You are here: Start » AVL.NET » AVL.TrainOcr_SVM(AvlNet.CharacterSample[], AvlNet.Size, float?, float?, float, float, bool, AvlNet.Size?, int?, AvlNet.CharacterFeatures, AvlNet.OcrModel, float)

AVL.TrainOcr_SVM(AvlNet.CharacterSample[], AvlNet.Size, float?, float?, float, float, bool, AvlNet.Size?, int?, AvlNet.CharacterFeatures, AvlNet.OcrModel, float)

Trains an OCR support vector machines classifier.

Namespace:AvlNet
Assembly:AVL.NET.dll

Syntax


public static void TrainOcr_SVM(
	AvlNet.CharacterSample[] inCharacterSamples,
	AvlNet.Size inNormalizationSize,
	float? inNu,
	float? inKernelGamma,
	float inRegularizationConstant,
	float inStopEpsilon,
	bool inUseShrinkingHeuristics,
	AvlNet.Size? inCharacterSize,
	int? inRandomSeed,
	AvlNet.CharacterFeatures inCharacterFeatures,
	out AvlNet.OcrModel outOcrModel,
	out float outTrainingAccuracy
)

Parameters

inCharacterSamples
Type: AvlNet.CharacterSample
Training font created from sample regions
inNormalizationSize
Type: AvlNet.Size
The character size after normalization
inNu
Type: System.Nullable<System.Single>
Trade-off between training accuracy and number of supported vectors, or null.
inKernelGamma
Type: System.Nullable<System.Single>
Gamma parameter for RBF kernel, or null.
inRegularizationConstant
Type: System.Single
Preventing overfitting
inStopEpsilon
Type: System.Single
Epsilon for stopping criterion
inUseShrinkingHeuristics
Type: System.Boolean
Heuristics may speed up computations
inCharacterSize
Type: System.Nullable<AvlNet.Size>
Size of fixed width font, or null.
inRandomSeed
Type: System.Nullable<System.Int32>
Random seed used to train classifier, or null.
inCharacterFeatures
Type: AvlNet.CharacterFeatures
Character features used to identify characters
outOcrModel
Type: AvlNet.OcrModel
Trained OcrSvmModel used to recognize characters
outTrainingAccuracy
Type: System.Single
The overall training score

See also