![]() |
GestureRecognitionToolkit
Version: 1.0 Revision: 04-03-15
The Gesture Recognition Toolkit (GRT) is a cross-platform, open-source, c++ machine learning library for real-time gesture recognition.
|
Public Types | |
typedef std::map< string, Classifier *(*)() > | StringClassifierMap |
![]() | |
enum | BaseTypes { BASE_TYPE_NOT_SET =0, CLASSIFIER, REGRESSIFIER, CLUSTERER } |
Public Member Functions | |
Classifier (void) | |
virtual | ~Classifier (void) |
virtual bool | deepCopyFrom (const Classifier *classifier) |
bool | copyBaseVariables (const Classifier *classifier) |
virtual bool | reset () |
virtual bool | clear () |
string | getClassifierType () const |
bool | getSupportsNullRejection () const |
bool | getNullRejectionEnabled () const |
double | getNullRejectionCoeff () const |
double | getMaximumLikelihood () const |
double | getBestDistance () const |
double | getPhase () const |
virtual UINT | getNumClasses () const |
UINT | getClassLabelIndexValue (UINT classLabel) const |
UINT | getPredictedClassLabel () const |
VectorDouble | getClassLikelihoods () const |
VectorDouble | getClassDistances () const |
VectorDouble | getNullRejectionThresholds () const |
vector< UINT > | getClassLabels () const |
vector< MinMax > | getRanges () const |
bool | enableNullRejection (bool useNullRejection) |
virtual bool | setNullRejectionCoeff (double nullRejectionCoeff) |
virtual bool | setNullRejectionThresholds (VectorDouble newRejectionThresholds) |
virtual bool | recomputeNullRejectionThresholds () |
bool | getTimeseriesCompatible () const |
Classifier * | createNewInstance () const |
Classifier * | deepCopy () const |
const Classifier * | getClassifierPointer () const |
const Classifier & | getBaseClassifier () const |
![]() | |
MLBase (void) | |
virtual | ~MLBase (void) |
bool | copyMLBaseVariables (const MLBase *mlBase) |
virtual bool | train (ClassificationData trainingData) |
virtual bool | train_ (ClassificationData &trainingData) |
virtual bool | train (RegressionData trainingData) |
virtual bool | train_ (RegressionData &trainingData) |
virtual bool | train (TimeSeriesClassificationData trainingData) |
virtual bool | train_ (TimeSeriesClassificationData &trainingData) |
virtual bool | train (TimeSeriesClassificationDataStream trainingData) |
virtual bool | train_ (TimeSeriesClassificationDataStream &trainingData) |
virtual bool | train (UnlabelledData trainingData) |
virtual bool | train_ (UnlabelledData &trainingData) |
virtual bool | train (MatrixDouble data) |
virtual bool | train_ (MatrixDouble &data) |
virtual bool | predict (VectorDouble inputVector) |
virtual bool | predict_ (VectorDouble &inputVector) |
virtual bool | predict (MatrixDouble inputMatrix) |
virtual bool | predict_ (MatrixDouble &inputMatrix) |
virtual bool | map (VectorDouble inputVector) |
virtual bool | map_ (VectorDouble &inputVector) |
virtual bool | print () const |
virtual bool | save (const string filename) const |
virtual bool | load (const string filename) |
virtual bool | saveModelToFile (string filename) const |
virtual bool | saveModelToFile (fstream &file) const |
virtual bool | loadModelFromFile (string filename) |
virtual bool | loadModelFromFile (fstream &file) |
virtual bool | getModel (ostream &stream) const |
double | scale (const double &x, const double &minSource, const double &maxSource, const double &minTarget, const double &maxTarget, const bool constrain=false) |
virtual string | getModelAsString () const |
UINT | getBaseType () const |
UINT | getNumInputFeatures () const |
UINT | getNumInputDimensions () const |
UINT | getNumOutputDimensions () const |
UINT | getMinNumEpochs () const |
UINT | getMaxNumEpochs () const |
UINT | getValidationSetSize () const |
UINT | getNumTrainingIterationsToConverge () const |
double | getMinChange () const |
double | getLearningRate () const |
double | getRootMeanSquaredTrainingError () const |
double | getTotalSquaredTrainingError () const |
bool | getUseValidationSet () const |
bool | getRandomiseTrainingOrder () const |
bool | getTrained () const |
bool | getModelTrained () const |
bool | getScalingEnabled () const |
bool | getIsBaseTypeClassifier () const |
bool | getIsBaseTypeRegressifier () const |
bool | getIsBaseTypeClusterer () const |
bool | enableScaling (bool useScaling) |
bool | setMaxNumEpochs (const UINT maxNumEpochs) |
bool | setMinNumEpochs (const UINT minNumEpochs) |
bool | setMinChange (const double minChange) |
bool | setLearningRate (double learningRate) |
bool | setUseValidationSet (const bool useValidationSet) |
bool | setValidationSetSize (const UINT validationSetSize) |
bool | setRandomiseTrainingOrder (const bool randomiseTrainingOrder) |
bool | registerTrainingResultsObserver (Observer< TrainingResult > &observer) |
bool | registerTestResultsObserver (Observer< TestInstanceResult > &observer) |
bool | removeTrainingResultsObserver (const Observer< TrainingResult > &observer) |
bool | removeTestResultsObserver (const Observer< TestInstanceResult > &observer) |
bool | removeAllTrainingObservers () |
bool | removeAllTestObservers () |
bool | notifyTrainingResultsObservers (const TrainingResult &data) |
bool | notifyTestResultsObservers (const TestInstanceResult &data) |
MLBase * | getMLBasePointer () |
const MLBase * | getMLBasePointer () const |
vector< TrainingResult > | getTrainingResults () const |
![]() | |
GRTBase (void) | |
virtual | ~GRTBase (void) |
bool | copyGRTBaseVariables (const GRTBase *GRTBase) |
string | getClassType () const |
string | getLastWarningMessage () const |
string | getLastErrorMessage () const |
string | getLastInfoMessage () const |
GRTBase * | getGRTBasePointer () |
const GRTBase * | getGRTBasePointer () const |
![]() | |
virtual void | notify (const TrainingResult &data) |
![]() | |
virtual void | notify (const TestInstanceResult &data) |
Static Public Member Functions | |
static Classifier * | createInstanceFromString (string const &classifierType) |
static vector< string > | getRegisteredClassifiers () |
![]() | |
static string | getGRTVersion (bool returnRevision=true) |
static string | getGRTRevison () |
Protected Types | |
enum | ClassifierModes { STANDARD_CLASSIFIER_MODE =0, TIMESERIES_CLASSIFIER_MODE } |
Protected Member Functions | |
bool | saveBaseSettingsToFile (fstream &file) const |
bool | loadBaseSettingsFromFile (fstream &file) |
![]() | |
bool | saveBaseSettingsToFile (fstream &file) const |
bool | loadBaseSettingsFromFile (fstream &file) |
![]() | |
double | SQR (const double &x) const |
Static Protected Member Functions | |
static StringClassifierMap * | getMap () |
Protected Attributes | |
string | classifierType |
bool | supportsNullRejection |
bool | useNullRejection |
UINT | numClasses |
UINT | predictedClassLabel |
UINT | classifierMode |
double | nullRejectionCoeff |
double | maxLikelihood |
double | bestDistance |
double | phase |
VectorDouble | classLikelihoods |
VectorDouble | classDistances |
VectorDouble | nullRejectionThresholds |
vector< UINT > | classLabels |
vector< MinMax > | ranges |
![]() | |
bool | trained |
bool | useScaling |
UINT | baseType |
UINT | numInputDimensions |
UINT | numOutputDimensions |
UINT | numTrainingIterationsToConverge |
UINT | minNumEpochs |
UINT | maxNumEpochs |
UINT | validationSetSize |
double | learningRate |
double | minChange |
double | rootMeanSquaredTrainingError |
double | totalSquaredTrainingError |
bool | useValidationSet |
bool | randomiseTrainingOrder |
Random | random |
vector< TrainingResult > | trainingResults |
TrainingResultsObserverManager | trainingResultsObserverManager |
TestResultsObserverManager | testResultsObserverManager |
![]() | |
string | classType |
DebugLog | debugLog |
ErrorLog | errorLog |
InfoLog | infoLog |
TrainingLog | trainingLog |
TestingLog | testingLog |
WarningLog | warningLog |
Definition at line 41 of file Classifier.h.
typedef std::map< string, Classifier*(*)() > GRT::Classifier::StringClassifierMap |
Defines a map between a string (which will contain the name of the classifier, such as ANBC) and a function returns a new instance of that classifier
Definition at line 248 of file Classifier.h.
GRT::Classifier::Classifier | ( | void | ) |
Default Classifier Constructor
Definition at line 67 of file Classifier.cpp.
|
virtual |
Default Classifier Destructor
Definition at line 84 of file Classifier.cpp.
|
virtual |
This function clears the classifier module, removing any trained model and setting all the base variables to their default values.
Reimplemented from GRT::MLBase.
Reimplemented in GRT::DTW, GRT::HMM, GRT::FiniteStateMachine, GRT::AdaBoost, GRT::ANBC, GRT::SVM, GRT::KNN, GRT::DecisionTree, GRT::BAG, GRT::GMM, GRT::RandomForests, GRT::ParticleClassifier, GRT::MinDist, GRT::Softmax, and GRT::SwipeDetector.
Definition at line 140 of file Classifier.cpp.
bool GRT::Classifier::copyBaseVariables | ( | const Classifier * | classifier | ) |
This copies the Classifier base class variables from the classifier pointer to this instance.
const | Classifier *classifier: a pointer to a classifier from which the values will be copied to this instance |
Definition at line 91 of file Classifier.cpp.
|
static |
Creates a new classifier instance based on the input string (which should contain the name of a valid classifier such as ANBC).
string | const &classifierType: the name of the classifier |
Definition at line 27 of file Classifier.cpp.
Classifier * GRT::Classifier::createNewInstance | ( | ) | const |
Creates a new classifier instance based on the current classifierType string value.
Definition at line 35 of file Classifier.cpp.
Classifier * GRT::Classifier::deepCopy | ( | ) | const |
This creates a new Classifier instance and deep copies the variables and models from this instance into the deep copy. The function will then return a pointer to the new instance. It is up to the user who calls this function to delete the dynamic instance when they are finished using it.
Definition at line 39 of file Classifier.cpp.
|
inlinevirtual |
This is the base deep copy function for the Classifier modules. This function should be overwritten by the derived class. This deep copies the variables and models from the classifier pointer to this classifier instance.
const | Classifier *classifier: a pointer to the Classifier base class, this should be pointing to another instance of a matching derived class |
Reimplemented in GRT::DTW, GRT::AdaBoost, GRT::SVM, GRT::LDA, GRT::KNN, GRT::DecisionTree, GRT::ANBC, GRT::HMM, GRT::RandomForests, GRT::GMM, GRT::BAG, GRT::MinDist, GRT::Softmax, GRT::SwipeDetector, GRT::FiniteStateMachine, and GRT::ParticleClassifier.
Definition at line 61 of file Classifier.h.
bool GRT::Classifier::enableNullRejection | ( | bool | useNullRejection | ) |
Sets if the classifier should use nullRejection.
If set to true then the classifier will reject a predicted class label if the likelihood of the prediction is below (or above depending on the algorithm) the models rejectionThreshold. If a prediction is rejected then the default null class label of 0 will be returned. If set to false then the classifier will simply return the most likely predicted class.
Definition at line 229 of file Classifier.cpp.
const Classifier & GRT::Classifier::getBaseClassifier | ( | ) | const |
Returns a pointer to this classifier. This is useful for a derived class so it can get easy access to this base classifier.
Definition at line 250 of file Classifier.cpp.
double GRT::Classifier::getBestDistance | ( | ) | const |
Returns the current bestDistance value. The bestDistance value is computed during the prediction phase and is either the minimum or maximum distance, depending on the algorithm. This value will return 0 if a prediction has not been made.
Definition at line 184 of file Classifier.cpp.
VectorDouble GRT::Classifier::getClassDistances | ( | ) | const |
Gets a vector of the class distances from the last prediction, this will be an N-dimensional vector, where N is the number of classes in the model. The exact form of these distances depends on the classification algorithm.
Definition at line 211 of file Classifier.cpp.
const Classifier * GRT::Classifier::getClassifierPointer | ( | ) | const |
Returns a pointer to the classifier.
Definition at line 52 of file Classifier.cpp.
string GRT::Classifier::getClassifierType | ( | ) | const |
Returns the classifier type as a string.
Definition at line 159 of file Classifier.cpp.
UINT GRT::Classifier::getClassLabelIndexValue | ( | UINT | classLabel | ) | const |
Gets the index of the query classLabel in the classLabels vector. If the query classLabel does not exist in the classLabels vector then the function will return zero.
UINT | classLabel: the query classLabel |
Definition at line 193 of file Classifier.cpp.
vector< UINT > GRT::Classifier::getClassLabels | ( | ) | const |
Gets a vector containing the label each class represents, this will be an N-dimensional vector, where N is the number of classes in the model. This is useful if the model was trained with non-monotonically class labels (i.e. class labels such as [1, 3, 6, 9, 12] instead of [1, 2, 3, 4, 5]).
Definition at line 221 of file Classifier.cpp.
VectorDouble GRT::Classifier::getClassLikelihoods | ( | ) | const |
Gets a vector of the class likelihoods from the last prediction, this will be an N-dimensional vector, where N is the number of classes in the model. The exact form of these likelihoods depends on the classification algorithm.
Definition at line 206 of file Classifier.cpp.
double GRT::Classifier::getMaximumLikelihood | ( | ) | const |
Returns the current maximumLikelihood value. The maximumLikelihood value is computed during the prediction phase and is the likelihood of the most likely model. This value will return 0 if a prediction has not been made.
Definition at line 175 of file Classifier.cpp.
double GRT::Classifier::getNullRejectionCoeff | ( | ) | const |
Returns the current nullRejectionCoeff value. The nullRejectionCoeff parameter is a multipler controlling the null rejection threshold for each class.
Definition at line 171 of file Classifier.cpp.
bool GRT::Classifier::getNullRejectionEnabled | ( | ) | const |
Returns true if nullRejection is enabled.
Definition at line 167 of file Classifier.cpp.
VectorDouble GRT::Classifier::getNullRejectionThresholds | ( | ) | const |
Gets a vector containing the null rejection thresholds for each class, this will be an N-dimensional vector, where N is the number of classes in the model.
Definition at line 216 of file Classifier.cpp.
|
virtual |
Gets the number of classes in trained model.
Reimplemented in GRT::SVM.
Definition at line 189 of file Classifier.cpp.
double GRT::Classifier::getPhase | ( | ) | const |
This function returns the estimated gesture phase from the most recent prediction. This value is only relevant if the classifier supports timeseries classification.
Definition at line 180 of file Classifier.cpp.
UINT GRT::Classifier::getPredictedClassLabel | ( | ) | const |
Gets the predicted class label from the last prediction.
Definition at line 201 of file Classifier.cpp.
vector< MinMax > GRT::Classifier::getRanges | ( | ) | const |
Gets a vector of the ranges used to scale the data for training and prediction, these ranges are only used if the classifier has been trained with the #useScaling flag set to true. This should be an N-dimensional vector, where N is the number of features in your data.
Definition at line 225 of file Classifier.cpp.
|
static |
Returns a vector of the names of all classifiers that have been registered with the base classifier.
Definition at line 56 of file Classifier.cpp.
bool GRT::Classifier::getSupportsNullRejection | ( | ) | const |
Returns true if the classifier instance supports null rejection, false otherwise.
Definition at line 163 of file Classifier.cpp.
|
inline |
Indicates if the classifier can be used to classify timeseries data. If true then the classifier can accept training data in the LabelledTimeSeriesClassificationData format.
return returns true if the classifier can be used to classify timeseries data, false otherwise
Definition at line 243 of file Classifier.h.
|
protected |
Loads the core base settings from a file.
Definition at line 301 of file Classifier.cpp.
|
inlinevirtual |
Recomputes the null rejection thresholds for each model.
Reimplemented in GRT::DTW, GRT::AdaBoost, GRT::ANBC, GRT::KNN, GRT::GMM, GRT::MinDist, and GRT::DecisionTree.
Definition at line 235 of file Classifier.h.
|
virtual |
This resets the classifier. This overrides the reset function in the MLBase base class.
Reimplemented from GRT::MLBase.
Reimplemented in GRT::DTW, GRT::HMM, GRT::FiniteStateMachine, GRT::ParticleClassifier, GRT::ANBC, GRT::SwipeDetector, and GRT::BAG.
Definition at line 121 of file Classifier.cpp.
|
protected |
Saves the core base settings to a file.
Definition at line 254 of file Classifier.cpp.
|
virtual |
Sets the nullRejectionCoeff, this is a multipler controlling the null rejection threshold for each class.
Reimplemented in GRT::KNN, GRT::ANBC, GRT::AdaBoost, and GRT::MinDist.
Definition at line 234 of file Classifier.cpp.
|
virtual |
Manually sets the nullRejectionThresholds, these are the thresholds used for null rejection for each class. This needs to be called after the model has been trained. Calling the setNullRejectionCoeff or recomputeNullRejectionThresholds functions will override these values. The size of the newRejectionThresholds vector must match the number of classes in the model.
VectorDouble | newRejectionThresholds: the new rejection thresholds |
Definition at line 242 of file Classifier.cpp.