gtsam::noiseModel::mEstimator::Base Class Reference

#include <LossFunctions.h>

Inheritance diagram for gtsam::noiseModel::mEstimator::Base:

Public Types
enum	ReweightScheme { Scalar, Block }
typedef std::shared_ptr< Base >	shared_ptr

Public Member Functions
	Base (const ReweightScheme reweight=Block)
virtual bool	equals (const Base &expected, double tol=1e-8) const =0
virtual double	loss (double distance) const
virtual void	print (const std::string &s) const =0
void	reweight (Vector &error) const
void	reweight (std::vector< Matrix > &A, Vector &error) const
void	reweight (Matrix &A, Vector &error) const
void	reweight (Matrix &A1, Matrix &A2, Vector &error) const
void	reweight (Matrix &A1, Matrix &A2, Matrix &A3, Vector &error) const
ReweightScheme	reweightScheme () const
	Returns the reweight scheme, as explained in ReweightScheme. More...
double	sqrtWeight (double distance) const
Vector	sqrtWeight (const Vector &error) const
virtual double	weight (double distance) const =0
Vector	weight (const Vector &error) const
virtual	~Base ()

Protected Attributes
ReweightScheme	reweight_
	Strategy for reweighting. More...

Detailed Description

Pure virtual class for all robust error function classes.

It provides the machinery for block vs scalar reweighting strategies, in addition to defining the interface of derived classes.

Definition at line 65 of file LossFunctions.h.

Member Typedef Documentation

shared_ptr

typedef std::shared_ptr<Base> gtsam::noiseModel::mEstimator::Base::shared_ptr

Definition at line 70 of file LossFunctions.h.

Member Enumeration Documentation

ReweightScheme

enum gtsam::noiseModel::mEstimator::Base::ReweightScheme

the rows can be weighted independently according to the error or uniformly with the norm of the right hand side

Enumerator
Scalar
Block

Definition at line 69 of file LossFunctions.h.

Constructor & Destructor Documentation

Base()

gtsam::noiseModel::mEstimator::Base::Base ( const ReweightScheme  reweight = Block )

inline

Definition at line 77 of file LossFunctions.h.

~Base()

virtual gtsam::noiseModel::mEstimator::Base::~Base ( )

inlinevirtual

Definition at line 78 of file LossFunctions.h.

Member Function Documentation

equals()

virtual bool gtsam::noiseModel::mEstimator::Base::equals ( const Base &  expected, double  tol = 1e-8  ) const

pure virtual

Implemented in gtsam::noiseModel::mEstimator::L2WithDeadZone, gtsam::noiseModel::mEstimator::DCS, gtsam::noiseModel::mEstimator::GemanMcClure, gtsam::noiseModel::mEstimator::Welsch, gtsam::noiseModel::mEstimator::Tukey, gtsam::noiseModel::mEstimator::Cauchy, gtsam::noiseModel::mEstimator::Huber, gtsam::noiseModel::mEstimator::Fair, and gtsam::noiseModel::mEstimator::Null.

loss()

virtual double gtsam::noiseModel::mEstimator::Base::loss ( double  distance ) const

inlinevirtual

This method is responsible for returning the total penalty for a given amount of error. For example, this method is responsible for implementing the quadratic function for an L2 penalty, the absolute value function for an L1 penalty, etc.

TODO(mikebosse): When the loss function has as input the norm of the error vector, then it prevents implementations of asymmeric loss functions. It would be better for this function to accept the vector and internally call the norm if necessary.

This returns (x) in mEstimator

Reimplemented in gtsam::noiseModel::mEstimator::L2WithDeadZone, gtsam::noiseModel::mEstimator::DCS, gtsam::noiseModel::mEstimator::GemanMcClure, gtsam::noiseModel::mEstimator::Welsch, gtsam::noiseModel::mEstimator::Tukey, gtsam::noiseModel::mEstimator::Cauchy, gtsam::noiseModel::mEstimator::Huber, gtsam::noiseModel::mEstimator::Fair, and gtsam::noiseModel::mEstimator::Null.

Definition at line 96 of file LossFunctions.h.

print()

virtual void gtsam::noiseModel::mEstimator::Base::print ( const std::string &  s ) const

pure virtual

Implemented in gtsam::noiseModel::mEstimator::L2WithDeadZone, gtsam::noiseModel::mEstimator::DCS, gtsam::noiseModel::mEstimator::GemanMcClure, gtsam::noiseModel::mEstimator::Welsch, gtsam::noiseModel::mEstimator::Tukey, gtsam::noiseModel::mEstimator::Cauchy, gtsam::noiseModel::mEstimator::Huber, gtsam::noiseModel::mEstimator::Fair, and gtsam::noiseModel::mEstimator::Null.

reweight() [1/5]

void Base::reweight

(

Vector &

error

)

const

reweight block matrices and a vector according to their weight implementation

Definition at line 50 of file LossFunctions.cpp.

reweight() [2/5]

void Base::reweight	(	std::vector< Matrix > &	A,
		Vector &	error
	)		const

Definition at line 60 of file LossFunctions.cpp.

reweight() [3/5]

void Base::reweight	(	Matrix &	A,
		Vector &	error
	)		const

Definition at line 78 of file LossFunctions.cpp.

reweight() [4/5]

void Base::reweight	(	Matrix &	A1,
		Matrix &	A2,
		Vector &	error
	)		const

Definition at line 92 of file LossFunctions.cpp.

reweight() [5/5]

void Base::reweight	(	Matrix &	A1,
		Matrix &	A2,
		Matrix &	A3,
		Vector &	error
	)		const

Definition at line 108 of file LossFunctions.cpp.

reweightScheme()

ReweightScheme gtsam::noiseModel::mEstimator::Base::reweightScheme ( ) const

inline

Returns the reweight scheme, as explained in ReweightScheme.

Definition at line 81 of file LossFunctions.h.

sqrtWeight() [1/2]

double gtsam::noiseModel::mEstimator::Base::sqrtWeight ( double  distance ) const

inline

Definition at line 113 of file LossFunctions.h.

sqrtWeight() [2/2]

Vector Base::sqrtWeight

(

const Vector &

error

)

const

square root version of the weight function

Definition at line 43 of file LossFunctions.cpp.

weight() [1/2]

virtual double gtsam::noiseModel::mEstimator::Base::weight ( double  distance ) const

pure virtual

This method is responsible for returning the weight function for a given amount of error. The weight function is related to the analytic derivative of the loss function. See https://members.loria.fr/MOBerger/Enseignement/Master2/Documents/ZhangIVC-97-01.pdf for details. This method is required when optimizing cost functions with robust penalties using iteratively re-weighted least squares.

This returns w(x) in mEstimator

Implemented in gtsam::noiseModel::mEstimator::L2WithDeadZone, gtsam::noiseModel::mEstimator::DCS, gtsam::noiseModel::mEstimator::GemanMcClure, gtsam::noiseModel::mEstimator::Welsch, gtsam::noiseModel::mEstimator::Tukey, gtsam::noiseModel::mEstimator::Cauchy, gtsam::noiseModel::mEstimator::Huber, gtsam::noiseModel::mEstimator::Fair, and gtsam::noiseModel::mEstimator::Null.

weight() [2/2]

Vector Base::weight

(

const Vector &

error

)

const

produce a weight vector according to an error vector and the implemented robust function

Definition at line 35 of file LossFunctions.cpp.

Member Data Documentation

reweight_

ReweightScheme gtsam::noiseModel::mEstimator::Base::reweight_

protected

Strategy for reweighting.

See also

ReweightScheme

Definition at line 74 of file LossFunctions.h.

---

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

• LossFunctions.h
• LossFunctions.cpp