ClassifierFactory Class Reference

Learner object that creates a classifier from an image list. More...

Public Member Functions
cvb.minos.Classifier	learn (self, cvb.minos.TrainingSet training_set)
	Learn a new classifier from the training_set using the parameters stored in the properties of this object.

Properties
	contrast_trigger = property
	int: Minimum contrast a Minos feature must achieve, before it is eligible to become part of the classifier.
	contrast_trigger_default = property
	int: Default value for the contrast_trigger.
	ensemble_size = property
	int: Maximum size of the Ensembles of similar instance images to be used for pair feature calculation.
	ensemble_size_default = property
	int: Default value for the ensemble_size.
	indifference_radius = property
	int: Minimum distance to be assumed between a (labeled) positive sample and a counter sample in a training set image.
	indifference_radius_default = property
	int: Default value for the indifference_radius.
	min_pair_features = property
	int: Minimum number of features to extract for each model when building a classifier.
	min_pair_features_default = property
	int: Default value for the min_pair_features.
	negative_density = property
	float: Scan density with which to extract counter samples from the training set images.
	negative_density_default = property
	float: Default value for the negative_density.
	polydromy = property
	int: Polydromy value controlling the complexity of the feature search tree in the classifier.
	polydromy_default = property
	int: Default value for the polydromy.

Detailed Description

Learner object that creates a classifier from an image list.

Instantiates a ClassifierFactory object.

Member Function Documentation

learn()

cvb.minos.Classifier learn	(		self,
		cvb.minos.TrainingSet	training_set )

Learn a new classifier from the training_set using the parameters stored in the properties of this object.

Parameters

training_set : cvb.minos.TrainingSet Training set from which to learn.

Returns

cvb.minos.Classifier Newly learned classifier object.

Property Documentation

contrast_trigger

contrast_trigger = property

static

int: Minimum contrast a Minos feature must achieve, before it is eligible to become part of the classifier.

Lower values will lead to a classifier that is more sensitive in low-contrast situation but might also increase the number of false positive results that need to be filtered out for example by means of their quality measure.

ensemble_size

ensemble_size = property

static

int: Maximum size of the Ensembles of similar instance images to be used for pair feature calculation.

Valid ensemble sizes must be greater than 0, the default value is 15.

negative_density

negative_density = property

static

float: Scan density with which to extract counter samples from the training set images.

Higher densities will lead to a classifier, that is potentially more robust versus false positive detections, but will also result in increased learning times.

polydromy

polydromy = property

static

int: Polydromy value controlling the complexity of the feature search tree in the classifier.

Valid polydromy values must be greater than 0, the default value is 2.