#include <SceneLearningEngine.h>

Public Member Functions
void	extractTracksFromDbFile (const std::string &dbFileName)

void	generateSceneModel ()

void	learn ()

void	readLearnerInput ()

void	saveSceneModel ()

	SceneLearningEngine ()

void	visualizeSceneModel ()

	~SceneLearningEngine ()

Private Member Functions
void	initializeSceneModel (double pWorkspaceVolume, double mStaticBreakRatio, double mTogetherRatio, double mMaxAngleDeviation)

void	initializeVisualizationChain ()

Private Attributes
bool	mAddTimestampsToFilename

std::string	mBaseFrameId

std::string	mDateTime

ros::NodeHandle	mGeneratorHandle

std::string	mInputDbFilename

bool	mIntermediateResults

ros::NodeHandle	mPrivateNamespaceHandle

double	mScaleFactor

std::string	mSceneModelDirectory

boost::shared_ptr< SceneModelLearner >	mSceneModelLearner

std::string	mSceneModelName

std::string	mSceneModelType

double	mSigmaMultiplicator

boost::shared_ptr< Visualization::ProbabilisticSceneModelVisualization >	mVisualizer

ISM::RecordedPatternPtr	recordedPattern

boost::shared_ptr< ISM::TableHelper >	tableHelper

Detailed Description

Class for learning the model required for probabilistic scene recognition.

Author: Joachim Gehrung

Version: See SVN

Definition at line 48 of file SceneLearningEngine.h.

Constructor & Destructor Documentation

ProbabilisticSceneRecognition::SceneLearningEngine::SceneLearningEngine ( )

Constructor.

Definition at line 25 of file SceneLearningEngine.cpp.

ProbabilisticSceneRecognition::SceneLearningEngine::~SceneLearningEngine ( )

Destructor.

Definition at line 102 of file SceneLearningEngine.cpp.

Member Function Documentation

void ProbabilisticSceneRecognition::SceneLearningEngine::extractTracksFromDbFile ( const std::string & dbFileName )

Open database and extract ISM::ObjectSets. All object sets that the database contains, are transfered to the distribution parameter learners.

Parameters

pPbdSceneGraphsBagPath Path of file to be parsed ISM::ObjectSets.

Definition at line 119 of file SceneLearningEngine.cpp.

void ProbabilisticSceneRecognition::SceneLearningEngine::generateSceneModel ( )

All learnt information is transfered into a scene model description.

Definition at line 140 of file SceneLearningEngine.cpp.

void ProbabilisticSceneRecognition::SceneLearningEngine::initializeSceneModel	(	double	pWorkspaceVolume,
		double	mStaticBreakRatio,
		double	mTogetherRatio,
		double	mMaxAngleDeviation
	)

private

Initializes the scene model with the parameters specified via the node handle.

Parameters

pWorkspaceVolume	Volume of the workspace the scene takes place in.
pStaticBreakRatio	The maximum ration the relationship between two objects may break.
pTogetherRatio	The minimum ratio that two objects must be together.
pMaxAngleDeviation	Maximum angle deviation between two objects before counting as break.

Definition at line 185 of file SceneLearningEngine.cpp.

void ProbabilisticSceneRecognition::SceneLearningEngine::initializeVisualizationChain ( )

private

Initializes the chain responsible for visualization.

Definition at line 190 of file SceneLearningEngine.cpp.

void ProbabilisticSceneRecognition::SceneLearningEngine::learn ( )

Learn recorded pattern and pass it to mSampleList.

void ProbabilisticSceneRecognition::SceneLearningEngine::readLearnerInput ( )

Extract ISM::ObjectSets from all databases given as CLI parameters. Transfer all measurements for each scene into distribution parameter learners.

Definition at line 107 of file SceneLearningEngine.cpp.

void ProbabilisticSceneRecognition::SceneLearningEngine::saveSceneModel ( )

Learnt scene model is written to an xml file whose name consists of a scene model identifier set before and a timestamp.

Definition at line 153 of file SceneLearningEngine.cpp.

void ProbabilisticSceneRecognition::SceneLearningEngine::visualizeSceneModel ( )

All distributions in the decomposition of the scene model and the learning data are plotted.

Definition at line 174 of file SceneLearningEngine.cpp.

Member Data Documentation

bool ProbabilisticSceneRecognition::SceneLearningEngine::mAddTimestampsToFilename

private

Set true to add timestamps to filename.

Definition at line 142 of file SceneLearningEngine.h.

std::string ProbabilisticSceneRecognition::SceneLearningEngine::mBaseFrameId

private

The coordinate frame in which the visualization should take place.

Definition at line 147 of file SceneLearningEngine.h.

std::string ProbabilisticSceneRecognition::SceneLearningEngine::mDateTime

private

The timestamp for the point in time the learner was started.

Definition at line 162 of file SceneLearningEngine.h.

ros::NodeHandle ProbabilisticSceneRecognition::SceneLearningEngine::mGeneratorHandle

private

Interface to basic ros node functionality enhanced by the given class.

Definition at line 118 of file SceneLearningEngine.h.

std::string ProbabilisticSceneRecognition::SceneLearningEngine::mInputDbFilename

private

Path to the directory the .sqlite file containing the training data.

Definition at line 182 of file SceneLearningEngine.h.

bool ProbabilisticSceneRecognition::SceneLearningEngine::mIntermediateResults

private

Set true to visualize intermediate results instead.

Definition at line 137 of file SceneLearningEngine.h.

ros::NodeHandle ProbabilisticSceneRecognition::SceneLearningEngine::mPrivateNamespaceHandle

private

Interface to the private namespace of the ros node.

Definition at line 123 of file SceneLearningEngine.h.

double ProbabilisticSceneRecognition::SceneLearningEngine::mScaleFactor

private

The visualization is pretty small, this scale factor enlarges it.

Definition at line 152 of file SceneLearningEngine.h.

std::string ProbabilisticSceneRecognition::SceneLearningEngine::mSceneModelDirectory

private

Path to the directory the XML file containing the model is stored after learning.

Definition at line 177 of file SceneLearningEngine.h.

boost::shared_ptr<SceneModelLearner> ProbabilisticSceneRecognition::SceneLearningEngine::mSceneModelLearner

private

The learner for the probabilistic scene model.

Definition at line 191 of file SceneLearningEngine.h.

std::string ProbabilisticSceneRecognition::SceneLearningEngine::mSceneModelName

private

Name of the scene model.

Definition at line 167 of file SceneLearningEngine.h.

std::string ProbabilisticSceneRecognition::SceneLearningEngine::mSceneModelType

private

Determines the type of representation of the scene model that should be learned.

Definition at line 172 of file SceneLearningEngine.h.

double ProbabilisticSceneRecognition::SceneLearningEngine::mSigmaMultiplicator

private

This factor determines the radii of the covariance ellipse.

Definition at line 157 of file SceneLearningEngine.h.

boost::shared_ptr<Visualization::ProbabilisticSceneModelVisualization> ProbabilisticSceneRecognition::SceneLearningEngine::mVisualizer

private

Class for coordinating the scene visualizers.

Definition at line 196 of file SceneLearningEngine.h.

ISM::RecordedPatternPtr ProbabilisticSceneRecognition::SceneLearningEngine::recordedPattern

private

Recorded Pattern Pointer

Definition at line 187 of file SceneLearningEngine.h.

boost::shared_ptr<ISM::TableHelper> ProbabilisticSceneRecognition::SceneLearningEngine::tableHelper

private

TableHelper to extract Objects from ".sqlite"-file.

Definition at line 128 of file SceneLearningEngine.h.

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

Public Member Functions

Private Member Functions

Private Attributes

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation

Member Data Documentation