ASR::GMM Class Reference

#include <GMM.h>

Public Member Functions
void	addGaussianModel (float weight, std::vector< float > mean, MatrixXf covariance)
ASR::MultidimensionalGaussianPtr	getModel (unsigned int index)
	GMM ()
void	sampleRandomValues (std::vector< float > &x)

Private Member Functions
float	gen_random_float (float min, float max)

Private Attributes
std::vector< float >	histogramm
std::vector< ASR::MultidimensionalGaussianPtr >	models
boost::mt19937	rng
std::vector< float >	weights

Detailed Description

This class contains a list of multidimensional gaussians. With weights they model a gaussian mixture model.

Definition at line 32 of file GMM.h.

Constructor & Destructor Documentation

ASR::GMM::GMM ( )

inline

Definition at line 55 of file GMM.h.

Member Function Documentation

void ASR::GMM::addGaussianModel ( float  weight, std::vector< float >  mean, MatrixXf  covariance  )

inline

Adds a new multidimensional gaussian to the gaussian mixture model.

Parameters

weight	- the weight of the new gaussian kernel
mean	- mean values of the new gaussian kernel
covariance	- the covariance matrix of the new gaussian kernel

Definition at line 68 of file GMM.h.

float ASR::GMM::gen_random_float ( float  min, float  max  )

inlineprivate

Get a random number between min and max.

Parameters

min	- min value
max	- max value

Returns

xi € [min, max)

Definition at line 47 of file GMM.h.

ASR::MultidimensionalGaussianPtr ASR::GMM::getModel ( unsigned int  index )

inline

Returns the model at the given index

Parameters

index

- the index of the model

Returns

the model at the given index

Definition at line 125 of file GMM.h.

void ASR::GMM::sampleRandomValues ( std::vector< float > &  x )

inline

Samples a random vector from the gaussian mixture model and stores the values in the given vector x. First a gaussian kernel is randomly selected. "Heavier" weighted kernels are more likely than "lighter" weighted kernels. The dimension of x depends on the dimension of the kernels which were given in the initialization.

Parameters

x	- the output vector which contains the sampled values.

Definition at line 86 of file GMM.h.

Member Data Documentation

std::vector<float> ASR::GMM::histogramm

private

Definition at line 37 of file GMM.h.

std::vector<ASR::MultidimensionalGaussianPtr> ASR::GMM::models

private

Definition at line 35 of file GMM.h.

boost::mt19937 ASR::GMM::rng

private

Definition at line 38 of file GMM.h.

std::vector<float> ASR::GMM::weights

private

Definition at line 36 of file GMM.h.

---

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

• GMM.h