Digikam::FaceClassifier Class Reference

Inheritance diagram for Digikam::FaceClassifier:

Signals
void	signalTrainingComplete ()
	Emitted when the training process is complete.

Public Member Functions
void	cancel ()
int	predict (const cv::Mat &target, const QList< int > &exclusionLabelList=QList< int >()) const override
int	predict (const cv::UMat &target, const QList< int > &exclusionLabelList=QList< int >()) const override
bool	ready () const
bool	retrain () override
void	setParameters (const FaceScanSettings &parameters)
	Tunes backend parameters.
Public Member Functions inherited from Digikam::FaceClassifierBase
QList< int >	predictMulti (const QList< cv::Mat > &targets) const override
QList< int >	predictMulti (const QList< cv::UMat > &targets) const override
Public Member Functions inherited from Digikam::MLClassifierFoundation
void	setThreshold (float _threshold)

Static Public Member Functions
static FaceClassifier *	instance ()

Protected Member Functions
bool	loadTrainingData () override

Protected Attributes
const float	MAX_DISTANCE = 1000.0F
	arbitrary max distance to start calculating distance.
Protected Attributes inherited from Digikam::MLClassifierFoundation
QReadWriteLock	lock
float	threshold = 0.0F

Additional Inherited Members
Static Public Attributes inherited from Digikam::FaceClassifierBase
static const int	UNKNOWN_LABEL_ID = -1

Member Function Documentation

loadTrainingData()

bool Digikam::FaceClassifier::loadTrainingData ( )

overrideprotectedvirtual

Implements Digikam::MLClassifierFoundation.

predict() [1/2]

int Digikam::FaceClassifier::predict ( const cv::Mat &  target, const QList< int > &  exclusionLabelList = QList<int>()  ) const

overridevirtual

Implements Digikam::MLClassifierFoundation.

predict() [2/2]

int Digikam::FaceClassifier::predict ( const cv::UMat &  target, const QList< int > &  exclusionLabelList = QList<int>()  ) const

overridevirtual

Implements Digikam::MLClassifierFoundation.

retrain()

bool Digikam::FaceClassifier::retrain ( )

overridevirtual

Implements Digikam::MLClassifierFoundation.

setParameters()

void Digikam::FaceClassifier::setParameters

(

const FaceScanSettings &

parameters

)

Available parameters:

"accuracy", synonymous: "threshold", range: 0-1, type: float Determines recognition threshold, 0->accept very insecure recognitions, 1-> be very sure about a recognition.

"recognizeModel" : sets the recognizer model used to instantiate the correct recognizer