Package qupath.opencv.ml

Class OpenCVClassifiers.LogisticRegressionClassifier

java.lang.Object

qupath.opencv.ml.OpenCVClassifiers.OpenCVStatModel

qupath.opencv.ml.OpenCVClassifiers.LogisticRegressionClassifier

Enclosing class:

OpenCVClassifiers

public static class OpenCVClassifiers.LogisticRegressionClassifier
extends OpenCVClassifiers.OpenCVStatModel

Classifier based on LogisticRegression.

Method Summary

Modifier and Type	Method	Description
TrainData	createTrainData(Mat samples, Mat targets, Mat weights, boolean doMulticlass)	Create training data in the format required by this classifier.
String	getName()	User-friendly, readable name for the classifier
ParameterList	getParameterList()	Retrieve a list of adjustable parameter that can be used to customize the classifier.
protected int	getTrainFlags()
boolean	isTrained()	Classifier has already been trained and is ready to predict.
void	predict(Mat samples, Mat results, Mat probabilities)	Default implementation calling
protected void	predictWithLock(Mat samples, Mat results, Mat probabilities)	Implement predictWithLock rather than predict to ensure predict is not called while training.
boolean	supportsAutoUpdate()	Returns true (the default value).
boolean	supportsMissingValues()	Tree classifiers in OpenCV support missing values, others do not.
boolean	supportsMulticlass()	Returns false (the default value).
boolean	supportsProbabilities()	Classifier can output a prediction confidence (expressed between 0 and 1), so may be interpreted as a probability...
String	toString()
void	train(TrainData trainData)	Train the classifier using data in an appropriate format.
void	trainWithLock(TrainData trainData)	Implement trainWithLock rather than train directly to ensure a lock is set when training, which can be used to prevent prediction occurring simultaneously.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

Method Details

createTrainData

public TrainData createTrainData(Mat samples,
 Mat targets,
 Mat weights,
 boolean doMulticlass)

Description copied from class: OpenCVClassifiers.OpenCVStatModel

Create training data in the format required by this classifier.

Parameters:

samples -

targets -

weights - optional weights

doMulticlass -

Returns:

See Also:

• OpenCVClassifiers.OpenCVStatModel.train(TrainData)

supportsMulticlass

public boolean supportsMulticlass()

Returns false (the default value).

Specified by:

supportsMulticlass in class OpenCVClassifiers.OpenCVStatModel

Returns:

true if multiclass classification is supported, false otherwise

supportsAutoUpdate

public boolean supportsAutoUpdate()

Returns true (the default value).

Specified by:

supportsAutoUpdate in class OpenCVClassifiers.OpenCVStatModel

Returns:

true if interactive classification is supported, false otherwise

supportsProbabilities

public boolean supportsProbabilities()

Description copied from class: OpenCVClassifiers.OpenCVStatModel

Classifier can output a prediction confidence (expressed between 0 and 1), so may be interpreted as a probability... even if it isn't necessarily one.

Specified by:

supportsProbabilities in class OpenCVClassifiers.OpenCVStatModel

Returns:

true if (pseudo-)probabilities can be provided

isTrained

public boolean isTrained()

Description copied from class: OpenCVClassifiers.OpenCVStatModel

Classifier has already been trained and is ready to predict.

Specified by:

isTrained in class OpenCVClassifiers.OpenCVStatModel

Returns:

true if the classifier is trained, false otherwise

getParameterList

public ParameterList getParameterList()

Description copied from class: OpenCVClassifiers.OpenCVStatModel

Retrieve a list of adjustable parameter that can be used to customize the classifier. After making changes to the ParameterList, the classifier should be retrained before being used.

Specified by:

getParameterList in class OpenCVClassifiers.OpenCVStatModel

Returns:

the parameter list for this classifier

toString

public String toString()

Overrides:

toString in class OpenCVClassifiers.OpenCVStatModel

train

public void train(TrainData trainData)

Description copied from class: OpenCVClassifiers.OpenCVStatModel

Train the classifier using data in an appropriate format.

Specified by:

train in class OpenCVClassifiers.OpenCVStatModel

Parameters:

trainData -

See Also:

• OpenCVClassifiers.OpenCVStatModel.createTrainData(Mat, Mat, Mat, boolean)

trainWithLock

public void trainWithLock(TrainData trainData)

Implement trainWithLock rather than train directly to ensure a lock is set when training, which can be used to prevent prediction occurring simultaneously.

Parameters:

trainData -

See Also:

• predictWithLock(org.bytedeco.opencv.opencv_core.Mat, org.bytedeco.opencv.opencv_core.Mat, org.bytedeco.opencv.opencv_core.Mat)

getTrainFlags

protected int getTrainFlags()

getName

public String getName()

Description copied from class: OpenCVClassifiers.OpenCVStatModel

User-friendly, readable name for the classifier

Specified by:

getName in class OpenCVClassifiers.OpenCVStatModel

Returns:

the classifier name

predict

public void predict(Mat samples,
 Mat results,
 Mat probabilities)

Default implementation calling

 statModel.predict(samples, results, 0);
 

before attempting to sanitize the outcome so that results always contains a signed int Mat containing classifications.

If results originally had more than 1 column, it will be returned as probabilities (if probabilities is not null); probabilities will be an empty matrix (i.e. no probabilities calculated).

Specified by:

predict in class OpenCVClassifiers.OpenCVStatModel

Parameters:

samples - the input samples

results - a Mat to receive the results

probabilities - a Mat to receive probability estimates, or null if probabilities are not needed

predictWithLock

protected void predictWithLock(Mat samples,
 Mat results,
 Mat probabilities)

Implement predictWithLock rather than predict to ensure predict is not called while training.

Parameters:

samples -

results -

probabilities -

See Also:

• trainWithLock(org.bytedeco.opencv.opencv_ml.TrainData)

supportsMissingValues

public boolean supportsMissingValues()

Tree classifiers in OpenCV support missing values, others do not.

Specified by:

supportsMissingValues in class OpenCVClassifiers.OpenCVStatModel

Returns:

true if NaNs are supported, false otherwise