CVB.Net 15.0
Loading...
Searching...
No Matches

Factory class for the generation of search predictors. More...

Inherits PredictorFactoryBase.

Public Member Functions

 SearchPredictorFactory ()
 Ctor.
 
void UseSettingsFromPredictor (SearchPredictor clf)
 Copy the predictor generation settings from a predictor.
 
SearchPredictor TrainPredictor (SampleRegressionImageList sil)
 Create a regression predictor from a sample image list.
 
SearchPredictor TrainPredictor (TrainingSet mts)
 Create a regression predictor from a Minos Training Set.
 
- Public Member Functions inherited from PredictorFactoryBase
void Dispose ()
 Disposes of the native data needed for marshalling.
 
void UseSettingsFromPredictor (PredictorBase clf)
 Copy the classifier generation settings from a classifier.
 

Static Public Attributes

static readonly ValueRange< double > ScaleFactorRangeDefault = new ValueRange<double>(1.0, 1.0)
 Default value for the scale factor range.
 
static readonly ValueRange< double > SingularValueRangeDefault = new ValueRange<double>(1.0, 1.0)
 Default value for the singular value range.
 
static readonly ValueRange< double > ScaleFactorRange = new ValueRange<double>(0.25, 10)
 Acceptable scale factor range for search classifier training.
 
const int NumClassificationStepsMin = 3
 Minimum classification step count.
 
const int SampleSizeMin = 1
 Absolute minimum value for SampleSize.
 
- Static Public Attributes inherited from PredictorFactoryBase
static readonly ValueRange< double > LambdaRange = new ValueRange<double>(0.0, 10.0)
 Acceptable scale factor range for search classifier training.
 
const double LambdaDefault = 0.1
 Default value for lambda.
 
const int FeatureResolutionDefault = 1
 Default value for feature resolution.
 
static readonly ValueRange< int > FeatureResolutionRange = new ValueRange<int>(0, 10)
 Valid range of feature resolution value.
 
const int PreprocessingMaxLength = InternalExtensions.PreprocessingMaxLength
 Maximum length of a preprocessing code (excluding the terminating zero).
 
const string PreprocessingValidCharacters = InternalExtensions.PreprocessingValidCharacters
 characters that a preprocessing string may contain.
 

Properties

int SampleSize [get, set]
 Sample size that has been used in each training set.
 
double ExtractionRadius [get, set]
 Radius for extracting positive search instances. The unit size is the size of the feature window, i.e. a value of 0.5 means that positive samples can be extracted from withing a range of +/- Feature Window Width and +/- Feature Window Height around each trained instance. Smaller values will necessitate a finer search grid but may yield better results on difficult search tasks.
 
ValueRange< AngleAngleRange [get, set]
 Range of angles to be covered during classifier training.
 
ValueRange< double > ScaleRange [get, set]
 Range of scale factors to be covered during classifier training. Only meaningful if the invariance type InvarianceType.RotationScaleTranslation has been set.
 
ValueRange< double > SingularValueRange [get, set]
 Range of permissible singular values to be covered during classifier training. Only meaningful if the invariance type InvarianceType.AffineGroup has been set.
 
InvarianceType Invariances [get, set]
 Invariances to be used for generating the classifier.
 
int NumClassificationSteps [get, set]
 Number of classification steps during search operation.
 
int FeatureResolutionStep1And2 [get, set]
 Feature resolution (determines the size of the classification retina) for search step 1 and 2.
 
int FeatureResolutionRest [get, set]
 Feature resolution (determines the size of the classification retina) for step 3 and up.
 
- Properties inherited from PredictorFactoryBase
double Lambda [get, set]
 Regularization value to be used for generating the object. Possible values range from 0 to 10, good starting values for experiments are usually around 0.01.
 
string Preprocessing [get, set]
 Preprocessing code with which the object is to be generated.
 

Additional Inherited Members

- Static Public Member Functions inherited from PredictorFactoryBase
static string FormatPreprocessingCode (string input)
 Correct a preprocessing code to make sure that no invalid characters are in the code and the code does not exceed the maximum length.
 
- Protected Member Functions inherited from PredictorFactoryBase
 PredictorFactoryBase ()
 ctor (internal).
 
virtual void Dispose (bool disposing)
 Clean-up of managed-to-native data structures.
 
virtual void OnStartProgressReport (int id, string caption, int stepsTotal)
 Event raiser.
 
virtual void OnProgressReport (int id, ref bool retval)
 Event raiser.
 
virtual void OnEndProgressReport (int id)
 Event raiser.
 
- Events inherited from PredictorFactoryBase
EventHandler< StartProgressReportEventArgsStartProgressReport
 Event that will be fired when a new progress stage was initiated.
 
EventHandler< ProgressReportEventArgsProgressReport
 Event that informs about a progress step in the given stage.
 
EventHandler< EndProgressReportEventArgsEndProgressReport
 Event that informs about the finalization of a given stage.
 

Detailed Description

Factory class for the generation of search predictors.

Member Function Documentation

◆ TrainPredictor() [1/2]

Create a regression predictor from a sample image list.

Parameters
silSample image list to train from.
Returns
Newly generated classifier.
Exceptions
PolimagoFactoryExceptionTraining error (usually due to unsuitable parameters).

◆ TrainPredictor() [2/2]

SearchPredictor TrainPredictor ( TrainingSet mts)

Create a regression predictor from a Minos Training Set.

Parameters
mtsTraining set to train from.
Returns
Newly generated classifier.
Exceptions
PolimagoFactoryExceptionTraining error (usually due to unsuitable parameters).

◆ UseSettingsFromPredictor()

void UseSettingsFromPredictor ( SearchPredictor clf)

Copy the predictor generation settings from a predictor.

Parameters
clfClassifier to take the settings from.