cpd Namespace Reference

Top-level cpd namespace. More...

Classes
class	Affine
	Affine coherent point drift. More...
struct	AffineResult
	The result of a affine coherent point drift run. More...
class	GaussTransform
	Abstract base class for Gauss transforms. More...
class	GaussTransformDirect
	The direct Gauss transform. More...
class	GaussTransformFgt
	The Gauss transform using the fgt library. More...
class	Nonrigid
	Nonrigid coherent point drift. More...
struct	NonrigidResult
	The result of a nonrigid coherent point drift run. More...
struct	Normalization
	The results of normalizing data to a unit cube (or whatever dimensionality). More...
struct	Probabilities
	Probability matrices produced by comparing two data sets with a GaussTransform. More...
struct	Result
	The result of a generic transform run. More...
class	Rigid
	Rigid coherent point drift. More...
struct	RigidResult
	The result of a rigid coherent point drift run. More...
class	Transform
	Generic coherent point drift transform. More...

Typedefs
typedef Eigen::MatrixXd	Matrix
	Our base matrix class. More...
typedef Eigen::VectorXd	Vector
	Typedef for our specific type of vector. More...
typedef Eigen::Matrix< Matrix::Index, Eigen::Dynamic, 1 >	IndexVector
	Typedef for an index vector, used to index other matrices. More...
typedef Eigen::ArrayXd	Array
	Typedef for our specific type of array. More...

Enumerations
enum	FgtMethod { DirectTree, Ifgt, Switched }
	The method(s) by which the fgt Gauss transform will compute the differences. More...

Functions
AffineResult	affine (const Matrix &fixed, const Matrix &moving)
	Runs a affine registration on two matrices. More...
Matrix	apply_transformation_matrix (Matrix points, const Matrix &transform)
	Apply a transformation matrix to a set of points. More...
NonrigidResult	nonrigid (const Matrix &fixed, const Matrix &moving)
	Runs a nonrigid registration on two matrices. More...
RigidResult	rigid (const Matrix &fixed, const Matrix &moving)
	Runs a rigid registration on two matrices. More...
Matrix	matrix_from_path (const std::string &path)
	Loads a matrix from a delimited text file. More...
double	default_sigma2 (const Matrix &fixed, const Matrix &moving)
	Computes the default sigma2 for the given matrices. More...
Matrix	affinity (const Matrix &fixed, const Matrix &moving, double beta)
	Computes the affinity matrix between the two matrices. More...
const char *	version ()
	Returns this library's version. More...

Variables
const FgtMethod	DEFAULT_FGT_METHOD = FgtMethod::DirectTree
	The default fgt method. More...
const double	DEFAULT_BREAKPOINT = 0.2
	The default switched fgt breakpoint. More...
const double	DEFAULT_EPSILON = 1e-4
	The default fgt epsilon. More...
const double	DEFAULT_BETA = 3.0
	Default value for beta. More...
const double	DEFAULT_LAMBDA = 3.0
	Default value for lambda. More...
const bool	DEFAULT_REFLECTIONS = false
	Should rigid registrations allow reflections by default? More...
const bool	DEFAULT_SCALE = !DEFAULT_LINKED
	Should rigid registrations scale the data by default? More...
const size_t	DEFAULT_MAX_ITERATIONS = 150
	The default number of iterations allowed. More...
const bool	DEFAULT_NORMALIZE = true
	Whether points should be normalized by default. More...
const double	DEFAULT_OUTLIERS = 0.1
	The default outlier weight. More...
const double	DEFAULT_TOLERANCE = 1e-5
	The default tolerance. More...
const double	DEFAULT_SIGMA2 = 0.0
	The default sigma2. More...
const bool	DEFAULT_CORRESPONDENCE = false
	Whether correspondence vector should be computed by default. More...
const bool	DEFAULT_LINKED = true
	Are the scalings of the two datasets linked by default? More...

Detailed Description

Top-level cpd namespace.

Typedef Documentation

typedef Eigen::ArrayXd cpd::Array

Typedef for our specific type of array.

typedef Eigen::Matrix<Matrix::Index, Eigen::Dynamic, 1> cpd::IndexVector

Typedef for an index vector, used to index other matrices.

typedef Eigen::MatrixXd cpd::Matrix

Our base matrix class.

typedef Eigen::VectorXd cpd::Vector

Typedef for our specific type of vector.

Enumeration Type Documentation

enum cpd::FgtMethod

The method(s) by which the fgt Gauss transform will compute the differences.

Enumerator
DirectTree	Use only the direct-tree fgt method.
Ifgt	Use only the improved fast Gauss transform method.
Switched	Switch between direct-tree and ifgt at a certain breakpoint.

Function Documentation

AffineResult cpd::affine	(	const Matrix &	fixed,
		const Matrix &	moving
	)

Runs a affine registration on two matrices.

Matrix cpd::affinity	(	const Matrix &	fixed,
		const Matrix &	moving,
		double	beta
	)

Computes the affinity matrix between the two matrices.

Matrix cpd::apply_transformation_matrix	(	Matrix	points,
		const Matrix &	transform
	)

Apply a transformation matrix to a set of points.

The transformation matrix should be one column wider than the point matrix.

double cpd::default_sigma2	(	const Matrix &	fixed,
		const Matrix &	moving
	)

Computes the default sigma2 for the given matrices.

Matrix cpd::matrix_from_path

(

const std::string &

path

)

Loads a matrix from a delimited text file.

NonrigidResult cpd::nonrigid	(	const Matrix &	fixed,
		const Matrix &	moving
	)

Runs a nonrigid registration on two matrices.

RigidResult cpd::rigid	(	const Matrix &	fixed,
		const Matrix &	moving
	)

Runs a rigid registration on two matrices.

const char* cpd::version

(

)

Returns this library's version.

Variable Documentation

const double cpd::DEFAULT_BETA = 3.0

Default value for beta.

const double cpd::DEFAULT_BREAKPOINT = 0.2

The default switched fgt breakpoint.

const bool cpd::DEFAULT_CORRESPONDENCE = false

Whether correspondence vector should be computed by default.

const double cpd::DEFAULT_EPSILON = 1e-4

The default fgt epsilon.

const FgtMethod cpd::DEFAULT_FGT_METHOD = FgtMethod::DirectTree

The default fgt method.

const double cpd::DEFAULT_LAMBDA = 3.0

Default value for lambda.

const bool cpd::DEFAULT_LINKED = true

Are the scalings of the two datasets linked by default?

const size_t cpd::DEFAULT_MAX_ITERATIONS = 150

The default number of iterations allowed.

const bool cpd::DEFAULT_NORMALIZE = true

Whether points should be normalized by default.

const double cpd::DEFAULT_OUTLIERS = 0.1

The default outlier weight.

const bool cpd::DEFAULT_REFLECTIONS = false

Should rigid registrations allow reflections by default?

const bool cpd::DEFAULT_SCALE = !DEFAULT_LINKED

Should rigid registrations scale the data by default?

const double cpd::DEFAULT_SIGMA2 = 0.0

The default sigma2.

const double cpd::DEFAULT_TOLERANCE = 1e-5

The default tolerance.