Base class for the definition of training sets. More...

#include <xmmTrainingSet.hpp>

Inheritance diagram for xmm::TrainingSet:

Collaboration diagram for xmm::TrainingSet:

Public Member Functions
	TrainingSet (MemoryMode memoryMode=MemoryMode::OwnMemory, Multimodality multimodality=Multimodality::Unimodal)
	Constructor. More...

	TrainingSet (TrainingSet const &src)
	Copy Constructor. More...

	TrainingSet (Json::Value const &root)
	Constructor from Json Structure. More...

TrainingSet &	operator= (TrainingSet const &src)
	Assignment Operator. More...

virtual	~TrainingSet ()
	Destructor. More...

Accessors
bool	ownMemory () const
	checks if the training set is owns the data More...

bool	bimodal () const
	checks if the training set is bimodal More...

Manipulate Phrases
bool	empty () const
	checks if the training set is empty More...

unsigned int	size () const
	Size of the training set. More...

std::map< int, std::shared_ptr< xmm::Phrase > >::iterator	begin ()
	iterator to the beginning of phrases More...

std::map< int, std::shared_ptr< xmm::Phrase > >::iterator	end ()
	iterator to the end of phrases More...

std::map< int, std::shared_ptr< xmm::Phrase > >::reverse_iterator	rbegin ()
	reverse iterator to the beginning of phrases More...

std::map< int, std::shared_ptr< xmm::Phrase > >::reverse_iterator	rend ()
	reverse iterator to the end of phrases More...

std::map< int, std::shared_ptr< xmm::Phrase > >::const_iterator	cbegin () const
	constant iterator to the beginning of phrases More...

std::map< int, std::shared_ptr< xmm::Phrase > >::const_iterator	cend () const
	constant iterator to the end of phrases More...

std::map< int, std::shared_ptr< xmm::Phrase > >::const_reverse_iterator	crbegin () const
	constant reverse iterator to the beginning of phrases More...

std::map< int, std::shared_ptr< xmm::Phrase > >::const_reverse_iterator	crend () const
	constant reverse iterator to the end of phrases More...

void	addPhrase (int phraseIndex, std::string label="")
	add a new phrase, or reset the phrase if existing More...

void	addPhrase (int phraseIndex, Phrase const &phrase)
	add a new phrase, or reset the phrase if existing More...

void	addPhrase (int phraseIndex, std::shared_ptr< Phrase > phrase)
	add a new phrase, or reset the phrase if existing More...

void	removePhrase (int phraseIndex)
	delete a phrase More...

void	removePhrasesOfClass (std::string const &label)
	delete all phrases of a given class More...

std::shared_ptr< xmm::Phrase >	getPhrase (int phraseIndex) const
	Access Phrase by index. More...

TrainingSet *	getPhrasesOfClass (std::string const &label)
	get the pointer to the sub-training set containing all phrases with a given label More...

void	clear ()
	delete all phrases More...

const std::set< std::string > &	labels () const
	get the list of labels currently in the training set More...

JSON I/O
Json::Value	toJson () const
	Write the object to a JSON Structure. More...

void	fromJson (Json::Value const &root)
	Read the object from a JSON Structure. More...

Utilities
std::vector< float >	mean () const
	Compute the global mean of all data phrases along the time axis. More...

std::vector< float >	standardDeviation () const
	Compute the global standard deviation of all data phrases along the time axis. More...

std::vector< std::pair< float, float > >	minmax () const
	Compute the global min/max of all data phrases along the time axis. More...

void	rescale (std::vector< float > offset, std::vector< float > gain)
	rescale a phrase given an offset and gain More...

void	normalize ()
	normalize the training set by rescaling all phrases to the mean/std of the whole training set More...

Python File I/O
void	writeFile (char *fileName) const
	write method for python wrapping ('write' keyword forbidden, name has to be different) More...

void	readFile (char *fileName)
	read method for python wrapping ('read' keyword forbidden, name has to be different) More...

std::string	__str__ () const
	"print" method for python => returns the results of write method More...

Public Attributes
Attribute< unsigned int >	dimension
	total dimension of the training data More...

Attribute< unsigned int >	dimension_input
	dimension of the input modality in bimodal mode More...

Attribute< std::vector< std::string > >	column_names
	labels of the columns of the training set (e.g. descriptor names) More...

Protected Member Functions
void	onPhraseEvent (PhraseEvent const &e)
	Monitors the training of each Model of the group. More...

virtual void	onAttributeChange (AttributeBase *attr_pointer)
	notification function called when a member attribute is changed More...

virtual void	update ()
	create all the sub-training sets: one for each label More...

Protected Attributes
bool	own_memory_
	defines if the phrase has its own memory More...

bool	bimodal_
	defines if the phrase is bimodal More...

std::set< std::string >	labels_
	Set containing all the labels present in the training set. More...

std::map< int, std::shared_ptr< Phrase > >	phrases_
	Training Phrases. More...

std::map< std::string, TrainingSet >	sub_training_sets_
	Sub-ensembles of the training set for specific classes. More...

Detailed Description

Base class for the definition of training sets.

Training sets host a collection of phrases

Constructor & Destructor Documentation

xmm::TrainingSet::TrainingSet	(	MemoryMode	memoryMode = `MemoryMode::OwnMemory`,
		Multimodality	multimodality = `Multimodality::Unimodal`
	)

Constructor.

Parameters

memoryMode	Memory mode (owned vs shared)
multimodality	Number of modalities

xmm::TrainingSet::TrainingSet ( TrainingSet const & src )

Copy Constructor.

Parameters

src	source Training Set

xmm::TrainingSet::TrainingSet ( Json::Value const & root )

explicit

Constructor from Json Structure.

Parameters

root	Json Value

xmm::TrainingSet::~TrainingSet ( )

virtual

Destructor.

Warning: phrases are only deleted if the training set is unlocked

See also: lock()

Member Function Documentation

std::string xmm::Writable::__str__ ( ) const

inlineinherited

"print" method for python => returns the results of write method

Warning: only defined if SWIGPYTHON is defined

void xmm::TrainingSet::addPhrase	(	int	phraseIndex,
		std::string	label = `""`
	)

add a new phrase, or reset the phrase if existing

The phrase is set to an empty phrase with the current attributes (dimensions, etc). The phrase is created if it does not exists at the given index.

Parameters

phraseIndex	index of the data phrase in the trainingSet
label	Label of the phrase

void xmm::TrainingSet::addPhrase	(	int	phraseIndex,
		Phrase const &	phrase
	)

add a new phrase, or reset the phrase if existing

The phrase is set to a copy of the phrase passed in argument The phrase is created if it does not exists at the given index.

Parameters

phraseIndex	index of the data phrase in the trainingSet
phrase	source phrase

void xmm::TrainingSet::addPhrase	(	int	phraseIndex,
		std::shared_ptr< Phrase >	phrase
	)

add a new phrase, or reset the phrase if existing

The phrase is set to point to the existing phrase passed in argument. The phrase is created if it does not exists at the given index.

Parameters

phraseIndex	index of the data phrase in the trainingSet
phrase	pointer to an existing phrase

std::map< int, std::shared_ptr< xmm::Phrase > >::iterator xmm::TrainingSet::begin ( )

iterator to the beginning of phrases

bool xmm::TrainingSet::bimodal ( ) const

checks if the training set is bimodal

Returns: true if the training set is bimodal (construction with Multimodality::Bimodal)

std::map< int, std::shared_ptr< xmm::Phrase > >::const_iterator xmm::TrainingSet::cbegin ( ) const

constant iterator to the beginning of phrases

std::map< int, std::shared_ptr< xmm::Phrase > >::const_iterator xmm::TrainingSet::cend ( ) const

constant iterator to the end of phrases

void xmm::TrainingSet::clear ( )

delete all phrases

Warning: if the training set is locked, the phrases themselves are not deleted (only their references)

std::map< int, std::shared_ptr< xmm::Phrase > >::const_reverse_iterator xmm::TrainingSet::crbegin ( ) const

constant reverse iterator to the beginning of phrases

std::map< int, std::shared_ptr< xmm::Phrase > >::const_reverse_iterator xmm::TrainingSet::crend ( ) const

constant reverse iterator to the end of phrases

bool xmm::TrainingSet::empty ( ) const

checks if the training set is empty

Returns: true if the training set is empty (no training phrases)

std::map< int, std::shared_ptr< xmm::Phrase > >::iterator xmm::TrainingSet::end ( )

iterator to the end of phrases

void xmm::TrainingSet::fromJson ( Json::Value const & root )

virtual

Read the object from a JSON Structure.

Parameters

root	JSON value containing the object's information

Exceptions

JsonException if the JSON value has a wrong format

Implements xmm::Writable.

std::shared_ptr< xmm::Phrase > xmm::TrainingSet::getPhrase ( int phraseIndex ) const

Access Phrase by index.

Parameters

phraseIndex index of the phrase

Returns: pointer to the phrase if it exists, else nullptr

xmm::TrainingSet * xmm::TrainingSet::getPhrasesOfClass ( std::string const & label )

get the pointer to the sub-training set containing all phrases with a given label

Warning: in order to protect the phrases in the current training set, the sub-training set returned is locked

Parameters

label target label

Returns: pointer to the sub-training set containing all phrases with a given label

Exceptions

out_of_range if the label does not exist

const std::set<std::string>& xmm::TrainingSet::labels ( ) const

inline

get the list of labels currently in the training set

Returns: reference to the set of labels in the training set

std::vector< float > xmm::TrainingSet::mean ( ) const

Compute the global mean of all data phrases along the time axis.

Returns: global mean of all phrases (along time axis, full-size)

std::vector< std::pair< float, float > > xmm::TrainingSet::minmax ( ) const

Compute the global min/max of all data phrases along the time axis.

Returns: vector of min/max pairs across all phrases (along time axis, full-size)

void xmm::TrainingSet::normalize ( )

normalize the training set by rescaling all phrases to the mean/std of the whole training set

void xmm::TrainingSet::onAttributeChange ( xmm::AttributeBase * attr_pointer )

protectedvirtual

notification function called when a member attribute is changed

void xmm::TrainingSet::onPhraseEvent ( PhraseEvent const & e )

protected

Monitors the training of each Model of the group.

xmm::TrainingSet & xmm::TrainingSet::operator= ( TrainingSet const & src )

Assignment Operator.

Parameters

src	source Training Set

bool xmm::TrainingSet::ownMemory ( ) const

checks if the training set is owns the data

Returns: true if the training set owns data (construction with MemoryMode::OwnMemory)

std::map< int, std::shared_ptr< xmm::Phrase > >::reverse_iterator xmm::TrainingSet::rbegin ( )

reverse iterator to the beginning of phrases

void xmm::Writable::readFile ( char * fileName )

inlineinherited

read method for python wrapping ('read' keyword forbidden, name has to be different)

Warning: only defined if SWIGPYTHON is defined

void xmm::TrainingSet::removePhrase ( int phraseIndex )

delete a phrase

Warning: if the training set is locked, the phrase iself is not deleted (only the reference)

Parameters

phraseIndex index of the phrase

Exceptions

out_of_bounds if the phrase does not exist

void xmm::TrainingSet::removePhrasesOfClass ( std::string const & label )

delete all phrases of a given class

Warning: if the training set is locked, the phrases themselves are not deleted (only the references)

Parameters

label label of the class to delete

Exceptions

out_of_bounds if the label does not exist

std::map< int, std::shared_ptr< xmm::Phrase > >::reverse_iterator xmm::TrainingSet::rend ( )

reverse iterator to the end of phrases

void xmm::TrainingSet::rescale	(	std::vector< float >	offset,
		std::vector< float >	gain
	)

rescale a phrase given an offset and gain

Parameters

offset	constant offset to be subtracted
gain	gain to be applied

unsigned int xmm::TrainingSet::size ( ) const

Size of the training set.

Returns: size of the training set (number of phrases)

std::vector< float > xmm::TrainingSet::standardDeviation ( ) const

Compute the global standard deviation of all data phrases along the time axis.

Returns: global standard deviation of all phrases (along time axis, full-size)

Json::Value xmm::TrainingSet::toJson ( ) const

virtual

Write the object to a JSON Structure.

Returns: Json value containing the object's information

Implements xmm::Writable.

void xmm::TrainingSet::update ( )

protectedvirtual

create all the sub-training sets: one for each label

each subset contains only the phrase for the given label

void xmm::Writable::writeFile ( char * fileName ) const

inlineinherited

write method for python wrapping ('write' keyword forbidden, name has to be different)

Warning: only defined if SWIGPYTHON is defined

Member Data Documentation

bool xmm::TrainingSet::bimodal_

protected

defines if the phrase is bimodal

Attribute<std::vector<std::string> > xmm::TrainingSet::column_names

labels of the columns of the training set (e.g. descriptor names)

Attribute<unsigned int> xmm::TrainingSet::dimension

total dimension of the training data

Attribute<unsigned int> xmm::TrainingSet::dimension_input

dimension of the input modality in bimodal mode

std::set<std::string> xmm::TrainingSet::labels_

protected

Set containing all the labels present in the training set.

bool xmm::TrainingSet::own_memory_

protected

defines if the phrase has its own memory

std::map<int, std::shared_ptr<Phrase> > xmm::TrainingSet::phrases_

protected

Training Phrases.

Phrases are stored in a map: allows the easy addition/deletion of phrases by index.

std::map<std::string, TrainingSet> xmm::TrainingSet::sub_training_sets_

protected

Sub-ensembles of the training set for specific classes.

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

XMM - Probabilistic Models for Motion Recognition and Mapping

Public Member Functions

Public Attributes

Protected Member Functions

Protected Attributes

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation

Member Data Documentation