Class that computes the covariance matrices for every point in a Surface used in the A-ICP algorithm. More...

#include <mitkCovarianceMatrixCalculator.h>

Inheritance diagram for mitk::CovarianceMatrixCalculator:

Collaboration diagram for mitk::CovarianceMatrixCalculator:

Public Member Functions
	mitkClassMacroItkParent (CovarianceMatrixCalculator, itk::Object) static Pointer New()

Pointer	Clone () const

void	SetVoronoiScalingFator (const double factor)

void	EnableNormalization (bool state)

double	GetMeanVariance () const

const CovarianceMatrixList &	GetCovarianceMatrices () const

void	SetInputSurface (Surface *input)

void	ComputeCovarianceMatrices ()

Protected Types
typedef itk::Matrix< double, 3, 3 >	CovarianceMatrix

typedef std::vector< CovarianceMatrix >	CovarianceMatrixList

typedef double	Vertex[3]

Protected Member Functions
void	ComputeOrthonormalCoordinateSystem (const int index, Vertex normal, CovarianceMatrix &principalComponents, Vertex variances, Vertex curVertex)

	CovarianceMatrixCalculator ()

	~CovarianceMatrixCalculator ()

Protected Attributes
CovarianceMatrixList	m_CovarianceMatrixList

Detailed Description

Class that computes the covariance matrices for every point in a Surface used in the A-ICP algorithm.

Computes a covariance matrix for every vertex in a given Surface based on it's direct neighbours and saves them into a CovarianceMatrixList. The Class implements the CM_PCA method presented by L. Maier-Hein et al. in "Convergent Iterative Closest-Point Algorithm to Accomodate Anisotropic and Inhomogenous Localization Error.", IEEE T Pattern Anal 34 (8), 1520-1532, 2012. The algorithm needs a clean Surface with non manifold edges and no duplicated vertices. To ensure a clean Surface representation use vtkCleanPolyData.

Definition at line 50 of file mitkCovarianceMatrixCalculator.h.

Member Typedef Documentation

typedef itk::Matrix<double, 3, 3> mitk::CovarianceMatrixCalculator::CovarianceMatrix

protected

Definition of the covariance matrix.

Definition at line 60 of file mitkCovarianceMatrixCalculator.h.

typedef std::vector<CovarianceMatrix> mitk::CovarianceMatrixCalculator::CovarianceMatrixList

protected

Definition of a list of covariance matrices

Definition at line 62 of file mitkCovarianceMatrixCalculator.h.

typedef double mitk::CovarianceMatrixCalculator::Vertex[3]

protected

Definition at line 63 of file mitkCovarianceMatrixCalculator.h.

Constructor & Destructor Documentation

mitk::CovarianceMatrixCalculator::CovarianceMatrixCalculator ( )

protected

Definition at line 71 of file mitkCovarianceMatrixCalculator.cpp.

mitk::CovarianceMatrixCalculator::~CovarianceMatrixCalculator ( )

protected

Definition at line 75 of file mitkCovarianceMatrixCalculator.cpp.

Member Function Documentation

Pointer mitk::CovarianceMatrixCalculator::Clone ( ) const

void mitk::CovarianceMatrixCalculator::ComputeCovarianceMatrices ( )

Method that computes the covariance matrices for the input surface.

Exceptions

std::exception If the input surface is not set.

Definition at line 106 of file mitkCovarianceMatrixCalculator.cpp.

References ComputeCovarianceMatrix(), and mitkThrow.

void mitk::CovarianceMatrixCalculator::ComputeOrthonormalCoordinateSystem	(	const int	index,
		Vertex	normal,
		CovarianceMatrix &	principalComponents,
		Vertex	variances,
		Vertex	curVertex
	)

protected

This method projects all surrounding vertices of given vertex in a Surface in the normal direction onto a plane and computes a primary component analysis on the projected vertices. In the next step a orthonormal system is created.

Parameters

The	index of the input Vertex in the Surface.
The	normal of the input Vertex.
Output	CovarianceMatrix of the principal component analysis.
Output.	Variances along the axes of the createt Orthonormal system.
Output.	The current Vertex in the surface