Classes
class	oevislib_net.ComputerVision.OCR.SVM.Model
	Model using Support Vector Machines. More...

Functions
static SVM.Model	oevislib_net.ComputerVision.OCR.OCR.Train (Array< CharacterSample > inCharacterSamples, Size inNormalizationSize=null, Optional< Size > inCharacterSize=null, KernelSVM inKernel=KernelSVM.Linear, Array< CharacterFeaturesType > inCharacterFeatures=null, Optional< float > inNu=null, Optional< float > inGamma=null, Optional< float > inDegree=null, float inC=1.0f, float inEpsilon=0.001f, Optional< float > outTrainingAccuracy=null, Optional< Array< Image > > outNormalizedCharacters=null)
	Trains Support Vector Machines classifier.

Detailed Description

Function Documentation

SVM.Model oevislib_net.ComputerVision.OCR.OCR.Train	(	Array< CharacterSample >	inCharacterSamples,
		Size	inNormalizationSize = null,
		Optional< Size >	inCharacterSize = null,
		KernelSVM	inKernel = KernelSVM::Linear,
		Array< CharacterFeaturesType >	inCharacterFeatures = null,
		Optional< float >	inNu = null,
		Optional< float >	inGamma = null,
		Optional< float >	inDegree = null,
		float	inC = 1::0f,
		float	inEpsilon = 0::001f,
		Optional< float >	outTrainingAccuracy = null,
		Optional< Array< Image > >	outNormalizedCharacters = null )

inlinestatic

Trains Support Vector Machines classifier.

Parameters

inCharacterSamples	Character samples used for training the model.
inNormalizationSize	Character size after normalization.
inCharacterSize	Size of internal font.
inKernel	Type of kernel.
inCharacterFeatures	Character features selected.
inNu	Trade-off between training accuracy and number of supported vectors. The larger the value, the smoother the decision boundary. Range: [0, 1].
inGamma	Gamma parameter for Polynomial, RBF, Sigmoid or ExponentialChi2 kernel. Range: (-inf, +inf).
inDegree	Degree parameter for Polynomial kernel. Range: [0, +inf).
inC	Prevents overfitting working as a penalty multiplier for outliers. Range: [0, +inf).
inEpsilon	How much the error could change between the iterations to make the algorithm continue. Range: (-inf, +inf).
outTrainingAccuracy	Training accuracy.
outNormalizedCharacters	Images of normalized characters.