You are here: Start » FIL.NET » FIS.TrainOcr_SVM Method

FIS.TrainOcr_SVM Method

Trains an OCR support vector machines classifier.

Namespace:FilNet
Assembly:FIL.NET.dll

Syntax

public static void TrainOcr_SVM
(
	IList<FilNet.CharacterSample> inCharacterSamples,
	FilNet.Size inNormalizationSize,
	float? inNu,
	float? inKernelGamma,
	float inRegularizationConstant,
	float inStopEpsilon,
	bool inUseShrinkingHeuristics,
	FilNet.Size? inCharacterSize,
	int? inRandomSeed,
	FilNet.CharacterFeatures inCharacterFeatures,
	FilNet.OcrModel outOcrModel,
	out float outTrainingAccuracy,
	IList<FilNet.Image> diagNormalizedCharacters
)

Parameters

Name Type Range Default Description
inCharacterSamplesSystem.Collections.Generic.IList<FilNet.CharacterSample>Training font created from sample regions.
inNormalizationSizeFilNet.Size(Width: 16, Height: 16)The character size after normalization. Default value: (Width: 16, Height: 16).
inNufloat?<0.0f, 1.0f>Trade-off between training accuracy and number of supported vectors. Default value: ftl::NIL.
inKernelGammafloat?Gamma parameter for RBF kernel. Default value: ftl::NIL.
inRegularizationConstantfloat<0.0f, INF>1.0fPreventing overfitting. Default value: 1.0f.
inStopEpsilonfloat0.001fEpsilon for stopping criterion. Default value: 0.001f.
inUseShrinkingHeuristicsboolTrueHeuristics may speed up computations. Default value: True.
inCharacterSizeFilNet.Size?Size of fixed width font. Default value: ftl::NIL.
inRandomSeedint?<0, +INF>Random seed used to train classifier. Default value: ftl::NIL.
inCharacterFeaturesFilNet.CharacterFeatures(Pixels: True)Character features used to identify characters. Default value: (Pixels: True).
outOcrModelFilNet.OcrModelTrained OcrSvmModel used to recognize characters.
outTrainingAccuracyfloatThe overall training score.
diagNormalizedCharactersSystem.Collections.Generic.IList<FilNet.Image>Images of normalized characters used to train classifier.

Function Overrides

See also