eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics Class Reference

Collection of metrics used to evaluate a defect detection problem where the detection is based on the thresholding of a score produced by the underlying deep learning tool, e.g. a EUnsupervisedSegmenter tool or a ESupervisedSegmenter tool. More...

#include <eVision/EasyDeepLearning/EDeepLearningDefectDetectionMetrics.h>

Inheritance diagram for eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics:

Collaboration diagram for eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics:

Public Member Functions
	EDeepLearningDefectDetectionMetrics ()
	Constructs an empty EDeepLearningDefectDetectionMetrics object.
	EDeepLearningDefectDetectionMetrics (const eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics &other)
	Copy constructor.
virtual	~EDeepLearningDefectDetectionMetrics () throw ()
eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics &	operator= (const eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics &other)
	Copy operator.
virtual void	Load (const std::string &path)
	Loads the defect detection metrics. The given ESerializer must have been created for reading.
virtual void	Load (eVision::ESerializer *serializer)
	Loads the defect detection metrics. The given ESerializer must have been created for reading.
virtual void	Save (const std::string &path)
	Saves the defect detection metrics. The given ESerializer must have been created for writing.
virtual void	Save (eVision::ESerializer *serializer)
	Saves the defect detection metrics. The given ESerializer must have been created for writing.
bool	IsDefectDetectionMetricsValid () const
	Whether the defect detection metrics are valid or not. Defect detection metrics are completely valid when the metrics has results for at least one good and one defective images. Some metrics might be valid with only defective or good results.
int	GetNumDefectiveSample () const
	Number of defective sample added to the metric.
int	GetNumGoodSample () const
	Number of good sample added to the metric.
float	GetAccuracy (int index) const
	The accuracy of the segmenter.
float	GetAccuracy () const
	The accuracy of the segmenter.
float	GetBestAccuracy () const
	Best achievable accuracy.
float	GetBestBalancedAccuracy () const
	Best achievable balanced accuracy.
float	GetBestWeightedAccuracy (float goodWeight, float badWeight) const
	Best achievable weighted accuracy.
float	GetBestAccuracyClassificationThreshold () const
	Classification threshold giving the best achievable accuracy (see EDeepLearningDefectDetectionMetrics::GetBestAccuracy).
float	GetBestBalancedAccuracyClassificationThreshold () const
	Classification threshold giving the best achievable balanced accuracy (see EDeepLearningDefectDetectionMetrics::GetBestBalancedAccuracy).
float	GetBestWeightedAccuracyClassificationThreshold (float goodWeight, float badWeight) const
	Classification threshold giving the best achievable weighted accuracy (see EDeepLearningDefectDetectionMetrics::GetBestWeightedAccuracy).
uint32_t	GetConfusion (eVision::EasyDeepLearning::EConfusionMatrixElement element, int index) const
	Confusion value of one label with another.
uint32_t	GetConfusion (eVision::EasyDeepLearning::EConfusionMatrixElement element) const
	Confusion value of one label with another.
float	GetAreaUnderROCCurve () const
	The area under ROC curve (AUC) of the classifier (see EDeepLearningDefectDetectionMetrics::GetROCPoint). It's value is between 0 and 1.
float	GetAveragePrecision () const
	Average precision.
float	GetBestFScore () const
	Best F1-Score.
float	GetBestFScoreThreshold () const
	Classification threshold that yields the best F1-Score.
float	GetRecall (int index) const
	Recall for the current threshold.
float	GetRecall () const
	Recall for the current threshold.
float	GetPrecision (int index) const
	Precision for the current threshold.
float	GetPrecision () const
	Precision for the current threshold.
float	GetFScore (int index) const
	The F1-Score for the current threshold.
float	GetFScore () const
	The F1-Score for the current threshold.
int	GetPrecisionRecallCurveIndex () const
	Index in the precision/recall curve for the current EDeepLearningDefectDetectionMetrics::GetClassificationThreshold.
int	GetNumPrecisionRecallCurvePoint () const
	Number of point in the precision/recall curve.
eVision::EasyDeepLearning::EROCPoint	GetPrecisionRecallCurvePoint (int index) const
	Get a point on the precision/recall curve.
eVision::EasyDeepLearning::EROCPoint	GetPrecisionRecallCurvePoint () const
	Get a point on the precision/recall curve.
int	GetNumberOfClassifiers () const
	Number of different possible classifiers.
eVision::EasyDeepLearning::EROCPoint	GetROCPoint (int index) const
	ROC (Receiver Operating Characteristic) point.
eVision::EasyDeepLearning::EROCPoint	GetROCPoint () const
	ROC (Receiver Operating Characteristic) point.
float	GetClassificationThreshold () const
	Classification threshold.
void	SetClassificationThreshold (float threshold)
	Classification threshold.
	EDeepLearningDefectDetectionMetrics (CppWrapperInternal::InternalConstructor)

Public Attributes
CppWrapperInternal::WrappedClass	wrappedClass_

Detailed Description

Collection of metrics used to evaluate a defect detection problem where the detection is based on the thresholding of a score produced by the underlying deep learning tool, e.g. a EUnsupervisedSegmenter tool or a ESupervisedSegmenter tool.

These metrics are only valid when results for good and defective images are included in the metrics (see EDeepLearningDefectDetectionMetrics::IsDefectDetectionMetricsValid). The definition of what is considered a good or a defective image depends on the deep learning tool used.

The defect detection metrics are separated in two main categories:

• Metrics dependent on the ROC (Receiver Operating Characteristic) curve. They require at least one good and one defective sample to be defined.
• Metrics dependent on the Precision/Recall curve. They require at least one defective sample to be defined.

The metrics related to the ROC curve are:

• The accuracy (see EDeepLearningDefectDetectionMetrics::GetAccuracy).
• The confusion matrix (see EDeepLearningDefectDetectionMetrics::GetConfusion and EConfusionMatrixElement)
• The ROC curve (see EDeepLearningDefectDetectionMetrics::GetROCPoint)
• The area Under ROC curve (see EDeepLearningDefectDetectionMetrics::GetAreaUnderROCCurve)

The metrics related to the precision/recall curve are:

• The average precision (see EDeepLearningDefectDetectionMetrics::GetAveragePrecision)
• Precision/Recall curve (see EDeepLearningDefectDetectionMetrics::GetPrecisionRecallCurvePoint)
• Precision (see EDeepLearningDefectDetectionMetrics::GetPrecision)
• Recall (see EDeepLearningDefectDetectionMetrics::GetRecall)
• F-Score (see EDeepLearningDefectDetectionMetrics::GetFScore).

The ROC and Precision/Recall curve are both obtained by computing some metrics for different values of the EDeepLearningDefectDetectionMetrics::GetClassificationThreshold. Thus, each point on the curves gives different metrics, except for the area under the ROC curve and the average precision.

By default, the value of the metrics corresponds to the classification threshold of the corresponding deep learning tool. However, you can specify an index to retrieve the value of the metrics for other value of the EDeepLearningDefectDetectionMetrics::GetClassificationThreshold. Metrics based on the ROC curve are indexed between 0 and EDeepLearningDefectDetectionMetrics::GetNumberOfClassifiers-1 and metrics based on the Precision/Recall curve are indexed between 0 and EDeepLearningDefectDetectionMetrics::GetNumPrecisionRecallCurvePoint-1.

License(s): BasicTypes

Constructor & Destructor Documentation

EDeepLearningDefectDetectionMetrics() [1/3]

eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::EDeepLearningDefectDetectionMetrics

(

)

Constructs an empty EDeepLearningDefectDetectionMetrics object.

EDeepLearningDefectDetectionMetrics() [2/3]

eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::EDeepLearningDefectDetectionMetrics

(

const eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics &

other

)

Copy constructor.

Parameters

other

Other object

~EDeepLearningDefectDetectionMetrics()

eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::~EDeepLearningDefectDetectionMetrics ( ) throw (  )

virtual

EDeepLearningDefectDetectionMetrics() [3/3]

eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::EDeepLearningDefectDetectionMetrics

(

CppWrapperInternal::InternalConstructor

)

Member Function Documentation

GetAccuracy() [1/2]

float eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetAccuracy

(

)

const

The accuracy of the segmenter.

The accuracy is the number of images that were correctly classified over the total number of images that was used to evaluate the classifier.

GetAccuracy() [2/2]

float eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetAccuracy

(

int

index

)

const

The accuracy of the segmenter.

The accuracy is the number of images that were correctly classified over the total number of images that was used to evaluate the classifier.

Parameters

index

The index of the classifier to use. If the index is equal to '-1', the index corresponding to EUnsupervisedSegmenter::SetClassificationThreshold will be used.

GetAreaUnderROCCurve()

float eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetAreaUnderROCCurve

(

)

const

The area under ROC curve (AUC) of the classifier (see EDeepLearningDefectDetectionMetrics::GetROCPoint). It's value is between 0 and 1.

In the context of unsupervised segmentation, the AUC is equal to the probability that good images will have a lower reconstruction error than defective images.

This metrics measure discrimination capacity of a model. It tells how much the model is capable of distinguishing between classes. The higher the AUC, the better the model is.

GetAveragePrecision()

float eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetAveragePrecision

(

)

const

Average precision.

The average precision is the area under the precision-recall curve. The precision is the ratio between the number of true positive and the number of predicted positive and the recall, also called the true positive rate, is the ratio between the number of true positive and the number of positive sample.

GetBestAccuracy()

float eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetBestAccuracy

(

)

const

Best achievable accuracy.

The classification threshold corresponding to this accuracy is given by EDeepLearningDefectDetectionMetrics::GetBestAccuracyClassificationThreshold.

GetBestAccuracyClassificationThreshold()

float eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetBestAccuracyClassificationThreshold

(

)

const

Classification threshold giving the best achievable accuracy (see EDeepLearningDefectDetectionMetrics::GetBestAccuracy).

GetBestBalancedAccuracy()

float eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetBestBalancedAccuracy

(

)

const

Best achievable balanced accuracy.

The classification threshold corresponding to this accuracy is given by EDeepLearningDefectDetectionMetrics::GetBestBalancedAccuracyClassificationThreshold.

GetBestBalancedAccuracyClassificationThreshold()

float eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetBestBalancedAccuracyClassificationThreshold

(

)

const

Classification threshold giving the best achievable balanced accuracy (see EDeepLearningDefectDetectionMetrics::GetBestBalancedAccuracy).

GetBestFScore()

float eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetBestFScore

(

)

const

Best F1-Score.

The F1-Score is the harmonic mean of the precision and recall (true positive rate).

GetBestFScoreThreshold()

float eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetBestFScoreThreshold

(

)

const

Classification threshold that yields the best F1-Score.

The F1-Score is the harmonic mean of the precision and recall (true positive rate).

GetBestWeightedAccuracy()

float eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetBestWeightedAccuracy	(	float	goodWeight,
		float	badWeight ) const

Best achievable weighted accuracy.

The weighted accuracy is the weighted average of the true positive rate and the true negative rate (which is equal to 1 minus the false positive rate). See EROCPoint.

The classification threshold corresponding to this accuracy is given by EDeepLearningDefectDetectionMetrics::GetBestWeightedAccuracyClassificationThreshold.

Parameters

goodWeight	Weight for the good label
badWeight	Weight for the bad label

When using a dataset as the source for the label weights, the good weight is the weight of the "good" label and the bad weight is the sum of the weights of all the other labels.

GetBestWeightedAccuracyClassificationThreshold()

float eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetBestWeightedAccuracyClassificationThreshold	(	float	goodWeight,
		float	badWeight ) const

Classification threshold giving the best achievable weighted accuracy (see EDeepLearningDefectDetectionMetrics::GetBestWeightedAccuracy).

Parameters

goodWeight	Weight for the good label
badWeight	Weight for the bad label

GetClassificationThreshold()

float eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetClassificationThreshold

(

)

const

Classification threshold.

By default, the classification threshold will be equal to the classification threshold of the last unsupervised segmenter used to produce the results that compose this metric.

Some metrics such as EDeepLearningDefectDetectionMetrics::GetROCPoint, EDeepLearningDefectDetectionMetrics::GetAccuracy or EDeepLearningDefectDetectionMetrics::GetConfusion depends on the classification threshold. By default, these methods will returns the metric corresponding to the classification threshold.

Modifying the classification threshold in this class doesn't modify the classification threshold of the unsupervised segmenter.

GetConfusion() [1/2]

uint32_t eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetConfusion

(

eVision::EasyDeepLearning::EConfusionMatrixElement

element

)

const

Confusion value of one label with another.

The confusion value of a label with another is the number of images belonging to this label that are classified as belonging to the other label.

For a EDeepLearningDefectDetectionMetrics there are only 2 labels (good and defective) so the confusion matrix is only composed of 4 values which are called matrix element (see EConfusionMatrixElement).

The confusion matrix is computed for a given threshold (see EUnsupervisedSegmenter::SetClassificationThreshold) which means an index can be passed to the method (see EDeepLearningDefectDetectionMetrics::GetNumberOfClassifiers).

Parameters

element

The element from which to obtain the confusion value

GetConfusion() [2/2]

uint32_t eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetConfusion	(	eVision::EasyDeepLearning::EConfusionMatrixElement	element,
		int	index ) const

Confusion value of one label with another.

The confusion value of a label with another is the number of images belonging to this label that are classified as belonging to the other label.

For a EDeepLearningDefectDetectionMetrics there are only 2 labels (good and defective) so the confusion matrix is only composed of 4 values which are called matrix element (see EConfusionMatrixElement).

The confusion matrix is computed for a given threshold (see EUnsupervisedSegmenter::SetClassificationThreshold) which means an index can be passed to the method (see EDeepLearningDefectDetectionMetrics::GetNumberOfClassifiers).

Parameters

element	The element from which to obtain the confusion value
index	The index of the classifier to use. If the index is '-1', the index corresponding to EUnsupervisedSegmenter::SetClassificationThreshold will be used.

GetFScore() [1/2]

float eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetFScore

(

)

const

The F1-Score for the current threshold.

The F1-Score is the harmonic mean of EDeepLearningDefectDetectionMetrics::GetRecall and EDeepLearningDefectDetectionMetrics::GetPrecision.

GetFScore() [2/2]

float eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetFScore

(

int

index

)

const

The F1-Score for the current threshold.

The F1-Score is the harmonic mean of EDeepLearningDefectDetectionMetrics::GetRecall and EDeepLearningDefectDetectionMetrics::GetPrecision.

Parameters

index

Index for the precision/recall curve point or -1 to use the index corresponding to the current EDeepLearningDefectDetectionMetrics::GetClassificationThreshold.

GetNumberOfClassifiers()

int eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetNumberOfClassifiers

(

)

const

Number of different possible classifiers.

Each classifier is obtained by choosing a different classification threshold (see EUnsupervisedSegmenter::SetClassificationThreshold and corresponds to a point in the ROC curve (see EDeepLearningDefectDetectionMetrics::GetROCPoint).

The number of classifiers is equal to the number of results added to the metrics that have a unique classification score (i.e. different from the classification score of all the other results) plus one classifier with a threshold below all the results's score. This classifier might be invalid if the lowest score is equal to 0 (i.e. the threshold will be negative which is invalid).

GetNumDefectiveSample()

int eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetNumDefectiveSample

(

)

const

Number of defective sample added to the metric.

GetNumGoodSample()

int eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetNumGoodSample

(

)

const

Number of good sample added to the metric.

GetNumPrecisionRecallCurvePoint()

int eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetNumPrecisionRecallCurvePoint

(

)

const

Number of point in the precision/recall curve.

The precision/recall curve is defined when there is at least one ground truth positive sample in the metric. The number of point in the precision/recall curve is equal to the number of positive sample plus 1.

GetPrecision() [1/2]

float eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetPrecision

(

)

const

Precision for the current threshold.

The precision is the proportion of detected defective instances that were correctly identified as such.

GetPrecision() [2/2]

float eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetPrecision

(

int

index

)

const

Precision for the current threshold.

The precision is the proportion of detected defective instances that were correctly identified as such.

Parameters

index

Index for the precision/recall curve point or -1 to use the index corresponding to the current EDeepLearningDefectDetectionMetrics::GetClassificationThreshold.

GetPrecisionRecallCurveIndex()

int eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetPrecisionRecallCurveIndex

(

)

const

Index in the precision/recall curve for the current EDeepLearningDefectDetectionMetrics::GetClassificationThreshold.

The value of the index is between 0 and EDeepLearningDefectDetectionMetrics::GetNumPrecisionRecallCurvePoint - 1. The index is -1 if the precision/recall curve is not defined.

GetPrecisionRecallCurvePoint() [1/2]

eVision::EasyDeepLearning::EROCPoint eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetPrecisionRecallCurvePoint

(

)

const

Get a point on the precision/recall curve.

The precision/recall curve is defined when there is at least one ground truth positive sample in the metric.

GetPrecisionRecallCurvePoint() [2/2]

eVision::EasyDeepLearning::EROCPoint eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetPrecisionRecallCurvePoint

(

int

index

)

const

Get a point on the precision/recall curve.

Parameters

index

Index for the precision/recall curve point or -1 to use the index corresponding to the current EDeepLearningDefectDetectionMetrics::GetClassificationThreshold.

The precision/recall curve is defined when there is at least one ground truth positive sample in the metric.

GetRecall() [1/2]

float eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetRecall

(

)

const

Recall for the current threshold.

It is the proportion of defective instances that were correctly identified as such. It is also called the true positive rate.

GetRecall() [2/2]

float eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetRecall

(

int

index

)

const

Recall for the current threshold.

It is the proportion of defective instances that were correctly identified as such. It is also called the true positive rate.

Parameters

index

Index for the precision/recall curve point or -1 to use the index corresponding to the current EDeepLearningDefectDetectionMetrics::GetClassificationThreshold.

GetROCPoint() [1/2]

eVision::EasyDeepLearning::EROCPoint eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetROCPoint

(

)

const

ROC (Receiver Operating Characteristic) point.

A ROC point is a point from the ROC curve which is the plot of the true positive rate against the false positive rate (see EConfusionMatrixElement) obtained at various classification threshold (see EUnsupervisedSegmenter::SetClassificationThreshold).

The ROC points are strictly ordered by decreasing threshold order meaning the true positive rate and false positive rate (see EConfusionMatrixElement) are sorted in increasing order.

Each ROC point corresponds to a different classifier (see EDeepLearningDefectDetectionMetrics::GetNumberOfClassifiers).

It means that the ROC curve is the perfect tool to choose a threshold depending on the false and true positive rate values that best suit your application.

GetROCPoint() [2/2]

eVision::EasyDeepLearning::EROCPoint eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::GetROCPoint

(

int

index

)

const

ROC (Receiver Operating Characteristic) point.

A ROC point is a point from the ROC curve which is the plot of the true positive rate against the false positive rate (see EConfusionMatrixElement) obtained at various classification threshold (see EUnsupervisedSegmenter::SetClassificationThreshold).

The ROC points are strictly ordered by decreasing threshold order meaning the true positive rate and false positive rate (see EConfusionMatrixElement) are sorted in increasing order.

Parameters

index

The index of the classifier to use. If the index is equal to '-1', the index corresponding to EUnsupervisedSegmenter::SetClassificationThreshold will be used.

Each ROC point corresponds to a different classifier (see EDeepLearningDefectDetectionMetrics::GetNumberOfClassifiers).

It means that the ROC curve is the perfect tool to choose a threshold depending on the false and true positive rate values that best suit your application.

IsDefectDetectionMetricsValid()

bool eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::IsDefectDetectionMetricsValid

(

)

const

Whether the defect detection metrics are valid or not. Defect detection metrics are completely valid when the metrics has results for at least one good and one defective images. Some metrics might be valid with only defective or good results.

Load() [1/2]

void eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::Load ( const std::string & path )

virtual

Loads the defect detection metrics. The given ESerializer must have been created for reading.

Parameters

path	The file path.

Reimplemented in eVision::EasyDeepLearning::ESupervisedSegmenterMetrics, and eVision::EasyDeepLearning::EUnsupervisedSegmenterMetrics.

Load() [2/2]

void eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::Load ( eVision::ESerializer * serializer )

virtual

Loads the defect detection metrics. The given ESerializer must have been created for reading.

Parameters

serializer

The serializer.

Reimplemented in eVision::EasyDeepLearning::ESupervisedSegmenterMetrics, and eVision::EasyDeepLearning::EUnsupervisedSegmenterMetrics.

operator=()

eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics & eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::operator=

(

const eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics &

other

)

Copy operator.

Parameters

other

Other object

Save() [1/2]

void eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::Save ( const std::string & path )

virtual

Saves the defect detection metrics. The given ESerializer must have been created for writing.

Parameters

path	The file path.

Reimplemented in eVision::EasyDeepLearning::ESupervisedSegmenterMetrics, and eVision::EasyDeepLearning::EUnsupervisedSegmenterMetrics.

Save() [2/2]

void eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::Save ( eVision::ESerializer * serializer )

virtual

Saves the defect detection metrics. The given ESerializer must have been created for writing.

Parameters

serializer

The ESerializer object that is written to.

Reimplemented in eVision::EasyDeepLearning::ESupervisedSegmenterMetrics, and eVision::EasyDeepLearning::EUnsupervisedSegmenterMetrics.

SetClassificationThreshold()

void eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::SetClassificationThreshold

(

float

threshold

)

Classification threshold.

By default, the classification threshold will be equal to the classification threshold of the last unsupervised segmenter used to produce the results that compose this metric.

Some metrics such as EDeepLearningDefectDetectionMetrics::GetROCPoint, EDeepLearningDefectDetectionMetrics::GetAccuracy or EDeepLearningDefectDetectionMetrics::GetConfusion depends on the classification threshold. By default, these methods will returns the metric corresponding to the classification threshold.

Modifying the classification threshold in this class doesn't modify the classification threshold of the unsupervised segmenter.

Member Data Documentation

wrappedClass_

CppWrapperInternal::WrappedClass eVision::EasyDeepLearning::EDeepLearningDefectDetectionMetrics::wrappedClass_