Convergence criteria of GlobalOptimization. More...

#include <GlobalOptimizationConvergenceCriteria.h>

Public Member Functions
	GlobalOptimizationConvergenceCriteria (int max_iteration=100, double min_relative_increment=1e-6, double min_relative_residual_increment=1e-6, double min_right_term=1e-6, double min_residual=1e-6, int max_iteration_lm=20, double upper_scale_factor=2./3., double lower_scale_factor=1./3.)
	Parameterized Constructor. More...

	~GlobalOptimizationConvergenceCriteria ()

Public Attributes
int	max_iteration_
	Maximum iteration number for iterative optimization module. More...

double	min_relative_increment_
	Minimum relative increments. More...

double	min_relative_residual_increment_
	Minimum relative residual increments. More...

double	min_right_term_
	Minimum right term value. More...

double	min_residual_
	Minimum residual value. More...

int	max_iteration_lm_
	Maximum iteration number for Levenberg Marquardt method. More...

double	upper_scale_factor_
	Upper scale factor value. More...

double	lower_scale_factor_
	Lower scale factor value. More...

Detailed Description

Convergence criteria of GlobalOptimization.

Definition at line 72 of file GlobalOptimizationConvergenceCriteria.h.

Constructor & Destructor Documentation

◆ GlobalOptimizationConvergenceCriteria()

cloudViewer::pipelines::registration::GlobalOptimizationConvergenceCriteria::GlobalOptimizationConvergenceCriteria	(	int	max_iteration = `100`,
		double	min_relative_increment = `1e-6`,
		double	min_relative_residual_increment = `1e-6`,
		double	min_right_term = `1e-6`,
		double	min_residual = `1e-6`,
		int	max_iteration_lm = `20`,
		double	upper_scale_factor = `2. / 3.`,
		double	lower_scale_factor = `1. / 3.`
	)

inline

Parameterized Constructor.

Parameters

max_iteration	Maximum iteration number.
min_relative_increment	Minimum relative increments.
min_relative_residual_increment	Minimum relative residual increments.
min_right_term	Minimum right term value.
min_residual	Minimum residual value.
max_iteration_lm	Maximum iteration number for Levenberg Marquardt method.
upper_scale_factor	Upper scale factor value.
lower_scale_factor	Lower scale factor value.

Definition at line 84 of file GlobalOptimizationConvergenceCriteria.h.

References lower_scale_factor_, and upper_scale_factor_.

◆ ~GlobalOptimizationConvergenceCriteria()

cloudViewer::pipelines::registration::GlobalOptimizationConvergenceCriteria::~GlobalOptimizationConvergenceCriteria ( )

inline

Definition at line 110 of file GlobalOptimizationConvergenceCriteria.h.

Member Data Documentation

◆ lower_scale_factor_

double cloudViewer::pipelines::registration::GlobalOptimizationConvergenceCriteria::lower_scale_factor_

Lower scale factor value.

Definition at line 138 of file GlobalOptimizationConvergenceCriteria.h.

Referenced by GlobalOptimizationConvergenceCriteria(), cloudViewer::pipelines::registration::GlobalOptimizationLevenbergMarquardt::OptimizePoseGraph(), and cloudViewer::pipelines::registration::pybind_global_optimization().

◆ max_iteration_

int cloudViewer::pipelines::registration::GlobalOptimizationConvergenceCriteria::max_iteration_

Maximum iteration number for iterative optimization module.

Definition at line 114 of file GlobalOptimizationConvergenceCriteria.h.

Referenced by cloudViewer::pipelines::registration::CheckMaxIteration(), and cloudViewer::pipelines::registration::pybind_global_optimization().

◆ max_iteration_lm_

int cloudViewer::pipelines::registration::GlobalOptimizationConvergenceCriteria::max_iteration_lm_

Maximum iteration number for Levenberg Marquardt method.

max_iteration_lm_ is used for additional Levenberg-Marquardt inner loop that automatically changes steepest gradient gain.

Definition at line 130 of file GlobalOptimizationConvergenceCriteria.h.

Referenced by cloudViewer::pipelines::registration::CheckMaxIterationLM(), and cloudViewer::pipelines::registration::pybind_global_optimization().

◆ min_relative_increment_

double cloudViewer::pipelines::registration::GlobalOptimizationConvergenceCriteria::min_relative_increment_

Minimum relative increments.

Several convergence criteria to determine stability of iterative optimization.

Definition at line 119 of file GlobalOptimizationConvergenceCriteria.h.

Referenced by cloudViewer::pipelines::registration::CheckRelativeIncrement(), and cloudViewer::pipelines::registration::pybind_global_optimization().

◆ min_relative_residual_increment_

double cloudViewer::pipelines::registration::GlobalOptimizationConvergenceCriteria::min_relative_residual_increment_

Minimum relative residual increments.

Definition at line 121 of file GlobalOptimizationConvergenceCriteria.h.

Referenced by cloudViewer::pipelines::registration::CheckRelativeResidualIncrement(), and cloudViewer::pipelines::registration::pybind_global_optimization().

◆ min_residual_

double cloudViewer::pipelines::registration::GlobalOptimizationConvergenceCriteria::min_residual_

Minimum residual value.

Definition at line 125 of file GlobalOptimizationConvergenceCriteria.h.

Referenced by cloudViewer::pipelines::registration::CheckResidual(), and cloudViewer::pipelines::registration::pybind_global_optimization().

◆ min_right_term_

double cloudViewer::pipelines::registration::GlobalOptimizationConvergenceCriteria::min_right_term_

Minimum right term value.

Definition at line 123 of file GlobalOptimizationConvergenceCriteria.h.

Referenced by cloudViewer::pipelines::registration::CheckRightTerm(), and cloudViewer::pipelines::registration::pybind_global_optimization().

◆ upper_scale_factor_

double cloudViewer::pipelines::registration::GlobalOptimizationConvergenceCriteria::upper_scale_factor_

Upper scale factor value.

Scaling factors are used for levenberg marquardt algorithm these are scaling factors that increase/decrease lambda used in H_LM = H + lambda * I

Definition at line 136 of file GlobalOptimizationConvergenceCriteria.h.