org.openimaj.image.processing.face.recognition.benchmarking

Class CrossValidationBenchmark<PERSON,IMAGE extends Image<?,IMAGE>,FACE extends DetectedFace>

java.lang.Object

org.openimaj.image.processing.face.recognition.benchmarking.CrossValidationBenchmark<PERSON,IMAGE,FACE>

Type Parameters:

PERSON - type representing a person or class

IMAGE - type of image containing the face of each person

FACE - type of DetectedFace

All Implemented Interfaces:

RunnableExperiment

@Experiment(author="Jonathon Hare",
            dateCreated="2012-07-26",
            description="Face recognition cross validation experiment")
public class CrossValidationBenchmark<PERSON,IMAGE extends Image<?,IMAGE>,FACE extends DetectedFace>
extends Object
implements RunnableExperiment

An RunnableExperiment for performing cross-validation experiments on face recognisers & classifiers.

Author:

Jonathon Hare (jsh2@ecs.soton.ac.uk)

Field Summary

Fields

Modifier and Type	Field and Description
protected CrossValidator<GroupedDataset<PERSON,ListDataset<FACE>,FACE>>	crossValidator
protected GroupedDataset<PERSON,? extends ListDataset<IMAGE>,IMAGE>	dataset
protected FaceRecogniserProvider<FACE,PERSON>	engine
protected FaceDetector<FACE,IMAGE>	faceDetector
protected AggregatedCMResult<PERSON>	result

Constructor Summary

Constructors

Constructor and Description
CrossValidationBenchmark(CrossValidator<GroupedDataset<PERSON,ListDataset<FACE>,FACE>> crossValidator, GroupedDataset<PERSON,? extends ListDataset<IMAGE>,IMAGE> dataset, FaceDetector<FACE,IMAGE> faceDetector, FaceRecogniserProvider<FACE,PERSON> engine) Construct the CrossValidationBenchmark experiment with the given dependent variables.

Method Summary

Modifier and Type	Method and Description
void	finish(ExperimentContext context) Finalise the experiment, possibly writing the results, cleaning up, etc.
void	perform() Perform the experiment
void	setup() Setup of the experiment

Methods inherited from class java.lang.Object

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

Field Detail

crossValidator

@IndependentVariable
protected CrossValidator<GroupedDataset<PERSON,ListDataset<FACE extends DetectedFace>,FACE extends DetectedFace>> crossValidator

dataset

@IndependentVariable
protected GroupedDataset<PERSON,? extends ListDataset<IMAGE extends Image<?,IMAGE>>,IMAGE extends Image<?,IMAGE>> dataset

faceDetector

@IndependentVariable
protected FaceDetector<FACE extends DetectedFace,IMAGE extends Image<?,IMAGE>> faceDetector

engine

@IndependentVariable
protected FaceRecogniserProvider<FACE extends DetectedFace,PERSON> engine

result

@DependentVariable
protected AggregatedCMResult<PERSON> result

Constructor Detail

CrossValidationBenchmark

public CrossValidationBenchmark(CrossValidator<GroupedDataset<PERSON,ListDataset<FACE>,FACE>> crossValidator,
                                GroupedDataset<PERSON,? extends ListDataset<IMAGE>,IMAGE> dataset,
                                FaceDetector<FACE,IMAGE> faceDetector,
                                FaceRecogniserProvider<FACE,PERSON> engine)

Construct the CrossValidationBenchmark experiment with the given dependent variables.

Parameters:

dataset - the dataset

crossValidator - the cross-validator

faceDetector - the face detector

engine - the recogniser

Method Detail

perform

public void perform()

Description copied from interface: RunnableExperiment

Perform the experiment

Specified by:

perform in interface RunnableExperiment

setup

public void setup()

Description copied from interface: RunnableExperiment

Setup of the experiment

Specified by:

setup in interface RunnableExperiment

finish

public void finish(ExperimentContext context)

Description copied from interface: RunnableExperiment

Finalise the experiment, possibly writing the results, cleaning up, etc. The context gathered by the ExperimentRunner is provided so that RunnableExperiment implementations can access information for generating their reports.

Note that at the point the context is provided, it is not "locked" and is still recording. The ExperimentRunner returns a completed locked context after the experiment has been completely run.

Specified by:

finish in interface RunnableExperiment

Parameters:

context - the context provided by the ExperimentRunner