eVision.EasyDeepLearning.EUnsupervisedSegmenter Class Reference

Unsupervised segmentation tool. More...

Inheritance diagram for eVision.EasyDeepLearning.EUnsupervisedSegmenter:

Collaboration diagram for eVision.EasyDeepLearning.EUnsupervisedSegmenter:

Public Member Functions
None	__init__ (self)
	Constructs a eVision.EasyDeepLearning.EUnsupervisedSegmenter object.
None	__init__ (self, eVision.EasyDeepLearning.EUnsupervisedSegmenter other)
	Copy constructor of a eVision.EasyDeepLearning.EUnsupervisedSegmenter object.
None	SerializeSettings (self, eVision.ESerializer|None serializer)
	Serializes the settings of the unsupervised segmenter.
eVision.EasyDeepLearning.EUnsupervisedSegmenterMetrics	GetTrainingMetrics (self, int iteration)
	Training metrics at the given iteration.
eVision.EasyDeepLearning.EUnsupervisedSegmenterMetrics	GetValidationMetrics (self, int iteration)
	Validation metrics at the given iteration.
None	Train (self, eVision.EasyDeepLearning.EDataset dataset)
	Trains the eVision.EasyDeepLearning.EUnsupervisedSegmenter with the given dataset for the specified number of iterations.
None	Train (self, eVision.EasyDeepLearning.EDataset dataset, eVision.EasyDeepLearning.EDataAugmentation dataAugmentation)
	Trains the eVision.EasyDeepLearning.EUnsupervisedSegmenter with the given dataset for the specified number of iterations.
None	Train (self, eVision.EasyDeepLearning.EDataset trainingDataset, eVision.EasyDeepLearning.EDataset validationDataset)
	Trains the eVision.EasyDeepLearning.EUnsupervisedSegmenter with the given dataset for the specified number of iterations.
None	Train (self, eVision.EasyDeepLearning.EDataset trainingDataset, eVision.EasyDeepLearning.EDataset validationDataset, eVision.EasyDeepLearning.EDataAugmentation dataAugmentation)
	Trains the eVision.EasyDeepLearning.EUnsupervisedSegmenter with the given dataset for the specified number of iterations.
None	Train (self, eVision.EasyDeepLearning.EDataset dataset, int iterations)
	Trains the eVision.EasyDeepLearning.EUnsupervisedSegmenter with the given dataset for the specified number of iterations.
None	Train (self, eVision.EasyDeepLearning.EDataset dataset, eVision.EasyDeepLearning.EDataAugmentation dataAugmentation, int iterations)
	Trains the eVision.EasyDeepLearning.EUnsupervisedSegmenter with the given dataset for the specified number of iterations.
None	Train (self, eVision.EasyDeepLearning.EDataset trainingDataset, eVision.EasyDeepLearning.EDataset validationDataset, int iterations)
	Trains the eVision.EasyDeepLearning.EUnsupervisedSegmenter with the given dataset for the specified number of iterations.
None	Train (self, eVision.EasyDeepLearning.EDataset trainingDataset, eVision.EasyDeepLearning.EDataset validationDataset, eVision.EasyDeepLearning.EDataAugmentation dataAugmentation, int iterations)
	Trains the eVision.EasyDeepLearning.EUnsupervisedSegmenter with the given dataset for the specified number of iterations.
eVision.EasyDeepLearning.EUnsupervisedSegmenterResult	Apply (self, eVision.EBaseROI img)
	Applies the unsupervised segmenter on the given image and its mask region.
eVision.EasyDeepLearning.EUnsupervisedSegmenterResult	Apply (self, eVision.EBaseROI img, eVision.ERegion mask)
	Applies the unsupervised segmenter on the given image and its mask region.
None	Apply (self, list[eVision._eVision.EBaseROI|None] imgs, list[eVision._eVision.EasyDeepLearning.EUnsupervisedSegmenterResult|None] results)
	Applies the unsupervised segmenter on the given image and its mask region.
None	Apply (self, list[eVision._eVision.EBaseROI|None] imgs, list[eVision._eVision.ERegion|None] masks, list[eVision._eVision.EasyDeepLearning.EUnsupervisedSegmenterResult|None] results)
	Applies the unsupervised segmenter on the given image and its mask region.
eVision.EasyDeepLearning.EUnsupervisedSegmenterMetrics	Evaluate (self, eVision.EasyDeepLearning.EDataset dataset)
	Evaluates the dataset.
int	GetNumPatchesForImage (self, int imageWidth, int imageHeight)
	Number of patches that will be extracted from an input image to perform inference.
int	Channels (self)
	Number of channels of input images.
eVision.EasyDeepLearning.EDeepLearningToolType	ToolType (self)
	Type of the deep learning tool.
eVision.EasyDeepLearning.EUnsupervisedSegmenterCapacity	Capacity (self)
	Capacity of the eVision.EasyDeepLearning.EUnsupervisedSegmenter.
	Capacity (self, eVision.EasyDeepLearning.EUnsupervisedSegmenterCapacity value)
	Capacity of the eVision.EasyDeepLearning.EUnsupervisedSegmenter.
bool	ForceGrayscale (self)
	Forces the eVision.EasyDeepLearning.EUnsupervisedSegmenter to convert all images to grayscale (default: true).
	ForceGrayscale (self, bool value)
	Forces the eVision.EasyDeepLearning.EUnsupervisedSegmenter to convert all images to grayscale (default: true).
int	PatchSize (self)
	Patch size (width and height of the patches processed by the neural network).
	PatchSize (self, int value)
	Patch size (width and height of the patches processed by the neural network).
float	Scale (self)
	Down-scaling applied to images before processing them.
	Scale (self, float value)
	Down-scaling applied to images before processing them.
float	ClassificationThreshold (self)
	Classification threshold for the score of an image (See eVision.EasyDeepLearning.EUnsupervisedSegmenterResult.ClassificationScore).
	ClassificationThreshold (self, float value)
	Classification threshold for the score of an image (See eVision.EasyDeepLearning.EUnsupervisedSegmenterResult.ClassificationScore).
float	SamplingDensity (self)
	Sampling density (default value: 2.0, its value must be equal or greater than 1).
	SamplingDensity (self, float value)
	Sampling density (default value: 2.0, its value must be equal or greater than 1).
str	GoodLabel (self)
	Name of the good label in the dataset that is used for training (default value: empty).
	GoodLabel (self, str value)
	Name of the good label in the dataset that is used for training (default value: empty).
bool	ScaleDisabledAtInference (self)
	Whether to apply the scale parameter at inference or not.
	ScaleDisabledAtInference (self, bool value)
	Whether to apply the scale parameter at inference or not.
float	InferenceScale (self)
	The effective scale applied when performing inference.
Public Member Functions inherited from eVision.EasyDeepLearning.EDeepLearningTool
None	Load (self, str filePath)
	Loads a deep learning tool.
None	Load (self, eVision.ESerializer|None serializer)
	Loads a deep learning tool.
eVision.EasyDeepLearning.EDeepLearningTool|None	Clone (self)
	Clones this tool.
None	Save (self, str filePath, bool includeTrainingModel)
	Saves the settings, inference and training model of the deep learning tool.
None	Save (self, str filePath)
	Saves the settings, inference and training model of the deep learning tool.
None	Save (self, eVision.ESerializer|None serializer, bool includeTrainingModel)
	Saves the settings, inference and training model of the deep learning tool.
None	Save (self, eVision.ESerializer|None serializer)
	Saves the settings, inference and training model of the deep learning tool.
None	SerializeInferenceModel (self, eVision.ESerializer|None serializer)
	Serializes the inference model of the tool.
None	SerializeTrainingModel (self, eVision.ESerializer|None serializer)
	Serializes the training model of the tool.
None	SaveSettings (self, str filePath)
	Saves the settings of the tool to a file.
None	SaveInferenceModel (self, str filePath)
	Saves the tool's inference model to a file.
None	SaveTrainingModel (self, str filePath)
	Saves the tool's training model to a file.
None	LoadSettings (self, str filePath)
	Loads the settings of a tool.
None	LoadInferenceModel (self, str filePath)
	Loads the tool's inference model.
None	LoadTrainingModel (self, str filePath)
	Loads the tool's training model.
None	UnloadInferenceModel (self)
	Unloads the inference model.
None	UnloadTrainingModel (self)
	Unloads the training model.
None	UnloadModels (self)
	Unloads the inference and training model.
bool	HasInferenceModel (self)
	Whether the inference model is loaded or created when the tool is not yet trained.
bool	HasTrainingModel (self)
	Whether the training model is loaded or created when the tool is not yet trained.
bool	IsTrained (self)
	Tells whether the deep learning tool has been trained.
str	GetAvailableDeviceName (self, int i)
	Name of the i-th device of the current engine.
eVision.EasyDeepLearning.EDeepLearningDevice	GetAvailableDevice (self, int i)
	i-th eVision.EasyDeepLearning.EDeepLearningDevice of the current engine.
None	SetActiveDevice (self, eVision.EasyDeepLearning.EDeepLearningDevice device)
	Sets the device with which to run or train the model.
eVision.EasyDeepLearning.EDeepLearningDevice	GetActiveDevice (self, int i)
	Gets the active device.
eVision.EasyDeepLearning.EDeepLearningDevice	GetActiveDevice (self)
	Gets the active device.
bool	HasEngineSetupData (self)
	Whether the inference model has engine setup data.
None	ClearEngineSetupData (self)
	Clears the engine data saved in the inference model.
bool	IsTraining (self)
	Indicates whether the object is currently training.
int	WaitForTrainingCompletion (self, int timeout)
	Waits until the training is complete or the timeout is expired.
int	WaitForTrainingCompletion (self)
	Waits until the training is complete or the timeout is expired.
int	WaitForIterationCompletion (self)
	Waits until an iteration is complete.
int	StopTraining (self, bool wait)
	Stops training and returns the last completed iteration.
None	SetLearningRate (self, eVision.EasyDeepLearning.EDeepLearningLearningRateType type, list[float] parameters)
	Specifies how the value of the learning rate and how it evolves during training.
None	InitializeInference (self, int width, int height, eVision.EImageType imageType, int numImages)
	Initializes the neural network for inference (i.e.
None	InitializeInference (self, eVision.EBaseROI img)
	Initializes the neural network for inference (i.e.
None	InitializeInference (self, list[eVision._eVision.EBaseROI|None] imgList)
	Initializes the neural network for inference (i.e.
str	GetLabel (self, int index)
	Gets the label from its index.
None	SetLabel (self, int index, str label)
	Changes a label predicted by the tool.
None	SetLabelWeight (self, int index, float weight)
	Sets the label weight.
float	GetLabelWeight (self, int index)
	Gets the label weight.
None	SetLabelColor (self, int index, eVision.ERGBColor c)
	Changes a label color.
eVision.ERGBColor	GetLabelColor (self, int index)
	Label color.
None	SetLabelOpacity (self, int index, int opacity)
	Changes a label opacity (between 0 for fully transparent and 255 for opaque color).
int	GetLabelOpacity (self, int index)
	Gets the label opacity (between 0 for fully transparent and 255 for opaque color).
int	GetOptimalNumImagesForBatchSize (self, eVision.EBaseROI sampleImage)
	Optimal number of images to process together for inference given the batch size and eVision.EasyDeepLearning.EDeepLearningTool.GetNumPatchesForImage.
int	GetOptimalNumImagesForBatchSize (self, int imageWidth, int imageHeight)
	Optimal number of images to process together for inference given the batch size and eVision.EasyDeepLearning.EDeepLearningTool.GetNumPatchesForImage.
None	Clean (self)
	Reset the engine and clean CPU and GPU memory allocated for running the model.
bool	EnableGPU (self)
	Enable the use of a GPU for training and inference of the deep learning tool.
	EnableGPU (self, bool value)
	Enable the use of a GPU for training and inference of the deep learning tool.
list[int]	GPUIndexes (self)
	Indexes of the GPUs to use for computations.
	GPUIndexes (self, list[int] value)
	Indexes of the GPUs to use for computations.
int	NumGPUs (self)
	Number of detected NVIDIA GPU.
str	Engine (self)
	Sets the engine with which to execute or train the neural network.
	Engine (self, str value)
	Sets the engine with which to execute or train the neural network.
eVision.EasyDeepLearning.EDeepLearningInferencePrecision	InferencePrecision (self)
	Inference precision.
	InferencePrecision (self, eVision.EasyDeepLearning.EDeepLearningInferencePrecision value)
	Inference precision.
int	NumAvailableDevices (self)
	Available devices for the current eVision.EasyDeepLearning.EDeepLearningTool.Engine.
list[eVision._eVision.EasyDeepLearning.EDeepLearningDevice]	AvailableDevices (self)
	Get all available device for the current engine.
int	NumActiveDevices (self)
	Gets the number of devices with which to run or train the model.
list[eVision._eVision.EasyDeepLearning.EDeepLearningDevice]	ActiveDevices (self)
	Devices with which to run or train the model.
	ActiveDevices (self, list[eVision._eVision.EasyDeepLearning.EDeepLearningDevice] value)
	Devices with which to run or train the model.
eVision.EasyDeepLearning.EDeepLearningExecutionSettings	ExecutionSettings (self)
	Execution settings of the tool (engine, device(s), inference precision, batch size settings).
	ExecutionSettings (self, eVision.EasyDeepLearning.EDeepLearningExecutionSettings value)
	Execution settings of the tool (engine, device(s), inference precision, batch size settings).
eVision.EasyDeepLearning.EDeepLearningDevice	EngineSetupDataDevice (self)
	Device corresponding to the engine data contained in the tool.
eVision.EasyDeepLearning.EDeepLearningInferencePrecision	EngineSetupDataInferencePrecision (self)
	Inference precision corresponding to the engine data contained in the tool.
int	EngineSetupDataBatchSize (self)
	Batch size corresponding to the engine data contained in the tool.
bool	OptimizeBatchSize (self)
	Indicates whether to optimize the batch size (see eVision.EasyDeepLearning.EDeepLearningTool.BatchSize) to maximize the training and inference speed according to the engine, the device and the available memory.
	OptimizeBatchSize (self, bool value)
	Indicates whether to optimize the batch size (see eVision.EasyDeepLearning.EDeepLearningTool.BatchSize) to maximize the training and inference speed according to the engine, the device and the available memory.
int	BatchSizeForMaximumInferenceSpeed (self)
	Computes the batch size that will maximize the inference speed on NVIDIA GPU.
int	BatchSize (self)
	Batch size.
	BatchSize (self, int value)
	Batch size.
float	CurrentTrainingProgression (self)
	Current training progression as a percentage (between 0 and 1).
int	CurrentTrainingFinishedIterations (self)
	Number of iterations that are already finished in the current training.
int	CurrentTrainingNumIterations (self)
	Total number of training iterations that will be performed during the current training of the deep learning tool.
int	NumTrainedIterations (self)
	Number of iterations that were performed to train this deep learning tool.
int	BestIteration (self)
	Iteration at which the minimum validation error was reached.
int	ImageCacheSize (self)
	Size in byte of the image cache.
	ImageCacheSize (self, int value)
	Size in byte of the image cache.
str	Path (self)
	Path of the tool when loaded or saved using eVision.EasyDeepLearning.EDeepLearningTool.Save or eVision.EasyDeepLearning.EDeepLearningTool.Load.
str	InferenceModelPath (self)
	Path of the inference model when loaded or saved using eVision.EasyDeepLearning.EDeepLearningTool.SaveInferenceModel, eVision.EasyDeepLearning.EDeepLearningTool.LoadInferenceModel or eVision.EasyDeepLearning.EDeepLearningTool.SerializeInferenceModel.
str	TrainingModelPath (self)
	Path of the training model when loaded or saved using eVision.EasyDeepLearning.EDeepLearningTool.SaveTrainingModel, eVision.EasyDeepLearning.EDeepLearningTool.LoadTrainingModel or eVision.EasyDeepLearning.EDeepLearningTool.SerializeTrainingModel.
str	SettingsPath (self)
	Path of the settings when loaded or saved using eVision.EasyDeepLearning.EDeepLearningTool.SaveSettings, eVision.EasyDeepLearning.EDeepLearningTool.LoadSettings or eVision.EasyDeepLearning.EDeepLearningTool.SerializeSettings.
int	AutoSavePeriod (self)
	Auto-save period.
	AutoSavePeriod (self, int value)
	Auto-save period.
eVision.EasyDeepLearning.EDeepLearningLearningRateType	LearningRateType (self)
	Learning rate type.
list[float]	LearningRateParameters (self)
	Parameters of the learning rate.
bool	EnableDeterministicTraining (self)
	Whether to use deterministic training algorithm that allows to repeatable training results.
	EnableDeterministicTraining (self, bool value)
	Whether to use deterministic training algorithm that allows to repeatable training results.
int	DeterministicTrainingRandomSeed (self)
	Random seed to use when eVision.EasyDeepLearning.EDeepLearningTool.EnableDeterministicTraining is set to true.
	DeterministicTrainingRandomSeed (self, int value)
	Random seed to use when eVision.EasyDeepLearning.EDeepLearningTool.EnableDeterministicTraining is set to true.
int	NumLabels (self)
	Number of labels of this tool.

Additional Inherited Members
Static Public Member Functions inherited from eVision.EasyDeepLearning.EDeepLearningTool
eVision.EasyDeepLearning.EDeepLearningTool|None	Create (str path)
	Factory method to open a deep learning tool.
eVision.EasyDeepLearning.EDeepLearningTool|None	Create (eVision.ESerializer|None serializer)
	Factory method to open a deep learning tool.
eVision.EasyDeepLearning.EDeepLearningTool|None	Create (eVision.EasyDeepLearning.EDeepLearningToolType type)
	Factory method to open a deep learning tool.
int	GetNumAvailableEngines ()
	Number of available engines.
str	GetAvailableEngineName (int id)
	Gets the name of the engine given its id.
list[str]	GetAvailableEngines ()
	Available engines.
list[eVision._eVision.EasyDeepLearning.EDeepLearningDevice]	GetAvailableDevicesForEngine (str engine)
	Available devices for the specified engine.
Protected Member Functions inherited from eVision.EasyDeepLearning.EDeepLearningTool
	_ActiveDeviceByName__setter (self, str value)
	Sets the device with which to run the model using its name.
	_ActiveDeviceById__setter (self, int value)
	Sets the device with which to run the model using its index in the list returned by eVision.EasyDeepLearning.EDeepLearningTool.AvailableDevices.
	_ActiveDevicesById__setter (self, list[int] value)
	Sets the devices with which to run the model using their index in the list returned by eVision.EasyDeepLearning.EDeepLearningTool.AvailableDevices.
Properties inherited from eVision.EasyDeepLearning.EDeepLearningTool
	ActiveDeviceByName
	ActiveDeviceById
	ActiveDevicesById

Detailed Description

Unsupervised segmentation tool.

The unsupervised segmentation tool learns a model of what is a good product and it can be used for classifying whether an image is from a good or defective product and for segmenting the defects within the image.

To learn a model of what is a good product, the tool is trained by considering only the images from the good label (eVision.EasyDeepLearning.EUnsupervisedSegmenter.GoodLabel). The tool labels (eVision.EasyDeepLearning.EDeepLearningTool.GetLabel) will always be empty.

The tool can work with images of any resolution higher than eVision.EasyDeepLearning.EUnsupervisedSegmenter.PatchSize by merging the results obtained by applying the deep neural network using a sliding window algorithm. The overlapping between the sliding windows is controlled by eVision.EasyDeepLearning.EUnsupervisedSegmenter.SamplingDensity.

The unsupervised segmenter offers a tradeoff between a high good detection rate and a high bad detection rate through a classification threshold that can be configured after training (eVision.EasyDeepLearning.EUnsupervisedSegmenter.ClassificationThreshold).

Remarks:

License(s): EasySegment or EasySegmentInference

Constructor & Destructor Documentation

__init__() [1/2]

None eVision.EasyDeepLearning.EUnsupervisedSegmenter.__init__

(

self

)

Constructs a eVision.EasyDeepLearning.EUnsupervisedSegmenter object.

__init__() [2/2]

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

Copy constructor of a eVision.EasyDeepLearning.EUnsupervisedSegmenter object.

Parameters

other

Reference to the eVision.EasyDeepLearning.EUnsupervisedSegmenter object that should be copied

Member Function Documentation

Apply() [1/4]

eVision.EasyDeepLearning.EUnsupervisedSegmenterResult eVision.EasyDeepLearning.EUnsupervisedSegmenter.Apply	(		self,
		eVision.EBaseROI	img )

Applies the unsupervised segmenter on the given image and its mask region.

We recommend to use pointer-based versions of this method to avoid unnecessary image and result copies.

Parameters

img	Image to classify and segment defects in.

Apply() [2/4]

eVision.EasyDeepLearning.EUnsupervisedSegmenterResult eVision.EasyDeepLearning.EUnsupervisedSegmenter.Apply	(		self,
		eVision.EBaseROI	img,
		eVision.ERegion	mask )

Applies the unsupervised segmenter on the given image and its mask region.

We recommend to use pointer-based versions of this method to avoid unnecessary image and result copies.

Parameters

img	Image to classify and segment defects in.
mask	Mask region of the image.

Apply() [3/4]

None eVision.EasyDeepLearning.EUnsupervisedSegmenter.Apply	(		self,
		list[eVision._eVision.EBaseROI | None]	imgs,
		list[eVision._eVision.EasyDeepLearning.EUnsupervisedSegmenterResult | None]	results )

Applies the unsupervised segmenter on the given image and its mask region.

We recommend to use pointer-based versions of this method to avoid unnecessary image and result copies.

Parameters

imgs	Vector of image to classify and segment defects in.
results	Vector of eVision.EasyDeepLearning.EUnsupervisedSegmenterResult to put the segmented defects.

Apply() [4/4]

None eVision.EasyDeepLearning.EUnsupervisedSegmenter.Apply	(		self,
		list[eVision._eVision.EBaseROI | None]	imgs,
		list[eVision._eVision.ERegion | None]	masks,
		list[eVision._eVision.EasyDeepLearning.EUnsupervisedSegmenterResult | None]	results )

Applies the unsupervised segmenter on the given image and its mask region.

We recommend to use pointer-based versions of this method to avoid unnecessary image and result copies.

Parameters

imgs	Vector of image to classify and segment defects in.
masks	Vector of mask regions for the images.
results	Vector of eVision.EasyDeepLearning.EUnsupervisedSegmenterResult to put the segmented defects.

Capacity() [1/2]

eVision.EasyDeepLearning.EUnsupervisedSegmenterCapacity eVision.EasyDeepLearning.EUnsupervisedSegmenter.Capacity

(

self

)

Capacity of the eVision.EasyDeepLearning.EUnsupervisedSegmenter.

A higher capacity makes the unsupervised segmenter capable of learning more information at the cost of a slower processing speed.

Capacity() [2/2]

eVision.EasyDeepLearning.EUnsupervisedSegmenter.Capacity	(		self,
		eVision.EasyDeepLearning.EUnsupervisedSegmenterCapacity	value )

Capacity of the eVision.EasyDeepLearning.EUnsupervisedSegmenter.

A higher capacity makes the unsupervised segmenter capable of learning more information at the cost of a slower processing speed.

Channels()

int eVision.EasyDeepLearning.EUnsupervisedSegmenter.Channels

(

self

)

Number of channels of input images.

It can either be 1 for grayscale images or 3 for color images.

Remarks:

If the segmenter tool is not trained or the value was not explicitly set, its value will be 0.

ClassificationThreshold() [1/2]

float eVision.EasyDeepLearning.EUnsupervisedSegmenter.ClassificationThreshold

(

self

)

Classification threshold for the score of an image (See eVision.EasyDeepLearning.EUnsupervisedSegmenterResult.ClassificationScore).

A image with a score smaller or equal to the classification threshold will be classified as good.

Remarks:

The classification threshold is optimized during training to maximize the balanced accuracy of the unsupervised segmenter.

The classification threshold can be changed after training.

ClassificationThreshold() [2/2]

eVision.EasyDeepLearning.EUnsupervisedSegmenter.ClassificationThreshold	(		self,
		float	value )

Classification threshold for the score of an image (See eVision.EasyDeepLearning.EUnsupervisedSegmenterResult.ClassificationScore).

A image with a score smaller or equal to the classification threshold will be classified as good.

Remarks:

The classification threshold is optimized during training to maximize the balanced accuracy of the unsupervised segmenter.

The classification threshold can be changed after training.

Evaluate()

eVision.EasyDeepLearning.EUnsupervisedSegmenterMetrics eVision.EasyDeepLearning.EUnsupervisedSegmenter.Evaluate	(		self,
		eVision.EasyDeepLearning.EDataset	dataset )

Evaluates the dataset.

Remarks:

License(s): EasySegment

Parameters

dataset

Dataset to evaluate

ForceGrayscale() [1/2]

bool eVision.EasyDeepLearning.EUnsupervisedSegmenter.ForceGrayscale

(

self

)

Forces the eVision.EasyDeepLearning.EUnsupervisedSegmenter to convert all images to grayscale (default: true).

Setting this property to false will make the underlying neural network operates with the same number of channels as the images in the dataset used for training. Otherwise, color images will be converted to grayscale before using them.

ForceGrayscale() [2/2]

eVision.EasyDeepLearning.EUnsupervisedSegmenter.ForceGrayscale	(		self,
		bool	value )

Forces the eVision.EasyDeepLearning.EUnsupervisedSegmenter to convert all images to grayscale (default: true).

Setting this property to false will make the underlying neural network operates with the same number of channels as the images in the dataset used for training. Otherwise, color images will be converted to grayscale before using them.

GetNumPatchesForImage()

int eVision.EasyDeepLearning.EUnsupervisedSegmenter.GetNumPatchesForImage	(		self,
		int	imageWidth,
		int	imageHeight )

Number of patches that will be extracted from an input image to perform inference.

For EasyClassify and EasyLocate, this will always be equal to 1. For EasySegment, the number of patches will depend on the scale, patch size and sampling density parameters.

Parameters

imageWidth	Width of the image for which to get the number of patch
imageHeight	Height of the image for which to get the number of patch

Reimplemented from eVision.EasyDeepLearning.EDeepLearningTool.

GetTrainingMetrics()

eVision.EasyDeepLearning.EUnsupervisedSegmenterMetrics eVision.EasyDeepLearning.EUnsupervisedSegmenter.GetTrainingMetrics	(		self,
		int	iteration )

Training metrics at the given iteration.

Remarks:

At a given iteration, the metrics will always contain the error (see eVision.EasyDeepLearning.EUnsupervisedSegmenterMetrics.Error). Other metrics (scores, accuracy, etc.) will be available only at iterations where the validation error is a new minimum.

Parameters

iteration

Iteration at which to get the metrics

GetValidationMetrics()

eVision.EasyDeepLearning.EUnsupervisedSegmenterMetrics eVision.EasyDeepLearning.EUnsupervisedSegmenter.GetValidationMetrics	(		self,
		int	iteration )

Validation metrics at the given iteration.

Remarks:

At a given iteration, the metrics will always contain the error (see eVision.EasyDeepLearning.EUnsupervisedSegmenterMetrics.Error). Other metrics (scores, accuracy, etc.) will be available only at iterations where the validation error is a new minimum.

Parameters

iteration

Iteration at which to get the metrics

GoodLabel() [1/2]

str eVision.EasyDeepLearning.EUnsupervisedSegmenter.GoodLabel

(

self

)

Name of the good label in the dataset that is used for training (default value: empty).

GoodLabel() [2/2]

eVision.EasyDeepLearning.EUnsupervisedSegmenter.GoodLabel	(		self,
		str	value )

Name of the good label in the dataset that is used for training (default value: empty).

InferenceScale()

float eVision.EasyDeepLearning.EUnsupervisedSegmenter.InferenceScale

(

self

)

The effective scale applied when performing inference.

If eVision.EasyDeepLearning.EUnsupervisedSegmenter.ScaleDisabledAtInference is true, it is equal to 1.0. Otherwise, it is equal to eVision.EasyDeepLearning.EUnsupervisedSegmenter.Scale.

PatchSize() [1/2]

int eVision.EasyDeepLearning.EUnsupervisedSegmenter.PatchSize

(

self

)

Patch size (width and height of the patches processed by the neural network).

Remarks:

There are three supported patch size: 64x64, 128x128, and 256x256. By default, the patch size is 0 and it means that the patch size will be 128x128 if all images in the training and validation dataset have a higher resolution or the patch size will be 64x64.

PatchSize() [2/2]

eVision.EasyDeepLearning.EUnsupervisedSegmenter.PatchSize	(		self,
		int	value )

Patch size (width and height of the patches processed by the neural network).

Remarks:

There are three supported patch size: 64x64, 128x128, and 256x256. By default, the patch size is 0 and it means that the patch size will be 128x128 if all images in the training and validation dataset have a higher resolution or the patch size will be 64x64.

SamplingDensity() [1/2]

float eVision.EasyDeepLearning.EUnsupervisedSegmenter.SamplingDensity

(

self

)

Sampling density (default value: 2.0, its value must be equal or greater than 1).

The sampling density is the parameter of the sliding window algorithm used for processing a whole image using subwindows of size eVision.EasyDeepLearning.EUnsupervisedSegmenter.PatchSize. It indicates how much overlap there will be between the image patches: the stride between two consecutive patches is eVision.EasyDeepLearning.EUnsupervisedSegmenter.PatchSize / eVision.EasyDeepLearning.EUnsupervisedSegmenter.SamplingDensity.

SamplingDensity() [2/2]

eVision.EasyDeepLearning.EUnsupervisedSegmenter.SamplingDensity	(		self,
		float	value )

Sampling density (default value: 2.0, its value must be equal or greater than 1).

The sampling density is the parameter of the sliding window algorithm used for processing a whole image using subwindows of size eVision.EasyDeepLearning.EUnsupervisedSegmenter.PatchSize. It indicates how much overlap there will be between the image patches: the stride between two consecutive patches is eVision.EasyDeepLearning.EUnsupervisedSegmenter.PatchSize / eVision.EasyDeepLearning.EUnsupervisedSegmenter.SamplingDensity.

Scale() [1/2]

float eVision.EasyDeepLearning.EUnsupervisedSegmenter.Scale

(

self

)

Down-scaling applied to images before processing them.

Its value is between 0 and 1 (default value: 1).

Scale() [2/2]

eVision.EasyDeepLearning.EUnsupervisedSegmenter.Scale	(		self,
		float	value )

Down-scaling applied to images before processing them.

Its value is between 0 and 1 (default value: 1).

ScaleDisabledAtInference() [1/2]

bool eVision.EasyDeepLearning.EUnsupervisedSegmenter.ScaleDisabledAtInference

(

self

)

Whether to apply the scale parameter at inference or not.

If you disable the scale at inference, make sure your images are already scaled. For example, if you trained with images of 1024x1024 and a scale of 0.5, the inference images should be 512 x 512 images with the same field of view as the training images.

ScaleDisabledAtInference() [2/2]

eVision.EasyDeepLearning.EUnsupervisedSegmenter.ScaleDisabledAtInference	(		self,
		bool	value )

Whether to apply the scale parameter at inference or not.

If you disable the scale at inference, make sure your images are already scaled. For example, if you trained with images of 1024x1024 and a scale of 0.5, the inference images should be 512 x 512 images with the same field of view as the training images.

SerializeSettings()

None eVision.EasyDeepLearning.EUnsupervisedSegmenter.SerializeSettings	(		self,
		eVision.ESerializer | None	serializer )

Serializes the settings of the unsupervised segmenter.

Parameters

serializer

Pointer to eVision.ESerializer

Reimplemented from eVision.EasyDeepLearning.EDeepLearningTool.

ToolType()

eVision.EasyDeepLearning.EDeepLearningToolType eVision.EasyDeepLearning.EUnsupervisedSegmenter.ToolType

(

self

)

Type of the deep learning tool.

Reimplemented from eVision.EasyDeepLearning.EDeepLearningTool.

Train() [1/8]

None eVision.EasyDeepLearning.EUnsupervisedSegmenter.Train	(		self,
		eVision.EasyDeepLearning.EDataset	dataset )

Trains the eVision.EasyDeepLearning.EUnsupervisedSegmenter with the given dataset for the specified number of iterations.

At the end of the training, the deep learning tool is in the state it was at the iteration that gave the minimum validation error. See eVision.EasyDeepLearning.EDeepLearningTool.BestIteration.

If no iterations is given, the training will stop automatically.

Remarks:

License(s): EasySegment

Parameters

dataset

eVision.EasyDeepLearning.EDataset with which to train and validate the deep learning tool

Train() [2/8]

None eVision.EasyDeepLearning.EUnsupervisedSegmenter.Train	(		self,
		eVision.EasyDeepLearning.EDataset	dataset,
		eVision.EasyDeepLearning.EDataAugmentation	dataAugmentation )

Trains the eVision.EasyDeepLearning.EUnsupervisedSegmenter with the given dataset for the specified number of iterations.

At the end of the training, the deep learning tool is in the state it was at the iteration that gave the minimum validation error. See eVision.EasyDeepLearning.EDeepLearningTool.BestIteration.

If no iterations is given, the training will stop automatically.

Remarks:

License(s): EasySegment

Parameters

dataset	eVision.EasyDeepLearning.EDataset with which to train and validate the deep learning tool
dataAugmentation	Data augmentation to use during training

Train() [3/8]

None eVision.EasyDeepLearning.EUnsupervisedSegmenter.Train	(		self,
		eVision.EasyDeepLearning.EDataset	dataset,
		eVision.EasyDeepLearning.EDataAugmentation	dataAugmentation,
		int	iterations )

Trains the eVision.EasyDeepLearning.EUnsupervisedSegmenter with the given dataset for the specified number of iterations.

At the end of the training, the deep learning tool is in the state it was at the iteration that gave the minimum validation error. See eVision.EasyDeepLearning.EDeepLearningTool.BestIteration.

If no iterations is given, the training will stop automatically.

Remarks:

License(s): EasySegment

Parameters

dataset	eVision.EasyDeepLearning.EDataset with which to train and validate the deep learning tool
dataAugmentation	Data augmentation to use during training
iterations	Number of iterations for training.

Train() [4/8]

None eVision.EasyDeepLearning.EUnsupervisedSegmenter.Train	(		self,
		eVision.EasyDeepLearning.EDataset	dataset,
		int	iterations )

Trains the eVision.EasyDeepLearning.EUnsupervisedSegmenter with the given dataset for the specified number of iterations.

At the end of the training, the deep learning tool is in the state it was at the iteration that gave the minimum validation error. See eVision.EasyDeepLearning.EDeepLearningTool.BestIteration.

If no iterations is given, the training will stop automatically.

Remarks:

License(s): EasySegment

Parameters

dataset	eVision.EasyDeepLearning.EDataset with which to train and validate the deep learning tool
iterations	Number of iterations for training.

Train() [5/8]

None eVision.EasyDeepLearning.EUnsupervisedSegmenter.Train	(		self,
		eVision.EasyDeepLearning.EDataset	trainingDataset,
		eVision.EasyDeepLearning.EDataset	validationDataset )

Trains the eVision.EasyDeepLearning.EUnsupervisedSegmenter with the given dataset for the specified number of iterations.

At the end of the training, the deep learning tool is in the state it was at the iteration that gave the minimum validation error. See eVision.EasyDeepLearning.EDeepLearningTool.BestIteration.

If no iterations is given, the training will stop automatically.

Remarks:

License(s): EasySegment

Parameters

trainingDataset	eVision.EasyDeepLearning.EDataset with which to train the deep learning tool
validationDataset	eVision.EasyDeepLearning.EDataset with which to validate the deep learning tool

Train() [6/8]

None eVision.EasyDeepLearning.EUnsupervisedSegmenter.Train	(		self,
		eVision.EasyDeepLearning.EDataset	trainingDataset,
		eVision.EasyDeepLearning.EDataset	validationDataset,
		eVision.EasyDeepLearning.EDataAugmentation	dataAugmentation )

Trains the eVision.EasyDeepLearning.EUnsupervisedSegmenter with the given dataset for the specified number of iterations.

At the end of the training, the deep learning tool is in the state it was at the iteration that gave the minimum validation error. See eVision.EasyDeepLearning.EDeepLearningTool.BestIteration.

If no iterations is given, the training will stop automatically.

Remarks:

License(s): EasySegment

Parameters

trainingDataset	eVision.EasyDeepLearning.EDataset with which to train the deep learning tool
validationDataset	eVision.EasyDeepLearning.EDataset with which to validate the deep learning tool
dataAugmentation	Data augmentation to use during training

Train() [7/8]

None eVision.EasyDeepLearning.EUnsupervisedSegmenter.Train	(		self,
		eVision.EasyDeepLearning.EDataset	trainingDataset,
		eVision.EasyDeepLearning.EDataset	validationDataset,
		eVision.EasyDeepLearning.EDataAugmentation	dataAugmentation,
		int	iterations )

Trains the eVision.EasyDeepLearning.EUnsupervisedSegmenter with the given dataset for the specified number of iterations.

At the end of the training, the deep learning tool is in the state it was at the iteration that gave the minimum validation error. See eVision.EasyDeepLearning.EDeepLearningTool.BestIteration.

If no iterations is given, the training will stop automatically.

Remarks:

License(s): EasySegment

Parameters

trainingDataset	eVision.EasyDeepLearning.EDataset with which to train the deep learning tool
validationDataset	eVision.EasyDeepLearning.EDataset with which to validate the deep learning tool
dataAugmentation	Data augmentation to use during training
iterations	Number of iterations for training.

Train() [8/8]

None eVision.EasyDeepLearning.EUnsupervisedSegmenter.Train	(		self,
		eVision.EasyDeepLearning.EDataset	trainingDataset,
		eVision.EasyDeepLearning.EDataset	validationDataset,
		int	iterations )

Trains the eVision.EasyDeepLearning.EUnsupervisedSegmenter with the given dataset for the specified number of iterations.

At the end of the training, the deep learning tool is in the state it was at the iteration that gave the minimum validation error. See eVision.EasyDeepLearning.EDeepLearningTool.BestIteration.

If no iterations is given, the training will stop automatically.

Remarks:

License(s): EasySegment

Parameters

trainingDataset	eVision.EasyDeepLearning.EDataset with which to train the deep learning tool
validationDataset	eVision.EasyDeepLearning.EDataset with which to validate the deep learning tool
iterations	Number of iterations for training.