ITK  4.2.0
Insight Segmentation and Registration Toolkit
Public Types | Static Public Member Functions
VNLSparseLUSolverTraits< T > Class Template Reference
Module ITKCommon

#include <VNLSparseLUSolverTraits.h>

List of all members.

Public Types

typedef vnl_sparse_matrix
< ValueType > 		MatrixType
typedef vnl_sparse_lu SolverType
typedef T ValueType
typedef vnl_vector< ValueType > VectorType

Static Public Member Functions

static void AddToMatrix (MatrixType &iA, const unsigned int &iR, const unsigned int &iC, const ValueType &iV)
static void FillMatrix (MatrixType &iA, const unsigned int &iR, const unsigned int &iC, const ValueType &iV)
static MatrixType InitializeSparseMatrix (const unsigned int &iN)
static MatrixType InitializeSparseMatrix (const unsigned int &iRow, const unsigned int &iCol)
static VectorType InitializeVector (const unsigned int &iN)
static bool IsDirectSolver ()
static bool Solve (const MatrixType &iA, const VectorType &iB, VectorType &oX)
static bool Solve (const MatrixType &iA, const VectorType &iBx, const VectorType &iBy, const VectorType &iBz, VectorType &oX, VectorType &oY, VectorType &oZ)
static bool Solve (const MatrixType &iA, const VectorType &iBx, const VectorType &iBy, VectorType &oX, VectorType &oY)

Detailed Description

template<typename T = double>
class VNLSparseLUSolverTraits< T >

Generic interface for sparse LU solver.

This generic interface (common to several sparse solvers), allow to interchange solver solutions when dealing with sparse linear systems. See itk::ParameterizationQuadEdgeMeshFilter for reference.

It internally uses the VNL library to represent and deal with vectors (vnl_vector) and sparse matrices (vnl_sparse_matrix). The solver by itself is made of a sparse LU decomposition followed by solving upper triangular system. see vnl_sparse_lu for more details on the method used.

See also:
VNLIterativeSparseSolverTraits:w

Definition at line 42 of file VNLSparseLUSolverTraits.h.

Member Typedef Documentation

template<typename T = double>
typedef vnl_sparse_matrix< ValueType > VNLSparseLUSolverTraits< T >::MatrixType

Definition at line 46 of file VNLSparseLUSolverTraits.h.

template<typename T = double>
typedef vnl_sparse_lu VNLSparseLUSolverTraits< T >::SolverType

Definition at line 48 of file VNLSparseLUSolverTraits.h.

template<typename T = double>
typedef T VNLSparseLUSolverTraits< T >::ValueType

Definition at line 45 of file VNLSparseLUSolverTraits.h.

template<typename T = double>
typedef vnl_vector< ValueType > VNLSparseLUSolverTraits< T >::VectorType

Definition at line 47 of file VNLSparseLUSolverTraits.h.

Member Function Documentation

template<typename T = double>
static void VNLSparseLUSolverTraits< T >::AddToMatrix ( MatrixType &  iA,
const unsigned int &  iR,
const unsigned int &  iC,
const ValueType &  iV 
)
inlinestatic

iA[iR][iC] += iV

Definition at line 81 of file VNLSparseLUSolverTraits.h.

template<typename T = double>
static void VNLSparseLUSolverTraits< T >::FillMatrix ( MatrixType &  iA,
const unsigned int &  iR,
const unsigned int &  iC,
const ValueType &  iV 
)
inlinestatic

iA[iR][iC] = iV

Definition at line 75 of file VNLSparseLUSolverTraits.h.

template<typename T = double>
static MatrixType VNLSparseLUSolverTraits< T >::InitializeSparseMatrix ( const unsigned int &  iN)
inlinestatic

initialize a square sparse matrix of size iN x iN

Definition at line 57 of file VNLSparseLUSolverTraits.h.

template<typename T = double>
static MatrixType VNLSparseLUSolverTraits< T >::InitializeSparseMatrix ( const unsigned int &  iRow,
const unsigned int &  iCol 
)
inlinestatic

initialize a sparse matrix of size iRow x iCol

Definition at line 63 of file VNLSparseLUSolverTraits.h.

template<typename T = double>
static VectorType VNLSparseLUSolverTraits< T >::InitializeVector ( const unsigned int &  iN)
inlinestatic

initialize a vector of size iN

Definition at line 69 of file VNLSparseLUSolverTraits.h.

template<typename T = double>
static bool VNLSparseLUSolverTraits< T >::IsDirectSolver ( )
inlinestatic
Returns:
false (it is not a direct solver, it is an iterative solver)

Definition at line 51 of file VNLSparseLUSolverTraits.h.

template<typename T = double>
static bool VNLSparseLUSolverTraits< T >::Solve ( const MatrixType &  iA,
const VectorType &  iB,
VectorType &  oX 
)
inlinestatic

Solve the linear system $ iA \cdot oX = iB $.

Definition at line 87 of file VNLSparseLUSolverTraits.h.

template<typename T = double>
static bool VNLSparseLUSolverTraits< T >::Solve ( const MatrixType &  iA,
const VectorType &  iBx,
const VectorType &  iBy,
const VectorType &  iBz,
VectorType &  oX,
VectorType &  oY,
VectorType &  oZ 
)
inlinestatic

Solve the linear systems: $ iA \cdot oX = iBx $, $ iA \cdot oY = iBy $, $ iA \cdot oZ = iBz $.

Definition at line 96 of file VNLSparseLUSolverTraits.h.

template<typename T = double>
static bool VNLSparseLUSolverTraits< T >::Solve ( const MatrixType &  iA,
const VectorType &  iBx,
const VectorType &  iBy,
VectorType &  oX,
VectorType &  oY 
)
inlinestatic

Solve the linear systems: $ iA \cdot oX = iBx $, $ iA \cdot oY = iBy $.

Definition at line 109 of file VNLSparseLUSolverTraits.h.

The documentation for this class was generated from the following file:
Tarballs of the nightly generated Doxygen documentation are available for the html, xml, and tag file.
Generated on Tue Jul 10 2012 23:50:08 for ITK by doxygen 1.8.1