CGAL::Shape_regularization::QP_regularization< GeomTraits, InputRange, NeighQuery, RegType, QPSolver > Class Template Reference

#include <CGAL/Shape_regularization/QP_regularization.h>

Definition

Shape regularization algorithm based on the quadratic programming global optimization.

Given a quadratic programming solver via the class QuadraticProgramTraits, this version of the shape regularization algorithm enables to regularize a set of user-specified geometric objects provided a way

• to access neighbors of each object via the NeighborQuery class;
• to obtain a maximum bound on a regularization characteristic (angle-orientation/ distance-offset/etc.) with respect to which an object is being regularized via the RegularizationType class;
• to obtain a target value for each pair of neighbor objects via the RegularizationType class.

This class is generic and forms a framework for different types of user-specified regularizations. Please refer to the user manual in order to learn how to use it. In particular, section Regularizing 2D Segments shows an example.

Template Parameters

GeomTraits	a model of Kernel
InputRange	a model of ConstRange
NeighQuery	a model of NeighborQuery
RegType	a model of RegularizationType
QPSolver	a model of QuadraticProgramTraits

Examples:

Shape_regularization/regularize_15_segments.cpp, and Shape_regularization/regularize_framework.cpp.

Initialization
template<typename NamedParameters = parameters::Default_named_parameters>
	QP_regularization (const InputRange &input_range, NeighQuery &neighbor_query, RegType &regularization_type, QPSolver &quadratic_program, const NamedParameters &np=parameters::default_values())
	initializes all internal data structures. More...

Regularization
void	regularize ()
	runs the shape regularization algorithm.

Memory Management
void	clear ()
	clears all internal data structures. More...
void	release_memory ()
	releases all memory that is used internally. More...

Constructor & Destructor Documentation

QP_regularization()

template<typename GeomTraits , typename InputRange , typename NeighQuery , typename RegType , typename QPSolver >

template<typename NamedParameters = parameters::Default_named_parameters>

CGAL::Shape_regularization::QP_regularization< GeomTraits, InputRange, NeighQuery, RegType, QPSolver >::QP_regularization	(	const InputRange &	input_range,
		NeighQuery &	neighbor_query,
		RegType &	regularization_type,
		QPSolver &	quadratic_program,
		const NamedParameters &	np = parameters::default_values()
	)

initializes all internal data structures.

Template Parameters

NamedParameters

a sequence of Named Parameters

Parameters

input_range	a const range of input objects for shape regularization
neighbor_query	an instance of NeighQuery that is used internally to access object neighbors
regularization_type	an instance of RegType that is used internally to obtain object bounds and target values
quadratic_program	an instance of QPSolver to solve the quadratic programming problem
np	an optional sequence of Named Parameters among the ones listed below; this parameter can be omitted, the default values are then used

Optional Named Parameters

geom_traits an instance of geometric traits class Type: a model of Kernel Default: GeomTraits()

Precondition

input_range.size() >= 2

Member Function Documentation

clear()

template<typename GeomTraits , typename InputRange , typename NeighQuery , typename RegType , typename QPSolver >

void CGAL::Shape_regularization::QP_regularization< GeomTraits, InputRange, NeighQuery, RegType, QPSolver >::clear

(

)

clears all internal data structures.

This method does not release the allocated memory but only removes the class internal data.

release_memory()

template<typename GeomTraits , typename InputRange , typename NeighQuery , typename RegType , typename QPSolver >

void CGAL::Shape_regularization::QP_regularization< GeomTraits, InputRange, NeighQuery, RegType, QPSolver >::release_memory

(

)

releases all memory that is used internally.

This method first calls the clear() method and then releases the allocated memory.