Sum product Loopy Belief Propagation inference class. More...

#include <InferLBP.h>

Inheritance diagram for DirectGraphicalModels::CInferLBP:

[legend]

Collaboration diagram for DirectGraphicalModels::CInferLBP:

[legend]

Public Member Functions
	CInferLBP (CGraph *pGraph)
	Constructor. More...

virtual	~CInferLBP (void)

Public Member Functions inherited from DirectGraphicalModels::CMessagePassing
	CMessagePassing (CGraph *pGraph)
	Constructor. More...

virtual	~CMessagePassing (void)

virtual void	infer (unsigned int nIt=1)
	Inference. More...

Public Member Functions inherited from DirectGraphicalModels::CInfer
	CInfer (CGraph *pGraph)
	Constructor. More...

virtual	~CInfer (void)

virtual vec_byte_t	decode (unsigned int nIt=0, Mat &lossMatrix=Mat())
	Approximate decoding. More...

vec_float_t	getConfidence (void) const
	Returns the confidence of the prediction. More...

vec_float_t	getPotentials (byte state) const
	Returns the potnetials for the selected state (class) More...

Protected Member Functions
virtual void	calculateMessages (unsigned int nIt)
	Calculates messages, associated with the edges of corresponding graphical model. More...

Protected Member Functions inherited from DirectGraphicalModels::CMessagePassing
void	calculateMessage (Edge edge, float temp, float *&dst, bool maxSum=false)
	Calculates one message for the specified edge edge. More...

void	createMessages (void)
	Allocates memory for Edge::msg and Edge::msg_temp containers for all edges in the graph. More...

void	deleteMessages (void)
	Deletes memory for Edge::msg and Edge::msg_temp containers for all edges in the graph. More...

void	swapMessages (void)
	Swaps Edge::msg and Edge::msg_temp for all edges in the graph. More...

Protected Member Functions inherited from DirectGraphicalModels::CDecode
	CDecode (CGraph *pGraph)
	Constructor. More...

void	setState (vec_byte_t &state, qword configuration) const
	Sets the state according to the configuration index configuration. More...

void	incState (vec_byte_t &state) const
	Increases the state by one, i.e. switches the state array to the consequent configuration. More...

vec_float_t	calculatePotentials (void) const
	Calculates potentials for all possible configurations. More...

virtual	~CDecode (void)

virtual vec_byte_t	decode (unsigned int nIt=0, Mat &lossMatrix=Mat()) const
	Approximate decoding. More...

Protected Attributes
bool	m_maxSum
	Flag indicating weather the max-sum LBP (Viterbi algorithm) should be applied. More...

Protected Attributes inherited from DirectGraphicalModels::CDecode
CGraph *	m_pGraph
	Pointer to the graph. More...

Additional Inherited Members
Static Protected Member Functions inherited from DirectGraphicalModels::CMessagePassing
static float	MatMul (const Mat &M, const float v, float &dst, bool maxSum=false)
	Specific matrix multiplication. More...

Static Protected Member Functions inherited from DirectGraphicalModels::CDecode
static vec_byte_t	decode (const CGraph *pGraph, Mat &lossMatrix=Mat())
	Approximate decoding. More...

static Mat	getDefaultLossMatrix (byte nStates)
	Returns a default loss matrix \(L\). More...

Detailed Description

Sum product Loopy Belief Propagation inference class.

Author: Sergey G. Kosov, serge.nosp@m.y.ko.nosp@m.sov@p.nosp@m.roje.nosp@m.ct-10.nosp@m..de

Definition at line 16 of file InferLBP.h.

Constructor & Destructor Documentation

DirectGraphicalModels::CInferLBP::CInferLBP ( CGraph * pGraph )

inline

Constructor.

Parameters

pGraph The graph

Definition at line 23 of file InferLBP.h.

virtual DirectGraphicalModels::CInferLBP::~CInferLBP ( void )

inlinevirtual

Definition at line 24 of file InferLBP.h.

void DirectGraphicalModels::CInferLBP::calculateMessages ( unsigned int nIt )

protectedvirtual

Calculates messages, associated with the edges of corresponding graphical model.

This function may modify Edge::msg and Edge::msg_temp containers of graph edges

Parameters

nIt	Number of iterations

Definition at line 6 of file InferLBP.cpp.

Here is the call graph for this function:

bool DirectGraphicalModels::CInferLBP::m_maxSum

protected

Flag indicating weather the max-sum LBP (Viterbi algorithm) should be applied.

Definition at line 32 of file InferLBP.h.

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