Collection of metrics used to evaluate the state of a eVision.EasyDeepLearning.ESupervisedSegmenter tool. More...

Inheritance diagram for eVision.EasyDeepLearning.ESupervisedSegmenterMetrics:

Collaboration diagram for eVision.EasyDeepLearning.ESupervisedSegmenterMetrics:

Public Member Functions
None	__init__ (self)
	Constructs an eVision.EasyDeepLearning.ESupervisedSegmenterMetrics object.
None	__init__ (self, eVision.EasyDeepLearning.ESupervisedSegmenterMetrics other)
	Copy constructor of an eVision.EasyDeepLearning.ESupervisedSegmenterMetrics object.
None	Load (self, str path)
	Loads an supervised segmentation metric.
None	Load (self, eVision.ESerializer\|None serializer)
	Loads an supervised segmentation metric.
None	Save (self, str path)
	Saves an supervised segmentation metric.
None	Save (self, eVision.ESerializer\|None serializer)
	Saves an supervised segmentation metric.
bool	IsValid (self)
	Indicates whether the object contains at least one result.
float	GetPixelLabelAccuracy (self, str label)
	Pixel label accuracy.
float	GetLabelError (self, str label)
	Label error.
int	GetPixelConfusion (self, str trueLabel, str predictedLabel)
	Pixel-wise confusion between the given true and predicted labels.
float	GetNormalizedPixelConfusion (self, str trueLabel, str predictedLabel)
	Pixel-wise normalized confusion between the given true and predicted labels.
float	GetIntersectionOverUnion (self, str label)
	Intersection over union.
int	GetGroundtruthBlobConfusion (self, str groundtruthLabel, str predictedLabel)
	Number of ground truth blobs of the given ground truth label that best match with blobs of the given predicted label.
int	GetGroundtruthBlobConfusion (self, int groundtruthLabelIndex, int predictedLabelIndex)
	Number of ground truth blobs of the given ground truth label that best match with blobs of the given predicted label.
int	GetPredictedBlobConfusion (self, str groundtruthLabel, str predictedLabel)
	Number of predicted blobs of the given predicted label that match to blobs of the given ground truth label.
int	GetPredictedBlobConfusion (self, int groundtruthLabelIndex, int predictedLabelIndex)
	Number of predicted blobs of the given predicted label that match to blobs of the given ground truth label.
str	GetLabel (self, int labelIdx)
	Label recognized by the segmenter that produced the results aggregated in the metrics.
Public Member Functions inherited from eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics
None	__init__ (self)
	Constructs an empty eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics object.
None	__init__ (self, eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics other)
	Copy constructor.
bool	IsDefectDetectionMetricsValid (self)
	Whether the defect detection metrics are valid or not.
float	GetAccuracy (self, int index)
	The accuracy of the segmenter.
float	GetAccuracy (self)
	The accuracy of the segmenter.
float	GetBestWeightedAccuracy (self, float goodWeight, float badWeight)
	Best achievable weighted accuracy.
float	GetBestWeightedAccuracyClassificationThreshold (self, float goodWeight, float badWeight)
	Classification threshold giving the best achievable weighted accuracy (see eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics.GetBestWeightedAccuracy).
int	GetConfusion (self, eVision.EasyDeepLearning.EConfusionMatrixElement element, int index)
	Confusion value of one label with another.
int	GetConfusion (self, eVision.EasyDeepLearning.EConfusionMatrixElement element)
	Confusion value of one label with another.
float	GetRecall (self, int index)
	Recall for the current threshold.
float	GetRecall (self)
	Recall for the current threshold.
float	GetPrecision (self, int index)
	Precision for the current threshold.
float	GetPrecision (self)
	Precision for the current threshold.
float	GetFScore (self, int index)
	The F1-Score for the current threshold.
float	GetFScore (self)
	The F1-Score for the current threshold.
eVision.EasyDeepLearning.EROCPoint	GetPrecisionRecallCurvePoint (self, int index)
	Get a point on the precision/recall curve.
eVision.EasyDeepLearning.EROCPoint	GetPrecisionRecallCurvePoint (self)
	Get a point on the precision/recall curve.
eVision.EasyDeepLearning.EROCPoint	GetROCPoint (self, int index)
	ROC (Receiver Operating Characteristic) point.
eVision.EasyDeepLearning.EROCPoint	GetROCPoint (self)
	ROC (Receiver Operating Characteristic) point.

Protected Member Functions
float	_WeightedPixelAccuracy__getter (self)
	\cond
float	_BalancedPixelAccuracy__getter (self)
	\cond
float	_PixelAccuracy__getter (self)
	\cond
float	_WeightedError__getter (self)
	\cond
float	_BalancedError__getter (self)
	\cond
float	_Error__getter (self)
	\cond
float	_WeightedIntersectionOverUnion__getter (self)
	\cond
float	_BalancedIntersectionOverUnion__getter (self)
	\cond
float	_BlobDetectionAveragePrecision__getter (self)
	\cond
float	_BlobDetectionBestFScore__getter (self)
	\cond
float	_BlobDetectionBestFScoreThreshold__getter (self)
	\cond
float	_BlobDetectionRecall__getter (self)
	\cond
float	_BlobDetectionPrecision__getter (self)
	\cond
float	_BlobDetectionFScore__getter (self)
	\cond
int	_NumLabels__getter (self)
	\cond
Protected Member Functions inherited from eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics
int	_NumDefectiveSample__getter (self)
	\cond
int	_NumGoodSample__getter (self)
	\cond
float	_BestAccuracy__getter (self)
	\cond
float	_BestBalancedAccuracy__getter (self)
	\cond
float	_BestAccuracyClassificationThreshold__getter (self)
	\cond
float	_BestBalancedAccuracyClassificationThreshold__getter (self)
	\cond
float	_AreaUnderROCCurve__getter (self)
	\cond
float	_AveragePrecision__getter (self)
	\cond
float	_BestFScore__getter (self)
	\cond
float	_BestFScoreThreshold__getter (self)
	\cond
int	_PrecisionRecallCurveIndex__getter (self)
	\cond
int	_NumPrecisionRecallCurvePoint__getter (self)
	\cond
int	_NumberOfClassifiers__getter (self)
	\cond
float	_ClassificationThreshold__getter (self)
	\cond
	_ClassificationThreshold__setter (self, float value)
	Classification threshold.

Properties
	WeightedPixelAccuracy = property(_WeightedPixelAccuracy__getter, None)
	\endcond
	BalancedPixelAccuracy = property(_BalancedPixelAccuracy__getter, None)
	\endcond
	PixelAccuracy = property(_PixelAccuracy__getter, None)
	\endcond
	WeightedError = property(_WeightedError__getter, None)
	\endcond
	BalancedError = property(_BalancedError__getter, None)
	\endcond
	Error = property(_Error__getter, None)
	\endcond
	WeightedIntersectionOverUnion = property(_WeightedIntersectionOverUnion__getter, None)
	\endcond
	BalancedIntersectionOverUnion = property(_BalancedIntersectionOverUnion__getter, None)
	\endcond
	BlobDetectionAveragePrecision = property(_BlobDetectionAveragePrecision__getter, None)
	\endcond
	BlobDetectionBestFScore = property(_BlobDetectionBestFScore__getter, None)
	\endcond
	BlobDetectionBestFScoreThreshold = property(_BlobDetectionBestFScoreThreshold__getter, None)
	\endcond
	BlobDetectionRecall = property(_BlobDetectionRecall__getter, None)
	\endcond
	BlobDetectionPrecision = property(_BlobDetectionPrecision__getter, None)
	\endcond
	BlobDetectionFScore = property(_BlobDetectionFScore__getter, None)
	\endcond
	NumLabels = property(_NumLabels__getter, None)
	\endcond
Properties inherited from eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics
	NumDefectiveSample = property(_NumDefectiveSample__getter, None)
	\endcond
	NumGoodSample = property(_NumGoodSample__getter, None)
	\endcond
	BestAccuracy = property(_BestAccuracy__getter, None)
	\endcond
	BestBalancedAccuracy = property(_BestBalancedAccuracy__getter, None)
	\endcond
	BestAccuracyClassificationThreshold = property(_BestAccuracyClassificationThreshold__getter, None)
	\endcond
	BestBalancedAccuracyClassificationThreshold = property(_BestBalancedAccuracyClassificationThreshold__getter, None)
	\endcond
	AreaUnderROCCurve = property(_AreaUnderROCCurve__getter, None)
	\endcond
	AveragePrecision = property(_AveragePrecision__getter, None)
	\endcond
	BestFScore = property(_BestFScore__getter, None)
	\endcond
	BestFScoreThreshold = property(_BestFScoreThreshold__getter, None)
	\endcond
	PrecisionRecallCurveIndex = property(_PrecisionRecallCurveIndex__getter, None)
	\endcond
	NumPrecisionRecallCurvePoint = property(_NumPrecisionRecallCurvePoint__getter, None)
	\endcond
	NumberOfClassifiers = property(_NumberOfClassifiers__getter, None)
	\endcond
	ClassificationThreshold = property(_ClassificationThreshold__getter, _ClassificationThreshold__setter)
	\endcond

Detailed Description

Collection of metrics used to evaluate the state of a eVision.EasyDeepLearning.ESupervisedSegmenter tool.

A metric is a value summarizing the quality of a collection of supervised segmentation results (see eVision.EasyDeepLearning.ESupervisedSegmenterResult) with respect to their ground truth. The eVision.EasyDeepLearning.ESupervisedSegmenterMetrics contains three types of metrics:

pixel based metrics that are related to the quality of the segmentation masks
blob based metrics that are related to the ability of the supervised segmentation tool to detect foreground blobs
defect detection metrics that are related to the ability of the supervised segmentation tool to differentiate between images that contains foreground pixels (defective images) and images that are entirely background (good images).

The pixel metrics are the error (see eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.Error), the pixel accuracy (see eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.PixelAccuracy), the pixel confusion (see eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetPixelConfusion), the pixel per-label accuracy (see eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetPixelLabelAccuracy) and the intersection over union (see eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetIntersectionOverUnion).

The blob based metrics are two confusion matrices (eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetGroundtruthBlobConfusion and eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetPredictedBlobConfusion), and various defect detection metrics (eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.BlobDetectionRecall, eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.BlobDetectionPrecision, and eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.BlobDetectionFScore). Note that the metrics for defective blob detection are different from the metrics for defective image detection because, in the case of blob detection, we have no "good" ground truth results.

See eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics for a description of the defect detection metrics. These metrics are available when eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics.IsDefectDetectionMetricsValid is true, i.e. when both images that are entirely background and images with non-background pixels have been added to the metrics.

Most metrics depends upon the eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics.ClassificationThreshold.

Remarks:

License(s): BasicTypes

Constructor & Destructor Documentation

◆ init() [1/2]

None eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.__init__ ( self )

Constructs an eVision.EasyDeepLearning.ESupervisedSegmenterMetrics object.

◆ init() [2/2]

None eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.__init__	(		self,
		eVision.EasyDeepLearning.ESupervisedSegmenterMetrics	other )

Copy constructor of an eVision.EasyDeepLearning.ESupervisedSegmenterMetrics object.

Parameters

other Reference to the eVision.EasyDeepLearning.ESupervisedSegmenterMetrics object that should be copied

Member Function Documentation

◆ _BalancedError__getter()

float eVision.EasyDeepLearning.ESupervisedSegmenterMetrics._BalancedError__getter ( self )

protected

\cond

Balanced error.

The balanced error is the weighted error (eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.WeightedError) where each label is given an equal weight.

Remarks:

The error is also called the cross-entropy loss.

◆ _BalancedIntersectionOverUnion__getter()

float eVision.EasyDeepLearning.ESupervisedSegmenterMetrics._BalancedIntersectionOverUnion__getter ( self )

protected

\cond

Balanced Intersection over Union (IoU).

The balanced intersection over union is the weighted intersection over union (eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.WeightedIntersectionOverUnion) where each label is given equal weight.

◆ _BalancedPixelAccuracy__getter()

float eVision.EasyDeepLearning.ESupervisedSegmenterMetrics._BalancedPixelAccuracy__getter ( self )

protected

\cond

Balanced accuracy.

The balanced accuracy is the weighted accuracy (eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.WeightedPixelAccuracy) where each label is given an equal weight.

◆ _BlobDetectionAveragePrecision__getter()

float eVision.EasyDeepLearning.ESupervisedSegmenterMetrics._BlobDetectionAveragePrecision__getter ( self )

protected

\cond

Average precision for blob detection.

Blob detection is the ability of the segmenter to correctly detect foreground ground truth blobs, regardless of their specific label. The average precision for blob detection is the average of the precision (see eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.BlobDetectionPrecision) over all the possible classification threshold values. As such, the average precision does not depend on a specific classification threshold.

◆ _BlobDetectionBestFScore__getter()

float eVision.EasyDeepLearning.ESupervisedSegmenterMetrics._BlobDetectionBestFScore__getter ( self )

protected

\cond

Best F1-Score of the segmenter for blob detection.

See eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.BlobDetectionBestFScoreThreshold for the corresponding threshold.

See eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics.GetFScore for a definition of the F1-Score. Blob detection is the ability of the segmenter to correctly detect foreground ground truth blobs, regardless of their specific label.

◆ _BlobDetectionBestFScoreThreshold__getter()

float eVision.EasyDeepLearning.ESupervisedSegmenterMetrics._BlobDetectionBestFScoreThreshold__getter ( self )

protected

\cond

Classification threshold that achieves the best F1-Score for blob detection.

See eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.BlobDetectionBestFScore for the corresponding F1-Score.

See eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics.GetFScore for a definition of the F1-Score. Blob detection is the ability of the segmenter to correctly detect foreground ground truth blobs, regardless of their specific label.

◆ _BlobDetectionFScore__getter()

float eVision.EasyDeepLearning.ESupervisedSegmenterMetrics._BlobDetectionFScore__getter ( self )

protected

\cond

F1-Score for defective blob detection given the current classification threshold (see eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics.ClassificationThreshold).

See eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.BlobDetectionBestFScore for the corresponding F1-Score.

See eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics.GetFScore for a definition of the F1-Score. Blob detection is the ability of the segmenter to correctly detect foreground ground truth blobs, regardless of their specific label.

◆ _BlobDetectionPrecision__getter()

float eVision.EasyDeepLearning.ESupervisedSegmenterMetrics._BlobDetectionPrecision__getter ( self )

protected

\cond

Precision for blob detection given the current classification threshold (see eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics.ClassificationThreshold).

The precision is the proportion of detected defective blobs that match ground truth defective blobs.

Blob detection is the ability of the segmenter to correctly detect foreground ground truth blobs, regardless of their specific label.

◆ _BlobDetectionRecall__getter()

float eVision.EasyDeepLearning.ESupervisedSegmenterMetrics._BlobDetectionRecall__getter ( self )

protected

\cond

Recall for blob detection given the current classification threshold (see eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics.ClassificationThreshold).

The recall is the proportion of ground truth defective blobs that are matched to predicted defective blobs.

Blob detection is the ability of the segmenter to correctly detect foreground ground truth blobs, regardless of their specific label.

◆ _Error__getter()

float eVision.EasyDeepLearning.ESupervisedSegmenterMetrics._Error__getter ( self )

protected

\cond

Error.

Remarks:

The error is also called the cross-entropy loss.

◆ _NumLabels__getter()

int eVision.EasyDeepLearning.ESupervisedSegmenterMetrics._NumLabels__getter ( self )

protected

\cond

Number of labels recognized by the segmenter that produced the results aggregated in the metrics.

◆ _PixelAccuracy__getter()

float eVision.EasyDeepLearning.ESupervisedSegmenterMetrics._PixelAccuracy__getter ( self )

protected

\cond

Accuracy.

The accuracy is the ratio of the number of correctly classified pixels to the total number of pixels.

Remarks:

The accuracy will be skewed towards the most represented labels in the dataset. For example, if your dataset has 1% of defective pixels, the accuracy will mostly reflect the results on the 99% of good pixels and will say nothing about the capability of the model to detect the defect labels.

◆ _WeightedError__getter()

float eVision.EasyDeepLearning.ESupervisedSegmenterMetrics._WeightedError__getter ( self )

protected

\cond

Weighted error.

The weighted error is the weighted average of each label error (eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetLabelError) with respect to the dataset label weights.

Remarks:

The error is also called the cross-entropy loss.

◆ _WeightedIntersectionOverUnion__getter()

float eVision.EasyDeepLearning.ESupervisedSegmenterMetrics._WeightedIntersectionOverUnion__getter ( self )

protected

\cond

Weighted Intersection over Union (IoU).

The weighted intersection over union is the weighted averaged of each label intersection over union (see eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetIntersectionOverUnion) with respect to the dataset label weights.

◆ _WeightedPixelAccuracy__getter()

float eVision.EasyDeepLearning.ESupervisedSegmenterMetrics._WeightedPixelAccuracy__getter ( self )

protected

\cond

Weighted accuracy.

The weighted accuracy is the weighted average of each label accuracy (eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetPixelLabelAccuracy) with respect to the dataset label weights.

◆ GetGroundtruthBlobConfusion() [1/2]

int eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetGroundtruthBlobConfusion	(		self,
		int	groundtruthLabelIndex,
		int	predictedLabelIndex )

Number of ground truth blobs of the given ground truth label that best match with blobs of the given predicted label.

See eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetPredictedBlobConfusion for the number of predicted blobs that match to these ground truth blobs.

Remarks:

A ground truth blob is matched to the subset of predicted blobs from the same predicted label that has the best intersection over union with the ground truth blob. Otherwise, the ground truth blob is matched to background.

Parameters

groundtruthLabelIndex	Ground truth label index
predictedLabelIndex	Predicted label index

◆ GetGroundtruthBlobConfusion() [2/2]

int eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetGroundtruthBlobConfusion	(		self,
		str	groundtruthLabel,
		str	predictedLabel )

Number of ground truth blobs of the given ground truth label that best match with blobs of the given predicted label.

See eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetPredictedBlobConfusion for the number of predicted blobs that match to these ground truth blobs.

Remarks:

A ground truth blob is matched to the subset of predicted blobs from the same predicted label that has the best intersection over union with the ground truth blob. Otherwise, the ground truth blob is matched to background.

Parameters

groundtruthLabel	Ground truth label
predictedLabel	Predicted label

◆ GetIntersectionOverUnion()

float eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetIntersectionOverUnion	(		self,
		str	label )

Intersection over union.

The intersection over union if the ratio between the number of correctly classified pixels from the given label to the total number of pixels that belongs to that label or are predicted as being from that label.

Assuming that the given label is the positive class, the intersection over union is expressed as IOU = TP / (TP + FP + FN) where TP is the number of true positive, FP the number of false positive (pixels that belongs to label but are not predicted as label) and FN the number of false negative (pixels predicted as label but that do not belong to label).

Remarks:

The intersection over union is also called the Jaccard index.

Parameters

label Label

◆ GetLabel()

str eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetLabel	(		self,
		int	labelIdx )

Label recognized by the segmenter that produced the results aggregated in the metrics.

Parameters

labelIdx Index of the label between 0 and eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.NumLabels - 1.

◆ GetLabelError()

float eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetLabelError	(		self,
		str	label )

Label error.

Parameters

label Label

◆ GetNormalizedPixelConfusion()

float eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetNormalizedPixelConfusion	(		self,
		str	trueLabel,
		str	predictedLabel )

Pixel-wise normalized confusion between the given true and predicted labels.

The normalized confusion is the ratio of the number of pixels belonging to the 'trueLabel' that are classified as 'predictedLabel' to the total number of pixels belonging to the 'trueLabel'.

Parameters

trueLabel	True label
predictedLabel	Predicted label

◆ GetPixelConfusion()

int eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetPixelConfusion	(		self,
		str	trueLabel,
		str	predictedLabel )

Pixel-wise confusion between the given true and predicted labels.

The confusion is the number of pixels belonging to the 'trueLabel' that are classified as 'predictedLabel'.

Parameters

trueLabel	True label
predictedLabel	Predicted label

◆ GetPixelLabelAccuracy()

float eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetPixelLabelAccuracy	(		self,
		str	label )

Pixel label accuracy.

The label accuracy is the ratio of the number of correctly classified pixels of the given class to the total number of ground truth pixels from that class. If there are no ground truth pixels from the requested label, the method returns '-1'.

Parameters

label Label

◆ GetPredictedBlobConfusion() [1/2]

int eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetPredictedBlobConfusion	(		self,
		int	groundtruthLabelIndex,
		int	predictedLabelIndex )

Number of predicted blobs of the given predicted label that match to blobs of the given ground truth label.

See eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetGroundtruthBlobConfusion for the number of ground truth blobs that match to these predicted blobs.

Remarks:

A predicted blob may be counted several times in the confusion matrix for predicted blobs:

Once if the more than 50% of the blob intersects ground truth background.
Once for each non-background label the predicted blob has an intersection with.

Parameters

groundtruthLabelIndex	Ground truth label index
predictedLabelIndex	Predicted label index

◆ GetPredictedBlobConfusion() [2/2]

int eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetPredictedBlobConfusion	(		self,
		str	groundtruthLabel,
		str	predictedLabel )

Number of predicted blobs of the given predicted label that match to blobs of the given ground truth label.

See eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetGroundtruthBlobConfusion for the number of ground truth blobs that match to these predicted blobs.

Remarks:

A predicted blob may be counted several times in the confusion matrix for predicted blobs:

Once if the more than 50% of the blob intersects ground truth background.
Once for each non-background label the predicted blob has an intersection with.

Parameters

groundtruthLabel	Ground truth label
predictedLabel	Predicted label

◆ IsValid()

bool eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.IsValid ( self )

Indicates whether the object contains at least one result.

◆ Load() [1/2]

None eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.Load	(		self,
		eVision.ESerializer \| None	serializer )

Loads an supervised segmentation metric.

The given eVision.ESerializer must have been created for reading.

Parameters

serializer The serializer.

Reimplemented from eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics.

◆ Load() [2/2]

None eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.Load	(		self,
		str	path )

Loads an supervised segmentation metric.

The given eVision.ESerializer must have been created for reading.

Parameters

path	The file path.

Reimplemented from eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics.

◆ Save() [1/2]

None eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.Save	(		self,
		eVision.ESerializer \| None	serializer )

Saves an supervised segmentation metric.

The given eVision.ESerializer must have been created for writing.

Parameters

serializer The serializer.

Reimplemented from eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics.

◆ Save() [2/2]

None eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.Save	(		self,
		str	path )

Saves an supervised segmentation metric.

The given eVision.ESerializer must have been created for writing.

Parameters

path	The file path.

Reimplemented from eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics.

Property Documentation

◆ BalancedError

eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.BalancedError = property(_BalancedError__getter, None)

static

\endcond

Balanced error.

The balanced error is the weighted error (eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.WeightedError) where each label is given an equal weight.

Remarks:

The error is also called the cross-entropy loss.

◆ BalancedIntersectionOverUnion

eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.BalancedIntersectionOverUnion = property(_BalancedIntersectionOverUnion__getter, None)

static

\endcond

Balanced Intersection over Union (IoU).

The balanced intersection over union is the weighted intersection over union (eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.WeightedIntersectionOverUnion) where each label is given equal weight.

◆ BalancedPixelAccuracy

eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.BalancedPixelAccuracy = property(_BalancedPixelAccuracy__getter, None)

static

\endcond

Balanced accuracy.

The balanced accuracy is the weighted accuracy (eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.WeightedPixelAccuracy) where each label is given an equal weight.

◆ BlobDetectionAveragePrecision

eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.BlobDetectionAveragePrecision = property(_BlobDetectionAveragePrecision__getter, None)

static

\endcond

Average precision for blob detection.

Blob detection is the ability of the segmenter to correctly detect foreground ground truth blobs, regardless of their specific label. The average precision for blob detection is the average of the precision (see eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.BlobDetectionPrecision) over all the possible classification threshold values. As such, the average precision does not depend on a specific classification threshold.

◆ BlobDetectionBestFScore

eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.BlobDetectionBestFScore = property(_BlobDetectionBestFScore__getter, None)

static

\endcond

Best F1-Score of the segmenter for blob detection.

See eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.BlobDetectionBestFScoreThreshold for the corresponding threshold.

See eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics.GetFScore for a definition of the F1-Score. Blob detection is the ability of the segmenter to correctly detect foreground ground truth blobs, regardless of their specific label.

◆ BlobDetectionBestFScoreThreshold

eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.BlobDetectionBestFScoreThreshold = property(_BlobDetectionBestFScoreThreshold__getter, None)

static

\endcond

Classification threshold that achieves the best F1-Score for blob detection.

See eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.BlobDetectionBestFScore for the corresponding F1-Score.

See eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics.GetFScore for a definition of the F1-Score. Blob detection is the ability of the segmenter to correctly detect foreground ground truth blobs, regardless of their specific label.

◆ BlobDetectionFScore

eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.BlobDetectionFScore = property(_BlobDetectionFScore__getter, None)

static

\endcond

F1-Score for defective blob detection given the current classification threshold (see eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics.ClassificationThreshold).

See eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.BlobDetectionBestFScore for the corresponding F1-Score.

See eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics.GetFScore for a definition of the F1-Score. Blob detection is the ability of the segmenter to correctly detect foreground ground truth blobs, regardless of their specific label.

◆ BlobDetectionPrecision

eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.BlobDetectionPrecision = property(_BlobDetectionPrecision__getter, None)

static

\endcond

Precision for blob detection given the current classification threshold (see eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics.ClassificationThreshold).

The precision is the proportion of detected defective blobs that match ground truth defective blobs.

Blob detection is the ability of the segmenter to correctly detect foreground ground truth blobs, regardless of their specific label.

◆ BlobDetectionRecall

eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.BlobDetectionRecall = property(_BlobDetectionRecall__getter, None)

static

\endcond

Recall for blob detection given the current classification threshold (see eVision.EasyDeepLearning.EDeepLearningDefectDetectionMetrics.ClassificationThreshold).

The recall is the proportion of ground truth defective blobs that are matched to predicted defective blobs.

Blob detection is the ability of the segmenter to correctly detect foreground ground truth blobs, regardless of their specific label.

◆ Error

eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.Error = property(_Error__getter, None)

static

\endcond

Error.

Remarks:

The error is also called the cross-entropy loss.

◆ NumLabels

eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.NumLabels = property(_NumLabels__getter, None)

static

\endcond

Number of labels recognized by the segmenter that produced the results aggregated in the metrics.

◆ PixelAccuracy

eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.PixelAccuracy = property(_PixelAccuracy__getter, None)

static

\endcond

Accuracy.

The accuracy is the ratio of the number of correctly classified pixels to the total number of pixels.

Remarks:

The accuracy will be skewed towards the most represented labels in the dataset. For example, if your dataset has 1% of defective pixels, the accuracy will mostly reflect the results on the 99% of good pixels and will say nothing about the capability of the model to detect the defect labels.

◆ WeightedError

eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.WeightedError = property(_WeightedError__getter, None)

static

\endcond

Weighted error.

The weighted error is the weighted average of each label error (eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetLabelError) with respect to the dataset label weights.

Remarks:

The error is also called the cross-entropy loss.

◆ WeightedIntersectionOverUnion

eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.WeightedIntersectionOverUnion = property(_WeightedIntersectionOverUnion__getter, None)

static

\endcond

Weighted Intersection over Union (IoU).

The weighted intersection over union is the weighted averaged of each label intersection over union (see eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetIntersectionOverUnion) with respect to the dataset label weights.

◆ WeightedPixelAccuracy

eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.WeightedPixelAccuracy = property(_WeightedPixelAccuracy__getter, None)

static

\endcond

Weighted accuracy.

The weighted accuracy is the weighted average of each label accuracy (eVision.EasyDeepLearning.ESupervisedSegmenterMetrics.GetPixelLabelAccuracy) with respect to the dataset label weights.

Public Member Functions

Protected Member Functions

Properties

Detailed Description

Constructor & Destructor Documentation

◆ __init__() [1/2]

◆ __init__() [2/2]

Member Function Documentation

◆ _BalancedError__getter()

◆ _BalancedIntersectionOverUnion__getter()

◆ _BalancedPixelAccuracy__getter()

◆ _BlobDetectionAveragePrecision__getter()

◆ _BlobDetectionBestFScore__getter()

◆ _BlobDetectionBestFScoreThreshold__getter()

◆ _BlobDetectionFScore__getter()

◆ _BlobDetectionPrecision__getter()

◆ _BlobDetectionRecall__getter()

◆ _Error__getter()

◆ _NumLabels__getter()

◆ _PixelAccuracy__getter()

◆ _WeightedError__getter()

◆ _WeightedIntersectionOverUnion__getter()

◆ _WeightedPixelAccuracy__getter()

◆ GetGroundtruthBlobConfusion() [1/2]

◆ GetGroundtruthBlobConfusion() [2/2]

◆ GetIntersectionOverUnion()

◆ GetLabel()

◆ GetLabelError()

◆ GetNormalizedPixelConfusion()

◆ GetPixelConfusion()

◆ GetPixelLabelAccuracy()

◆ GetPredictedBlobConfusion() [1/2]

◆ GetPredictedBlobConfusion() [2/2]

◆ IsValid()

◆ Load() [1/2]

◆ Load() [2/2]

◆ Save() [1/2]

◆ Save() [2/2]

Property Documentation

◆ BalancedError

◆ BalancedIntersectionOverUnion

◆ BalancedPixelAccuracy

◆ BlobDetectionAveragePrecision

◆ BlobDetectionBestFScore

◆ BlobDetectionBestFScoreThreshold

◆ BlobDetectionFScore

◆ BlobDetectionPrecision

◆ BlobDetectionRecall

◆ Error

◆ NumLabels

◆ PixelAccuracy

◆ WeightedError

◆ WeightedIntersectionOverUnion

◆ WeightedPixelAccuracy

◆ init() [1/2]

◆ init() [2/2]