor) will be the one which maximizes the likelihood in the given image window, i.e. the model which achieves the maximal average probability in the given window.

The window size and position can be defined once in the parameters, or can be be given directly to the required apply methods.

It is also possible to get the lti::principalComponents object that will zoom into the model colors. This is mainly used to zoom into skin colors, to allow the segmentation between skin-similar and skin colors. The computation of the PCA will be delayed until the first apply requires it, or when the user explicitelly calls initPCA. This is done this way to avoid waiting times in those cases where the color zoom is not required.

Constructor & Destructor Documentation

lti::colorModelSelector::colorModelSelector ( )

default constructor

lti::colorModelSelector::colorModelSelector ( const colorModelSelector & other )

copy constructor

Parameters:

other

the object to be copied

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

destructor

Member Function Documentation

bool lti::colorModelSelector::addColorModel	(	const std::string &	name,
		const probabilityMap &	colorMap
	)

add a further color model to a intern list

Parameters:

	name	the name for this model
	colorMap	a probabilityMap functor with the color and non-color models.

bool lti::colorModelSelector::addColorModel	(	const std::string &	name,
		const thistogram< double > &	colorHist,
		const thistogram< double > &	nonColorHist
	)

add a further color model to a intern list

Parameters:

	name	the name for this model
	colorHist	thistogram<double> with the color model
	nonColorHist	thistogram<double> with the non-color model.

bool lti::colorModelSelector::addColorModel	(	const std::string &	name,
		const std::string &	colorHistFileName,
		const std::string &	nonColorHistFileName
	)

add a further color model to a intern list.

The model is assumed to be stored in a file of a thistogram<double> written with a lispStreamHandler.

If you want to write the histograms by yourself do something like:

 thistogram<double> myHist;  // the histogram to be stored in a file

 std::ofstream out("myFile.hist"); // the file stream
 lti::lispStreamHandler lsh(out);  // the lisp stream handler
 myHist.write(lsh);                // write to the lisp stream handler
 out.close();

Please note that you can also add all histogram models directly, and at the end you can write this functor with all its models (and pca) to be readed later. This can save lots of time, specially, because the pca do not need to be computed again.

Parameters:

	name	the name for this model
	colorHistFileName	complete file name (path+filename) for the thistogram<double> file with the color model
	nonColorHistFileName	complete file name (path+filename) for the thistogram<double> file with the non-color model.

bool lti::colorModelSelector::apply	(	const image &	img,
		const rectangle &	window,
		const probabilityMap *&	pColorMap,
		std::string &	modelName
	)			const

get a pointer to the internal stored color model functor that maximizes the likelihood for the image window specified in the parameters object.

Note that if this object is destroyed, the pointers returned will be invalid.

Parameters:

	img	image with the source data.
	window	the region of the image to be considered.
	pColorMap	pointer to the internal model which maximizes the color probability for the given image window.
	modelName	name of the model used

Returns:: true if apply successful or false otherwise.

bool lti::colorModelSelector::apply	(	const image &	img,
		const rectangle &	window,
		const probabilityMap *&	pColorMap
	)			const

get a pointer to the internal stored color model functor that maximizes the likelihood for the image window specified in the parameters object.

Note that if this object is destroyed, the pointers returned will be invalid.

Parameters:

	img	image with the source data.
	window	the region of the image to be considered.
	pColorMap	pointer to the internal model which maximizes the color probability for the given image window.

Returns:: true if apply successful or false otherwise.

bool lti::colorModelSelector::apply	(	const image &	img,
		const rectangle &	window,
		probabilityMap &	colorMap
	)			const

Copy into colorMap the internal functor for the color model that maximizes the likelihood for the image window specified in the parameters object.

Parameters:

	img	image with the source data.
	window	the region of the image to be considered.
	colorMap	the internal model which maximizes the color probability for the given image window will be copied here.

Returns:: true if apply successful or false otherwise.

Note that this member will copy the color model functor, and this takes some time. If you can ensure that this object will exist while using the model functor, you better get just a pointer to the internal objects.

bool lti::colorModelSelector::apply	(	const image &	img,
		const rectangle &	window,
		const probabilityMap *&	pColorMap,
		const principalComponents< float > *&	pPca,
		std::string &	modelName
	)

Get a pointer to the internal stored model and transformation functors that maximize the likelihood for the image window specified in the arguments.

Note that if this object is destroyed, the pointers returned will be invalid.

Parameters:

	img	image with the source data.
	window	the region of the image to be considered.
	pColorMap	pointer to the internal model which maximizes the color probability for the given image window.
	pPca	pointer to the internal PCA functor (whitening transform) computed from the colorMap to be used in color zooming.
	modelName	the name of the model used

Returns:: true if apply successful or false otherwise.

bool lti::colorModelSelector::apply	(	const image &	img,
		const probabilityMap *&	pColorMap,
		std::string &	modelName
	)			const

get a pointer to the internal stored color model functor that maximizes the likelihood for the image window specified in the parameters object.

Note that if this object is destroyed, the pointers returned will be invalid.

Parameters:

	img	image with the source data.
	pColorMap	pointer to the internal model which maximizes the color probability for the given image window.
	modelName	the name of the model used

Returns:: true if apply successful or false otherwise.

bool lti::colorModelSelector::apply	(	const image &	img,
		const probabilityMap *&	pColorMap
	)			const

get a pointer to the internal stored color model functor that maximizes the likelihood for the image window specified in the parameters object.

Note that if this object is destroyed, the pointers returned will be invalid.

Parameters:

	img	image with the source data.
	pColorMap	pointer to the internal model which maximizes the color probability for the given image window.

Returns:: true if apply successful or false otherwise.

bool lti::colorModelSelector::apply	(	const image &	img,
		probabilityMap &	colorMap
	)			const

copy into colorMap the internal functor for the color model that maximizes the likelihood for the image window specified in the parameters object.

Parameters:

	img	image with the source data.
	colorMap	the internal model which maximizes the color probability for the given image window will be copied here.

Returns:: true if apply successful or false otherwise.

Note that this member will copy the color model functor, and this takes some time. If you can ensure that this object will exist while using the model functor, you better get just a pointer to the internal objects.

bool lti::colorModelSelector::apply	(	const image &	img,
		const rectangle &	window,
		const probabilityMap *&	pColorMap,
		const principalComponents< float > *&	pPca
	)

get a pointer to the internal stored model and transformation functors that maximize the likelihood for the image window specified in the parameters.

Note that if this object is destroyed, the pointers returned will be invalid.

Parameters:

	img	image with the source data.
	window	the region of the image to be considered.
	pColorMap	pointer to the internal model which maximizes the color probability for the given image window.
	pPca	pointer to the internal PCA functor (whitening transform) computed from the colorMap to be used in color zooming.

Returns:: true if apply successful or false otherwise.

bool lti::colorModelSelector::apply	(	const image &	img,
		const rectangle &	window,
		probabilityMap &	colorMap,
		principalComponents< float > &	pca
	)

copy into colorMap and pca the internal functors for the color models that maximize the likelihood for the image window specified in the parameters.

Parameters:

	img	image with the source data.
	window	the region of the image to be considered.
	colorMap	the internal model which maximizes the color probability for the given image window will be copied here.
	pca	functor with the whitening transform used to zoom the color model into the whole RGB color space.

Returns:: true if apply successful or false otherwise.

Note that this member will copy the models and pca transformation functors, and this takes some time. If you can ensure that this object will exist while using the models and pca functors, you better get just a pointer to the internal objects.

bool lti::colorModelSelector::apply	(	const image &	img,
		const probabilityMap *&	pColorMap,
		const principalComponents< float > *&	pPca
	)

get a pointer into the internal stored model and transformation functors that maximize the likelihood for the image window specified in the parameters object.

Note that if this object is destroyed, the pointers returned will be invalid.

Parameters:

	img	image with the source data.
	pColorMap	pointer to the internal model which maximizes the color probability for the given image window.
	pPca	pointer to the internal PCA functor (whitening transform) computed from the colorMap to be used in color zooming.

Returns:: true if apply successful or false otherwise.

bool lti::colorModelSelector::apply	(	const image &	img,
		probabilityMap &	colorMap,
		principalComponents< float > &	pca
	)

copy into colorMap and pca the internal functors for the color models that maximize the likelihood for the image window specified in the parameters object.

Parameters:

	img	image with the source data.
	colorMap	the internal model which maximizes the color probability for the given image window will be copied here.
	pca	functor with the whitening transform used to zoom the color model into the whole RGB color space.

Returns:: true if apply successful or false otherwise.

Note that this member will copy the models and pca transformation functors, and this takes some time. If you can ensure that this object will exist while using the models and pca functors, you better get just a pointer to the internal objects.

bool lti::colorModelSelector::clearColorModels ( )

remove all color models

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

returns a pointer to a clone of this functor.

Implements lti::functor.

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

copy data of "other" functor.

Parameters:

other

the functor to be copied

Returns:: a reference to this functor object

Reimplemented from lti::functor.

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

returns used parameters

Reimplemented from lti::transform.

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

returns the name of this type ("colorModelSelector")

Reimplemented from lti::transform.

bool lti::colorModelSelector::initPCA ( )

call if you want to precompute the PCA (Whitening Transformation) for the color models, before calling the corresponding apply methods.

Otherwise, the pca will be initialized on-the-fly (when required).

colorModelSelector& lti::colorModelSelector::operator= ( const colorModelSelector & other )

alias for copy member

Parameters:

other

the functor to be copied

Returns:: a reference to this functor object

Reimplemented from lti::functor.

bool lti::colorModelSelector::read	(	ioHandler &	handler,
		const bool	complete = `true`
	)			`[virtual]`

read the colorModelSelector from the given ioHandler

Parameters:

	handler	the ioHandler to be used
	complete	if true (the default) the enclosing begin/end will be also written, otherwise only the data block will be written.

Returns:: true if write was successful

Reimplemented from lti::functor.

bool lti::colorModelSelector::removeColorModel ( const std::string & name )

remove only the color models with the given name.

bool lti::colorModelSelector::updateModels ( const probabilityMap::parameters & newAttribs )

Update parameters for all models.

The color models are considered here as state, and not as part of the probabilityMap::parameters. I.e. all attributes except the histograms themselves will be updated in each of the availabe functors.

bool lti::colorModelSelector::write	(	ioHandler &	handler,
		const bool	complete = `true`
	)			const `[virtual]`

write the colorModelSelector in the given ioHandler

Parameters:

	handler	the ioHandler to be used
	complete	if true (the default) the enclosing begin/end will be also written, otherwise only the data block will be written.

Returns:: true if write was successful

Reimplemented from lti::functor.

Member Data Documentation

std::list<namedModel> lti::colorModelSelector::modelList [protected]

list with all color models

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

ltiColorModelSelector.h

lti::colorModelSelector Class Reference

Classes

Public Member Functions

Protected Attributes

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation

Member Data Documentation