cv::ml::SVM Class Reference

Support Vector Machines. More...

#include <opencv2/ml.hpp>

Inheritance diagram for cv::ml::SVM:

Classes
class	Kernel

Public Types
enum	KernelTypes {   CUSTOM =-1,   LINEAR =0,   POLY =1,   RBF =2,   SIGMOID =3,   CHI2 =4,   INTER =5 }
	SVM kernel type More...
enum	ParamTypes {   C =0,   GAMMA =1,   P =2,   NU =3,   COEF =4,   DEGREE =5 }
	SVM params type More...
enum	Types {   C_SVC =100,   NU_SVC =101,   ONE_CLASS =102,   EPS_SVR =103,   NU_SVR =104 }
	SVM type More...
Public Types inherited from cv::ml::StatModel
enum	Flags {   UPDATE_MODEL = 1,   RAW_OUTPUT =1,   COMPRESSED_INPUT =2,   PREPROCESSED_INPUT =4 }

Public Member Functions
virtual double	getC () const =0
virtual cv::Mat	getClassWeights () const =0
virtual double	getCoef0 () const =0
virtual double	getDecisionFunction (int i, OutputArray alpha, OutputArray svidx) const =0
	Retrieves the decision function. More...
virtual double	getDegree () const =0
virtual double	getGamma () const =0
virtual int	getKernelType () const =0
virtual double	getNu () const =0
virtual double	getP () const =0
virtual Mat	getSupportVectors () const =0
	Retrieves all the support vectors. More...
virtual cv::TermCriteria	getTermCriteria () const =0
virtual int	getType () const =0
virtual Mat	getUncompressedSupportVectors () const =0
	Retrieves all the uncompressed support vectors of a linear SVM. More...
virtual void	setC (double val)=0
virtual void	setClassWeights (const cv::Mat &val)=0
virtual void	setCoef0 (double val)=0
virtual void	setCustomKernel (const Ptr< Kernel > &_kernel)=0
virtual void	setDegree (double val)=0
virtual void	setGamma (double val)=0
virtual void	setKernel (int kernelType)=0
virtual void	setNu (double val)=0
virtual void	setP (double val)=0
virtual void	setTermCriteria (const cv::TermCriteria &val)=0
virtual void	setType (int val)=0
virtual bool	trainAuto (const Ptr< TrainData > &data, int kFold=10, ParamGrid Cgrid=getDefaultGrid(C), ParamGrid gammaGrid=getDefaultGrid(GAMMA), ParamGrid pGrid=getDefaultGrid(P), ParamGrid nuGrid=getDefaultGrid(NU), ParamGrid coeffGrid=getDefaultGrid(COEF), ParamGrid degreeGrid=getDefaultGrid(DEGREE), bool balanced=false)=0
	Trains an SVM with optimal parameters. More...
virtual bool	trainAuto (InputArray samples, int layout, InputArray responses, int kFold=10, Ptr< ParamGrid > Cgrid=SVM::getDefaultGridPtr(SVM::C), Ptr< ParamGrid > gammaGrid=SVM::getDefaultGridPtr(SVM::GAMMA), Ptr< ParamGrid > pGrid=SVM::getDefaultGridPtr(SVM::P), Ptr< ParamGrid > nuGrid=SVM::getDefaultGridPtr(SVM::NU), Ptr< ParamGrid > coeffGrid=SVM::getDefaultGridPtr(SVM::COEF), Ptr< ParamGrid > degreeGrid=SVM::getDefaultGridPtr(SVM::DEGREE), bool balanced=false)=0
	Trains an SVM with optimal parameters. More...
Public Member Functions inherited from cv::ml::StatModel
virtual float	calcError (const Ptr< TrainData > &data, bool test, OutputArray resp) const
	Computes error on the training or test dataset. More...
virtual bool	empty () const CV_OVERRIDE
	Returns true if the Algorithm is empty (e.g. in the very beginning or after unsuccessful read. More...
virtual int	getVarCount () const =0
	Returns the number of variables in training samples. More...
virtual bool	isClassifier () const =0
	Returns true if the model is classifier. More...
virtual bool	isTrained () const =0
	Returns true if the model is trained. More...
virtual float	predict (InputArray samples, OutputArray results=noArray(), int flags=0) const =0
	Predicts response(s) for the provided sample(s) More...
virtual bool	train (const Ptr< TrainData > &trainData, int flags=0)
	Trains the statistical model. More...
virtual bool	train (InputArray samples, int layout, InputArray responses)
	Trains the statistical model. More...
Public Member Functions inherited from cv::Algorithm
	Algorithm ()
virtual	~Algorithm ()
virtual void	clear ()
	Clears the algorithm state. More...
virtual String	getDefaultName () const
virtual void	read (const FileNode &fn)
	Reads algorithm parameters from a file storage. More...
virtual void	save (const String &filename) const
virtual void	write (FileStorage &fs) const
	Stores algorithm parameters in a file storage. More...
void	write (const Ptr< FileStorage > &fs, const String &name=String()) const
	simplified API for language bindings This is an overloaded member function, provided for convenience. It differs from the above function only in what argument(s) it accepts. More...

Static Public Member Functions
static Ptr< SVM >	create ()
static ParamGrid	getDefaultGrid (int param_id)
	Generates a grid for SVM parameters. More...
static Ptr< ParamGrid >	getDefaultGridPtr (int param_id)
	Generates a grid for SVM parameters. More...
static Ptr< SVM >	load (const String &filepath)
	Loads and creates a serialized svm from a file. More...
Static Public Member Functions inherited from cv::ml::StatModel
template<typename _Tp >
static Ptr< _Tp >	train (const Ptr< TrainData > &data, int flags=0)
	Create and train model with default parameters. More...
Static Public Member Functions inherited from cv::Algorithm
template<typename _Tp >
static Ptr< _Tp >	load (const String &filename, const String &objname=String())
	Loads algorithm from the file. More...
template<typename _Tp >
static Ptr< _Tp >	loadFromString (const String &strModel, const String &objname=String())
	Loads algorithm from a String. More...
template<typename _Tp >
static Ptr< _Tp >	read (const FileNode &fn)
	Reads algorithm from the file node. More...

Additional Inherited Members
Protected Member Functions inherited from cv::Algorithm
void	writeFormat (FileStorage &fs) const

Detailed Description

Support Vector Machines.

See also

Support Vector Machines

Member Enumeration Documentation

KernelTypes

enum cv::ml::SVM::KernelTypes

SVM kernel type

A comparison of different kernels on the following 2D test case with four classes. Four SVM::C_SVC SVMs have been trained (one against rest) with auto_train. Evaluation on three different kernels (SVM::CHI2, SVM::INTER, SVM::RBF). The color depicts the class with max score. Bright means max-score > 0, dark means max-score < 0.

image

Enumerator
CUSTOM	Returned by SVM::getKernelType in case when custom kernel has been set
LINEAR	Linear kernel. No mapping is done, linear discrimination (or regression) is done in the original feature space. It is the fastest option. \(K(x_i, x_j) = x_i^T x_j\).
POLY	Polynomial kernel: \(K(x_i, x_j) = (\gamma x_i^T x_j + coef0)^{degree}, \gamma > 0\).
RBF	Radial basis function (RBF), a good choice in most cases. \(K(x_i, x_j) = e^{-\gamma ||x_i - x_j||^2}, \gamma > 0\).
SIGMOID	Sigmoid kernel: \(K(x_i, x_j) = \tanh(\gamma x_i^T x_j + coef0)\).
CHI2	Exponential Chi2 kernel, similar to the RBF kernel: \(K(x_i, x_j) = e^{-\gamma \chi^2(x_i,x_j)}, \chi^2(x_i,x_j) = (x_i-x_j)^2/(x_i+x_j), \gamma > 0\).
INTER	Histogram intersection kernel. A fast kernel. \(K(x_i, x_j) = min(x_i,x_j)\).

ParamTypes

enum cv::ml::SVM::ParamTypes

SVM params type

Enumerator
C
GAMMA
P
NU
COEF
DEGREE

Types

enum cv::ml::SVM::Types

SVM type

Enumerator
C_SVC	C-Support Vector Classification. n-class classification (n \(\geq\) 2), allows imperfect separation of classes with penalty multiplier C for outliers.
NU_SVC	\(\nu\)-Support Vector Classification. n-class classification with possible imperfect separation. Parameter \(\nu\) (in the range 0..1, the larger the value, the smoother the decision boundary) is used instead of C.
ONE_CLASS	Distribution Estimation (One-class SVM). All the training data are from the same class, SVM builds a boundary that separates the class from the rest of the feature space.
EPS_SVR	\(\epsilon\)-Support Vector Regression. The distance between feature vectors from the training set and the fitting hyper-plane must be less than p. For outliers the penalty multiplier C is used.
NU_SVR	\(\nu\)-Support Vector Regression. \(\nu\) is used instead of p. See [37] for details.

Member Function Documentation

create()

static Ptr<SVM> cv::ml::SVM::create ( )

static

Python:
	retval	=	cv.ml.SVM_create(		)

Creates empty model. Use StatModel::train to train the model. Since SVM has several parameters, you may want to find the best parameters for your problem, it can be done with SVM::trainAuto.

Examples:

samples/cpp/train_HOG.cpp.

getC()

virtual double cv::ml::SVM::getC ( ) const

pure virtual

Python:
	retval	=	cv.ml_SVM.getC(		)

Parameter C of a SVM optimization problem. For SVM::C_SVC, SVM::EPS_SVR or SVM::NU_SVR. Default value is 0.

See also

setC

getClassWeights()

virtual cv::Mat cv::ml::SVM::getClassWeights ( ) const

pure virtual

Python:
	retval	=	cv.ml_SVM.getClassWeights(		)

Optional weights in the SVM::C_SVC problem, assigned to particular classes. They are multiplied by C so the parameter C of class i becomes classWeights(i) * C. Thus these weights affect the misclassification penalty for different classes. The larger weight, the larger penalty on misclassification of data from the corresponding class. Default value is empty Mat.

See also

setClassWeights

getCoef0()

virtual double cv::ml::SVM::getCoef0 ( ) const

pure virtual

Python:
	retval	=	cv.ml_SVM.getCoef0(		)

Parameter coef0 of a kernel function. For SVM::POLY or SVM::SIGMOID. Default value is 0.

See also

setCoef0

getDecisionFunction()

virtual double cv::ml::SVM::getDecisionFunction ( int  i, OutputArray  alpha, OutputArray  svidx  ) const

pure virtual

Python:
	retval, alpha, svidx	=	cv.ml_SVM.getDecisionFunction(	i[, alpha[, svidx]]	)

Retrieves the decision function.

Parameters

i	the index of the decision function. If the problem solved is regression, 1-class or 2-class classification, then there will be just one decision function and the index should always be 0. Otherwise, in the case of N-class classification, there will be \(N(N-1)/2\) decision functions.
alpha	the optional output vector for weights, corresponding to different support vectors. In the case of linear SVM all the alpha's will be 1's.
svidx	the optional output vector of indices of support vectors within the matrix of support vectors (which can be retrieved by SVM::getSupportVectors). In the case of linear SVM each decision function consists of a single "compressed" support vector.

The method returns rho parameter of the decision function, a scalar subtracted from the weighted sum of kernel responses.

getDefaultGrid()

static ParamGrid cv::ml::SVM::getDefaultGrid ( int  param_id )

static

Generates a grid for SVM parameters.

Parameters

param_id

SVM parameters IDs that must be one of the SVM::ParamTypes. The grid is generated for the parameter with this ID.

The function generates a grid for the specified parameter of the SVM algorithm. The grid may be passed to the function SVM::trainAuto.

getDefaultGridPtr()

static Ptr<ParamGrid> cv::ml::SVM::getDefaultGridPtr ( int  param_id )

static

Python:
	retval	=	cv.ml.SVM_getDefaultGridPtr(	param_id	)

Generates a grid for SVM parameters.

Parameters

param_id

SVM parameters IDs that must be one of the SVM::ParamTypes. The grid is generated for the parameter with this ID.

The function generates a grid pointer for the specified parameter of the SVM algorithm. The grid may be passed to the function SVM::trainAuto.

getDegree()

virtual double cv::ml::SVM::getDegree ( ) const

pure virtual

Python:
	retval	=	cv.ml_SVM.getDegree(		)

Parameter degree of a kernel function. For SVM::POLY. Default value is 0.

See also

setDegree

getGamma()

virtual double cv::ml::SVM::getGamma ( ) const

pure virtual

Python:
	retval	=	cv.ml_SVM.getGamma(		)

Parameter \(\gamma\) of a kernel function. For SVM::POLY, SVM::RBF, SVM::SIGMOID or SVM::CHI2. Default value is 1.

See also

setGamma

getKernelType()

virtual int cv::ml::SVM::getKernelType ( ) const

pure virtual

Python:
	retval	=	cv.ml_SVM.getKernelType(		)

Type of a SVM kernel. See SVM::KernelTypes. Default value is SVM::RBF.

getNu()

virtual double cv::ml::SVM::getNu ( ) const

pure virtual

Python:
	retval	=	cv.ml_SVM.getNu(		)

Parameter \(\nu\) of a SVM optimization problem. For SVM::NU_SVC, SVM::ONE_CLASS or SVM::NU_SVR. Default value is 0.

See also

setNu

getP()

virtual double cv::ml::SVM::getP ( ) const

pure virtual

Python:
	retval	=	cv.ml_SVM.getP(		)

Parameter \(\epsilon\) of a SVM optimization problem. For SVM::EPS_SVR. Default value is 0.

See also

setP

getSupportVectors()

virtual Mat cv::ml::SVM::getSupportVectors ( ) const

pure virtual

Python:
	retval	=	cv.ml_SVM.getSupportVectors(		)

Retrieves all the support vectors.

The method returns all the support vectors as a floating-point matrix, where support vectors are stored as matrix rows.

getTermCriteria()

virtual cv::TermCriteria cv::ml::SVM::getTermCriteria ( ) const

pure virtual

Python:
	retval	=	cv.ml_SVM.getTermCriteria(		)

Termination criteria of the iterative SVM training procedure which solves a partial case of constrained quadratic optimization problem. You can specify tolerance and/or the maximum number of iterations. Default value is TermCriteria( TermCriteria::MAX_ITER + TermCriteria::EPS, 1000, FLT_EPSILON );

See also

setTermCriteria

getType()

virtual int cv::ml::SVM::getType ( ) const

pure virtual

Python:
	retval	=	cv.ml_SVM.getType(		)

Type of a SVM formulation. See SVM::Types. Default value is SVM::C_SVC.

See also

setType

getUncompressedSupportVectors()

virtual Mat cv::ml::SVM::getUncompressedSupportVectors ( ) const

pure virtual

Python:
	retval	=	cv.ml_SVM.getUncompressedSupportVectors(		)

Retrieves all the uncompressed support vectors of a linear SVM.

The method returns all the uncompressed support vectors of a linear SVM that the compressed support vector, used for prediction, was derived from. They are returned in a floating-point matrix, where the support vectors are stored as matrix rows.

load()

static Ptr<SVM> cv::ml::SVM::load ( const String &  filepath )

static

Python:
	retval	=	cv.ml.SVM_load(	filepath	)

Loads and creates a serialized svm from a file.

Use SVM::save to serialize and store an SVM to disk. Load the SVM from this file again, by calling this function with the path to the file.

Parameters

filepath

path to serialized svm

setC()

virtual void cv::ml::SVM::setC ( double  val )

pure virtual

Python:
	None	=	cv.ml_SVM.setC(	val	)

See also

getC

setClassWeights()

virtual void cv::ml::SVM::setClassWeights ( const cv::Mat &  val )

pure virtual

Python:
	None	=	cv.ml_SVM.setClassWeights(	val	)

See also

getClassWeights

setCoef0()

virtual void cv::ml::SVM::setCoef0 ( double  val )

pure virtual

Python:
	None	=	cv.ml_SVM.setCoef0(	val	)

See also

getCoef0

setCustomKernel()

virtual void cv::ml::SVM::setCustomKernel ( const Ptr< Kernel > &  _kernel )

pure virtual

Initialize with custom kernel. See SVM::Kernel class for implementation details

setDegree()

virtual void cv::ml::SVM::setDegree ( double  val )

pure virtual

Python:
	None	=	cv.ml_SVM.setDegree(	val	)

See also

getDegree

setGamma()

virtual void cv::ml::SVM::setGamma ( double  val )

pure virtual

Python:
	None	=	cv.ml_SVM.setGamma(	val	)

See also

getGamma

setKernel()

virtual void cv::ml::SVM::setKernel ( int  kernelType )

pure virtual

Python:
	None	=	cv.ml_SVM.setKernel(	kernelType	)

Initialize with one of predefined kernels. See SVM::KernelTypes.

setNu()

virtual void cv::ml::SVM::setNu ( double  val )

pure virtual

Python:
	None	=	cv.ml_SVM.setNu(	val	)

See also

getNu

setP()

virtual void cv::ml::SVM::setP ( double  val )

pure virtual

Python:
	None	=	cv.ml_SVM.setP(	val	)

See also

getP

setTermCriteria()

virtual void cv::ml::SVM::setTermCriteria ( const cv::TermCriteria &  val )

pure virtual

Python:
	None	=	cv.ml_SVM.setTermCriteria(	val	)

See also

getTermCriteria

setType()

virtual void cv::ml::SVM::setType ( int  val )

pure virtual

Python:
	None	=	cv.ml_SVM.setType(	val	)

See also

getType

trainAuto() [1/2]

virtual bool cv::ml::SVM::trainAuto ( const Ptr< TrainData > &  data, int  kFold = 10, ParamGrid  Cgrid = getDefaultGrid(C), ParamGrid  gammaGrid = getDefaultGrid(GAMMA), ParamGrid  pGrid = getDefaultGrid(P), ParamGrid  nuGrid = getDefaultGrid(NU), ParamGrid  coeffGrid = getDefaultGrid(COEF), ParamGrid  degreeGrid = getDefaultGrid(DEGREE), bool  balanced = false  )

pure virtual

Python:
	retval	=	cv.ml_SVM.trainAuto(	samples, layout, responses[, kFold[, Cgrid[, gammaGrid[, pGrid[, nuGrid[, coeffGrid[, degreeGrid[, balanced]]]]]]]]	)

Trains an SVM with optimal parameters.

Parameters

data	the training data that can be constructed using TrainData::create or TrainData::loadFromCSV.
kFold	Cross-validation parameter. The training set is divided into kFold subsets. One subset is used to test the model, the others form the train set. So, the SVM algorithm is executed kFold times.
Cgrid	grid for C
gammaGrid	grid for gamma
pGrid	grid for p
nuGrid	grid for nu
coeffGrid	grid for coeff
degreeGrid	grid for degree
balanced	If true and the problem is 2-class classification then the method creates more balanced cross-validation subsets that is proportions between classes in subsets are close to such proportion in the whole train dataset.

The method trains the SVM model automatically by choosing the optimal parameters C, gamma, p, nu, coef0, degree. Parameters are considered optimal when the cross-validation estimate of the test set error is minimal.

If there is no need to optimize a parameter, the corresponding grid step should be set to any value less than or equal to 1. For example, to avoid optimization in gamma, set gammaGrid.step = 0, gammaGrid.minVal, gamma_grid.maxVal as arbitrary numbers. In this case, the value Gamma is taken for gamma.

And, finally, if the optimization in a parameter is required but the corresponding grid is unknown, you may call the function SVM::getDefaultGrid. To generate a grid, for example, for gamma, call SVM::getDefaultGrid(SVM::GAMMA).

This function works for the classification (SVM::C_SVC or SVM::NU_SVC) as well as for the regression (SVM::EPS_SVR or SVM::NU_SVR). If it is SVM::ONE_CLASS, no optimization is made and the usual SVM with parameters specified in params is executed.

trainAuto() [2/2]

virtual bool cv::ml::SVM::trainAuto ( InputArray  samples, int  layout, InputArray  responses, int  kFold = 10, Ptr< ParamGrid >  Cgrid = SVM::getDefaultGridPtr(SVM::C), Ptr< ParamGrid >  gammaGrid = SVM::getDefaultGridPtr(SVM::GAMMA), Ptr< ParamGrid >  pGrid = SVM::getDefaultGridPtr(SVM::P), Ptr< ParamGrid >  nuGrid = SVM::getDefaultGridPtr(SVM::NU), Ptr< ParamGrid >  coeffGrid = SVM::getDefaultGridPtr(SVM::COEF), Ptr< ParamGrid >  degreeGrid = SVM::getDefaultGridPtr(SVM::DEGREE), bool  balanced = false  )

pure virtual

Python:
	retval	=	cv.ml_SVM.trainAuto(	samples, layout, responses[, kFold[, Cgrid[, gammaGrid[, pGrid[, nuGrid[, coeffGrid[, degreeGrid[, balanced]]]]]]]]	)

Trains an SVM with optimal parameters.

Parameters

samples	training samples
layout	See ml::SampleTypes.
responses	vector of responses associated with the training samples.
kFold	Cross-validation parameter. The training set is divided into kFold subsets. One subset is used to test the model, the others form the train set. So, the SVM algorithm is
Cgrid	grid for C
gammaGrid	grid for gamma
pGrid	grid for p
nuGrid	grid for nu
coeffGrid	grid for coeff
degreeGrid	grid for degree
balanced	If true and the problem is 2-class classification then the method creates more balanced cross-validation subsets that is proportions between classes in subsets are close to such proportion in the whole train dataset.

The method trains the SVM model automatically by choosing the optimal parameters C, gamma, p, nu, coef0, degree. Parameters are considered optimal when the cross-validation estimate of the test set error is minimal.

This function only makes use of SVM::getDefaultGrid for parameter optimization and thus only offers rudimentary parameter options.

This function works for the classification (SVM::C_SVC or SVM::NU_SVC) as well as for the regression (SVM::EPS_SVR or SVM::NU_SVR). If it is SVM::ONE_CLASS, no optimization is made and the usual SVM with parameters specified in params is executed.

---

The documentation for this class was generated from the following file:

• opencv2/ml.hpp