cpt::StateBase< N, T > Class Template Reference

Hold the state for a closest point transform. More...

#include <StateBase.h>

Inheritance diagram for cpt::StateBase< N, T >:

[legend]

Collaboration diagram for cpt::StateBase< N, T >:

[legend]

List of all members.

Public Member Functions
Constructors, etc.
	StateBase ()
	Default constructor.
	~StateBase ()
	Destructor.
Accessors.
std::size_t	getNumberOfGrids () const
	Return the number of grids.
const Grid &	getGrid (const std::size_t n) const
	Get the specified grid.
bool	hasBRepBeenSet () const
	Return true if the b-rep has been set.
const BBox &	getDomain () const
	Return the domain that contains all grids.
Number	getMaximumDistance () const
	Return how far the distance is being computed.
Manipulators.
void	setParameters (const BBox &domain, Number maximumDistance)
	Set the parameters for the Closest Point Transform.
void	setParameters (const Number *domainLowerData, const Number *domainUpperData, const Number maximumDistance)
	Set parameters for the closest point transform.
void	setParameters (const Number maximumDistance)
	Set the parameters for the Closest Point Transform.
void	setLattice (const SizeList &extents, const BBox &domain)
	Set the grid geometry.
void	setLattice (const int *extentsData, const Number *lowerData, const Number *upperData)
	Set the lattice geometry.
void	insertGrid (const SizeList &extents, const IndexList &bases, bool useGradientOfDistance, bool useClosestPoint, bool useClosestFace)
	Add a grid for the Closest Point Transform.
void	insertGrid (const SizeList &extents, const IndexList &bases, Number *distance, Number *gradientOfDistance, Number *closestPoint, int *closestFace)
	Add a grid for the Closest Point Transform.
void	insertGrid (const int *lowerBounds, const int *upperBounds, Number *distance, Number *gradientOfDistance, Number *closestPoint, int *closestFace)
	Add a grid for the Closest Point Transform.
void	clearGrids ()
	Clear the grids.
std::pair< std::size_t, std::size_t >	computeClosestPointTransformUsingBBox ()
	Compute the closest point transform for signed distance.
std::pair< std::size_t, std::size_t >	computeClosestPointTransformUsingBruteForce ()
	Compute the closest point transform for signed distance.
std::pair< std::size_t, std::size_t >	computeClosestPointTransformUsingTree ()
	Compute the closest point transform for signed distance.
std::pair< std::size_t, std::size_t >	computeClosestPointTransformUnsignedUsingBBox ()
	Compute the closest point transform for unsigned distance.
std::pair< std::size_t, std::size_t >	computeClosestPointTransformUnsignedUsingBruteForce ()
	Compute the closest point transform for unsigned distance.
void	floodFillAtBoundary (Number farAway)
	Flood fill the distance.
void	floodFillDetermineSign (Number farAway)
	Flood fill the distance.
void	floodFillUnsigned (Number farAway)
	Flood fill the unsigned distance.
bool	areGridsValid ()
	Check the grids.
bool	areGridsValidUnsigned ()
	Check the grids.
void	setBRepWithNoClipping (const std::vector< std::tr1::array< Number, N > > &vertices, const std::vector< std::tr1::array< std::size_t, N > > &faces)
	Set the b-rep.
void	setBRepWithNoClipping (std::size_t numVertices, const Number *vertices, std::size_t numFaces, const int *faces)
	Wrapper for the above function.
void	setBRep (const std::vector< std::tr1::array< Number, N > > &vertices, const std::vector< std::tr1::array< std::size_t, N > > &faces)
	Set the b-rep.
void	setBRep (std::size_t numVertices, const Number *vertices, std::size_t numFaces, const int *faces)
	Wrapper for the above function.
I/O.
void	displayInformation (std::ostream &out) const
	Display information about the state of the closest point transform.
Protected Types
typedef T	Number
	The number type.
typedef geom::BBox< N, Number >	BBox
	A bounding box.
typedef geom::RegularGrid< N, T >	Lattice
	The lattice defines a domain and index extents.
typedef cpt::Grid< N, T >	Grid
	The grid.
typedef cpt::BRep< N, T >	BRep
	The b-rep.
typedef Grid::Point	Point
	A point in N-D.
typedef Grid::SizeList	SizeList
	An extent in N-D.
typedef Grid::IndexList	IndexList
	A multi-index in N-D.
typedef Grid::IndexList	IndexedFace
	The indices of a face.
typedef Grid::Index	Index
	A single index.
typedef Grid::Range	Range
	A multi-index range in N-D.
Protected Member Functions
Grid operations.
void	initializeGrids ()
	Initialize the grids.
Protected Attributes
bool	_hasBRepBeenSet
	Has the b-rep been set.
bool	_hasCptBeenComputed
	Has the CPT been computed.
BBox	_domain
	The domain containing all grids for which the CPT will be computed.
Number	_maximumDistance
	How far (in Cartesian space) to compute the distance.
Lattice	_lattice
	The lattice.
std::vector< Grid >	_grids
	The grids.
BRep	_brep
	The b-rep.

---

Detailed Description

template<std::size_t N, typename T = double>
class cpt::StateBase< N, T >

Hold the state for a closest point transform.

Implements the dimension-independent functionality.

---

Member Function Documentation

template<std::size_t N, typename T = double>

bool cpt::StateBase< N, T >::areGridsValid

(

)

Check the grids.

Verify that the distance grids are valid. The known distances should be between +-maximumDistance. The difference between adjacent grid points should not be more than the grid spacing. Verify that the closest point and closest face grids are valid. Return true if the grids are valid. Return false and print a message to stderr otherwise.

template<std::size_t N, typename T = double>

bool cpt::StateBase< N, T >::areGridsValidUnsigned

(

)

Check the grids.

Verify that the distance grids are valid. The known distances should be between 0 and maximumDistance. The difference between adjacent grid points should not be more than the grid spacing. Verify that the closest point and closest face grids are valid. Return true if the grids are valid. Return false and print a message to stderr otherwise.

template<std::size_t N, typename T = double>

void cpt::StateBase< N, T >::clearGrids

(

)

[inline]

Clear the grids.

If the grids change, call this function and then add all the new grids with insertGrid().

template<std::size_t N, typename T = double>

std::pair<std::size_t, std::size_t> cpt::StateBase< N, T >::computeClosestPointTransformUnsignedUsingBBox

(

)

Compute the closest point transform for unsigned distance.

Compute the unsigned distance. Compute the gradient of the distance, the closest face and closest point if their arrays specified in insertGrid() are nonzero.

This algorithm does not use polyhedron scan conversion. Instead, it builds bounding boxes around the characteristic polyhedra.

The unknown distances, gradients, and closest points are set to std::numeric_limits<Number>::max(). The unknown closest faces are set to std::numeric_limits<std::size_t>::max().

Returns:

Return the number of points for which the distance was computed and the number of points for which the distance was set.

template<std::size_t N, typename T = double>

std::pair<std::size_t, std::size_t> cpt::StateBase< N, T >::computeClosestPointTransformUnsignedUsingBruteForce

(

)

Compute the closest point transform for unsigned distance.

Compute the unsigned distance. Compute the gradient of the distance, the closest face and closest point if their arrays specified in insertGrid() are nonzero.

This algorithm does not use polyhedron scan conversion or the characteristic polyhedra. Instead, it builds bounding boxes around the faces, edges and vertices.

The unknown distances, gradients, and closest points are set to std::numeric_limits<Number>::max(). The unknown closest faces are set to std::numeric_limits<std::size_t>::max().

Returns:

Return the number of points for which the distance was computed and the number of points for which the distance was set.

template<std::size_t N, typename T = double>

std::pair<std::size_t, std::size_t> cpt::StateBase< N, T >::computeClosestPointTransformUsingBBox

(

)

Compute the closest point transform for signed distance.

Compute the signed distance. Compute the gradient of the distance, the closest face and closest point if their arrays specified in insertGrid() are nonzero.

This algorithm does not use polyhedron scan conversion. Instead, it builds bounding boxes around the characteristic polyhedra.

The unknown distances, gradients, and closest points are set to std::numeric_limits<Number>::max(). The unknown closest faces are set to std::numeric_limits<std::size_t>::max().

Returns:

Return the number of points for which the distance was computed and the number of points for which the distance was set.

template<std::size_t N, typename T = double>

std::pair<std::size_t, std::size_t> cpt::StateBase< N, T >::computeClosestPointTransformUsingBruteForce

(

)

Compute the closest point transform for signed distance.

Compute the signed distance. Compute the gradient of the distance, the closest face and closest point if their arrays specified in insertGrid() are nonzero.

This algorithm does not use polyhedron scan conversion or the characteristic polyhedra. Instead, it builds bounding boxes around the faces, edges and vertices.

The unknown distances, gradients, and closest points are set to std::numeric_limits<Number>::max(). The unknown closest faces are set to std::numeric_limits<std::size_t>::max().

Returns:

Return the number of points for which the distance was computed and the number of points for which the distance was set.

template<std::size_t N, typename T = double>

std::pair<std::size_t, std::size_t> cpt::StateBase< N, T >::computeClosestPointTransformUsingTree

(

)

Compute the closest point transform for signed distance.

Compute the signed distance. Compute the gradient of the distance, the closest face and closest point if their arrays specified in insertGrid() are nonzero.

This algorithm uses a bounding box tree to store the mesh and a lower-upper-bound queries to determine the distance.

The unknown distances, gradients, and closest points are set to std::numeric_limits<Number>::max(). The unknown closest faces are set to std::numeric_limits<std::size_t>::max().

Returns:

Return the number of points for which the distance was computed and the number of points for which the distance was set.

template<std::size_t N, typename T = double>

void cpt::StateBase< N, T >::floodFillAtBoundary

(

Number

farAway

)

Flood fill the distance.

This function is used to prepare the distance for visualization. The signed distance is flood filled. If any of the distances are known in a particular grid, set the unknown distances to +-farAway. If no distances are known, set all distances to +farAway. Thus note that if no points in a particular grid have known distance, then the sign of the distance is not determined.

template<std::size_t N, typename T = double>

void cpt::StateBase< N, T >::floodFillDetermineSign

(

Number

farAway

)

Flood fill the distance.

This function is used to prepare the distance for visualization. The signed distance is flood filled. If any of the distances are known in a particular grid, set the unknown distances to +-farAway. If no distances are known, determine the correct sign by computing the signed distance to the boundary for a single point in the grid.

Note:

In order to determine the sign of the distance for far away grids, this algorithm needs that portion of the boundary that is closest to those grids. Using setBRep() may clip away the needed portion (or for that matter all) of the boundary. You should use setBRepWithNoClipping() before calling this function. To be on the safe side, you should give the entire boundary to setBRepWithNoClipping() .

template<std::size_t N, typename T = double>

void cpt::StateBase< N, T >::floodFillUnsigned

(

Number

farAway

)

Flood fill the unsigned distance.

The unsigned distance is flood filled. Unknown distances are set to farAway.

template<std::size_t N, typename T = double>

void cpt::StateBase< N, T >::insertGrid	(	const int *	lowerBounds,
		const int *	upperBounds,
		Number *	distance,
		Number *	gradientOfDistance,
		Number *	closestPoint,
		int *	closestFace
	)			[inline]

Add a grid for the Closest Point Transform.

This is a wrapper for the above insertGrid function().

template<std::size_t N, typename T = double>

void cpt::StateBase< N, T >::insertGrid	(	const SizeList &	extents,
		const IndexList &	bases,
		Number *	distance,
		Number *	gradientOfDistance,
		Number *	closestPoint,
		int *	closestFace
	)

Add a grid for the Closest Point Transform.

Parameters:

	extents	The array extents.
	bases	The index bases for the arrays. For each of the gradientOfDistance, closestPoint and closestFace arrays: if the pointer is non-zero, that quantity will be computed.
	distance	Pointer to the distance array.
	gradientOfDistance	Pointer to the gradient of distance array.
	closestPoint	Pointer to the closest point array.
	closestFace	Pointer to the closest face array.

template<std::size_t N, typename T = double>

void cpt::StateBase< N, T >::insertGrid	(	const SizeList &	extents,
		const IndexList &	bases,
		bool	useGradientOfDistance,
		bool	useClosestPoint,
		bool	useClosestFace
	)

Add a grid for the Closest Point Transform.

This function must be called at least once before calls to computeClosestPointTransform().

Parameters:

	extents	The array extents.
	bases	The index bases for the arrays.
	useGradientOfDistance	Whether to compute the gradient of the distance. The gradient of the distance is computed from the geometric primitives and not by differencing the distance array. Thus it is accurate to within machine precision.
	useClosestPoint	Whether to compute the closest points.
	useClosestFace	Whether to compute the closet face on the mesh.

Referenced by cpt::StateBase< 2, T >::insertGrid().

template<std::size_t N, typename T = double>

void cpt::StateBase< N, T >::setBRep	(	const std::vector< std::tr1::array< Number, N > > &	vertices,
		const std::vector< std::tr1::array< std::size_t, N > > &	faces
	)

Set the b-rep.

Clip the mesh to use only points that affect the cartesian domain.

Either this function or setBRepWithNoClipping() must be called at least once before calls to computeClosestPointTransform(). This version is more efficient if the b-rep extends beyond the domain spanned by the grid.

Parameters:

	vertices	are the locations of the vertices.
	faces	is a vector of tuples of vertex indices that describe the faces.

template<std::size_t N, typename T = double>

void cpt::StateBase< N, T >::setBRepWithNoClipping	(	const std::vector< std::tr1::array< Number, N > > &	vertices,
		const std::vector< std::tr1::array< std::size_t, N > > &	faces
	)

Set the b-rep.

Do not use the Cartesian domain to clip the mesh.

Either this function or setBRep() must be called at least once before calls to computeClosestPointTransform().

Parameters:

	vertices	are the locations of the vertices.
	faces	is a vector of tuples of vertex indices that describe the faces.

template<std::size_t N, typename T = double>

void cpt::StateBase< N, T >::setLattice	(	const int *	extentsData,
		const Number *	lowerData,
		const Number *	upperData
	)			[inline]

Set the lattice geometry.

This is a wrapper for the above setLattice function().

template<std::size_t N, typename T = double>

void cpt::StateBase< N, T >::setLattice	(	const SizeList &	extents,
		const BBox &	domain
	)

Set the grid geometry.

Parameters:

	extents	are the grid geometry extents.
	domain	is the grid geometry domain.

Referenced by cpt::StateBase< 2, T >::setLattice().

template<std::size_t N, typename T = double>

void cpt::StateBase< N, T >::setParameters

(

const Number

maximumDistance

)

[inline]

Set the parameters for the Closest Point Transform.

This calls the first setParameters() function. The domain containing all grids is set to all space. This means that clipping the mesh will have no effect.

template<std::size_t N, typename T = double>

void cpt::StateBase< N, T >::setParameters	(	const Number *	domainLowerData,
		const Number *	domainUpperData,
		const Number	maximumDistance
	)			[inline]

Set parameters for the closest point transform.

This is a wrapper for the above setParameters function.

template<std::size_t N, typename T = double>

void cpt::StateBase< N, T >::setParameters	(	const BBox &	domain,
		Number	maximumDistance
	)

Set the parameters for the Closest Point Transform.

This function must be called at least once before calls to computeClosestPointTransform().

Parameters:

	domain	is the Cartesian domain that contains all grids.
	maximumDistance	The distance will be computed up to maximumDistance away from the surface.

The distance for grid points whose distance is larger than maximumDistance will be set to std::numeric_limits<Number>::max(). Each component of the closest point of these far away points will be set to std::numeric_limits<Number>::max(). The closest face of far away points will be set to std::numeric_limits<std::size_t>::max().

Referenced by cpt::StateBase< 2, T >::setParameters().

---

Member Data Documentation

template<std::size_t N, typename T = double>

BBox cpt::StateBase< N, T >::_domain [protected]

The domain containing all grids for which the CPT will be computed.

This may be used in clipping the mesh.

Referenced by cpt::StateBase< 2, T >::getDomain().

---

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

• StateBase.h