Package rawDeepLearningClassifier.dlClassification.deepAcoustics

Class DeepAcousticsClassifier

java.lang.Object

rawDeepLearningClassifier.dlClassification.StandardClassifierModel

rawDeepLearningClassifier.dlClassification.archiveModel.ArchiveModelClassifier

rawDeepLearningClassifier.dlClassification.deepAcoustics.DeepAcousticsClassifier

All Implemented Interfaces:

PamSettings, SettingsNameProvider, DLClassiferModel

public class DeepAcousticsClassifier
extends ArchiveModelClassifier

A classifier based on the Deep Acoustics method which uses object detection models within spectrograms to predict dolphin whistle detections and then classify to species. This extends the ArchiveModelClassifier to provide functionality for interpreting the results from the Deep Acoustics model, which detects objects within input segments. Specifically the deep acoustics model can return multiple results per segment and bounding boxes (i.e. time and frequency limits of input data is not necassarily the same as the output) which is different from most other models

Author:

Jamie Macaulay

Nested Class Summary

Nested classes/interfaces inherited from class rawDeepLearningClassifier.dlClassification.StandardClassifierModel

StandardClassifierModel.StandardDLTaskThread

Field Summary

Fields

Modifier and Type	Field	Description
static String	MODEL_NAME

Fields inherited from class rawDeepLearningClassifier.dlClassification.archiveModel.ArchiveModelClassifier

archiveModelUI, archiveModelWorker

Fields inherited from class rawDeepLearningClassifier.dlClassification.StandardClassifierModel

dlControl

Constructor Summary

Constructors

Constructor	Description
DeepAcousticsClassifier(DLControl dlControl)

Method Summary

Modifier and Type	Method	Description
ArchiveModelWorker	getModelWorker()	Get the KetosWorker.
String	getName()	Get the name of the model.
StandardModelParams	makeParams()	Create the parameters class for the model.
void	prepModel()	Prepare the model.
protected ArrayList<ArrayList<? extends PredictionResult>>	processModelResults(ArrayList<? extends PamDataUnit> groupedRawData, List<StandardPrediction> modelResult)	Process the model results - for example to add class names and time stamps.

Methods inherited from class rawDeepLearningClassifier.dlClassification.archiveModel.ArchiveModelClassifier

getDLModelSettings, getDLParams, getDLWorker, getFileExtensions, getKetosParams, getModelUI, getSettingsReference, getSettingsVersion, getUnitName, getUnitType, isModelType, restoreSettings, setDLParams

Methods inherited from class rawDeepLearningClassifier.dlClassification.StandardClassifierModel

checkSettingsOK, closeModel, getAllowedDataTypes, getClassNameIDs, getClassNames, getDLControl, getModelStatus, getNumClasses, getRawSettingsPane, isDecision, isModelExtensions, millis2Samples, runModel, setModel

Methods inherited from class java.lang.Object

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

Field Details

MODEL_NAME

public static String MODEL_NAME

Constructor Details

DeepAcousticsClassifier

public DeepAcousticsClassifier(DLControl dlControl)

Method Details

getName

public String getName()

Description copied from interface: DLClassiferModel

Get the name of the model.

Specified by:

getName in interface DLClassiferModel

Overrides:

getName in class ArchiveModelClassifier

Returns:

the name of the model.

makeParams

public StandardModelParams makeParams()

Description copied from class: ArchiveModelClassifier

Create the parameters class for the model. This can be overridden for bespoke parameters. classes.

Overrides:

makeParams in class ArchiveModelClassifier

Returns:

a new parameters class object.

getModelWorker

public ArchiveModelWorker getModelWorker()

Get the KetosWorker. this handles loading and running the Ketos model.

Overrides:

getModelWorker in class ArchiveModelClassifier

Returns:

the Ketos worker.

prepModel

public void prepModel()

Description copied from interface: DLClassiferModel

Prepare the model. This is called on PAMGuard start up and before processing in viewer mode.

Specified by:

prepModel in interface DLClassiferModel

Overrides:

prepModel in class StandardClassifierModel

processModelResults

protected ArrayList<ArrayList<? extends PredictionResult>> processModelResults(ArrayList<? extends PamDataUnit> groupedRawData,
 List<StandardPrediction> modelResult)

Description copied from class: StandardClassifierModel

Process the model results - for example to add class names and time stamps.

Overrides:

processModelResults in class StandardClassifierModel

Parameters:

groupedRawData - - the grouped raw data used for input data into the model

modelResult - - the model results.