itk::Statistics::KdTree< TSample > Class Template Reference

This class provides methods for k-nearest neighbor search and related data structures for a k-d tree. More...

#include <itkKdTree.h>

Inheritance diagram for itk::Statistics::KdTree:

[legend]Collaboration diagram for itk::Statistics::KdTree< TSample >:

[legend]List of all members.

[NOHEADER]
int	SearchLoop (KdTreeNodeType *node, MeasurementVectorType &query, MeasurementVectorType &lowerBound, MeasurementVectorType &upperBound)
Public Types
typedef KdTree	Self
typedef Object	Superclass
typedef SmartPointer< Self >	Pointer
typedef TSample	SampleType
typedef TSample::MeasurementVectorType	MeasurementVectorType
typedef TSample::MeasurementType	MeasurementType
typedef TSample::InstanceIdentifier	InstanceIdentifier
typedef TSample::FrequencyType	FrequencyType
typedef EuclideanDistance< MeasurementVectorType >	DistanceMetricType
typedef KdTreeNode< TSample >	KdTreeNodeType
typedef std::pair< InstanceIdentifier, double >	NeighborType
typedef std::vector< InstanceIdentifier >	InstanceIdentifierVectorType
typedef TSample::Iterator	Iterator
Public Methods
virtual const char *	GetClassName () const
	itkStaticConstMacro (MeasurementVectorSize, unsigned int, TSample::MeasurementVectorSize)
void	SetBucketSize (unsigned int size)
void	SetSample (TSample *sample)
TSample *	GetSample ()
unsigned long	Size ()
KdTreeNodeType *	GetEmptyTerminalNode ()
void	SetRoot (KdTreeNodeType *root)
KdTreeNodeType *	GetRoot ()
MeasurementVectorType	GetMeasurementVector (InstanceIdentifier id)
FrequencyType	GetFrequency (InstanceIdentifier id)
DistanceMetricType *	GetDistanceMetric ()
void	Search (MeasurementVectorType &query, unsigned int k, InstanceIdentifierVectorType &result)
void	Search (MeasurementVectorType &query, double radius, InstanceIdentifierVectorType &result)
int	GetNumberOfVisits ()
bool	BallWithinBounds (MeasurementVectorType &query, MeasurementVectorType &lowerBound, MeasurementVectorType &upperBound, double radius)
bool	BoundsOverlapBall (MeasurementVectorType &query, MeasurementVectorType &lowerBound, MeasurementVectorType &upperBound, double radius)
void	DeleteNode (KdTreeNodeType *node)
void	PrintTree (KdTreeNodeType *node, int level, unsigned int activeDimension)
Iterator	Begin ()
Iterator	End ()
Static Public Methods
Pointer	New ()
Protected Methods
	KdTree ()
virtual	~KdTree ()
void	PrintSelf (std::ostream &os, Indent indent) const
int	NearestNeighborSearchLoop (KdTreeNodeType *node, MeasurementVectorType &query, MeasurementVectorType &lowerBound, MeasurementVectorType &upperBound)

---

Detailed Description

template<class TSample>
class itk::Statistics::KdTree< TSample >

This class provides methods for k-nearest neighbor search and related data structures for a k-d tree.

An object of this class stores instance identifiers in a k-d tree that is a binary tree with childrens split along a dimension among k-dimensions. The dimension of the split (or partition) is determined for each nonterminal node that has two children. The split process is terminated when the node has no children (when the number of measurement vectors is less than or equal to the size set by the SetBucketSize. That is The split process is a recursive process in nature and in implementation. This implementation doesn't support dynamic insert and delete operations for the tree. Instead, we can use the KdTreeGenerator or WeightedCentroidKdTreeGenerator to generate a static KdTree object.

To search k-nearest neighbor, call the Search method with the query point in a k-d space and the number of nearest neighbors. The GetSearchResult method returns a pointer to a NearestNeighbors object with k-nearest neighbors.

See also:

KdTreeNode, KdTreeNonterminalNode, KdTreeWeightedCentroidNonterminalNode, KdTreeTerminalNode, KdTreeGenerator, WeightedCentroidKdTreeNode

Definition at line 322 of file itkKdTree.h.

---

Member Typedef Documentation

template<class TSample> typedef EuclideanDistance< MeasurementVectorType > itk::Statistics::KdTree< TSample >::DistanceMetricType

DistanceMetric type for the distance calculation and comparison Definition at line 348 of file itkKdTree.h.

template<class TSample> typedef TSample::FrequencyType itk::Statistics::KdTree< TSample >::FrequencyType

Definition at line 341 of file itkKdTree.h.

template<class TSample> typedef TSample::InstanceIdentifier itk::Statistics::KdTree< TSample >::InstanceIdentifier

Definition at line 340 of file itkKdTree.h.

template<class TSample> typedef std::vector< InstanceIdentifier > itk::Statistics::KdTree< TSample >::InstanceIdentifierVectorType

Definition at line 358 of file itkKdTree.h.

template<class TSample> typedef TSample::Iterator itk::Statistics::KdTree< TSample >::Iterator

Definition at line 523 of file itkKdTree.h.

template<class TSample> typedef KdTreeNode< TSample > itk::Statistics::KdTree< TSample >::KdTreeNodeType

Node type of the KdTree Definition at line 351 of file itkKdTree.h.

template<class TSample> typedef TSample::MeasurementType itk::Statistics::KdTree< TSample >::MeasurementType

Definition at line 339 of file itkKdTree.h.

template<class TSample> typedef TSample::MeasurementVectorType itk::Statistics::KdTree< TSample >::MeasurementVectorType

Definition at line 338 of file itkKdTree.h.

template<class TSample> typedef std::pair< InstanceIdentifier, double > itk::Statistics::KdTree< TSample >::NeighborType

Neighbor type. The first element of the std::pair is the instance identifier and the second one is the distance between the measurement vector identified by the first element and the query point. Definition at line 356 of file itkKdTree.h.

template<class TSample> typedef SmartPointer<Self> itk::Statistics::KdTree< TSample >::Pointer

Reimplemented from itk::Object. Definition at line 328 of file itkKdTree.h.

template<class TSample> typedef TSample itk::Statistics::KdTree< TSample >::SampleType

typedef alias for the source data container Definition at line 337 of file itkKdTree.h.

template<class TSample> typedef KdTree itk::Statistics::KdTree< TSample >::Self

Standard class typedefs Reimplemented from itk::Object. Definition at line 326 of file itkKdTree.h.

template<class TSample> typedef Object itk::Statistics::KdTree< TSample >::Superclass

Reimplemented from itk::Object. Definition at line 327 of file itkKdTree.h.

---

Constructor & Destructor Documentation

template<class TSample> itk::Statistics::KdTree< TSample >::KdTree (    )  [protected]

Constructor

template<class TSample> virtual itk::Statistics::KdTree< TSample >::~KdTree (    )  [protected, virtual]

Destructor: deletes the root node and the empty terminal node.

---

Member Function Documentation

template<class TSample> bool itk::Statistics::KdTree< TSample >::BallWithinBounds (  MeasurementVectorType &    query, MeasurementVectorType &    lowerBound, MeasurementVectorType &    upperBound, double    radius )

Returns true if the intermediate k-nearest neighbors exist within the the bounding box defined by the lowerBound and the upperBound. Otherwise returns false. Returns false if the ball defined by the distance between the query point and the farthest neighbor touch the surface of the bounding box.

template<class TSample> Iterator itk::Statistics::KdTree< TSample >::Begin (  void    )  [inline]

Definition at line 525 of file itkKdTree.h.

template<class TSample> bool itk::Statistics::KdTree< TSample >::BoundsOverlapBall (  MeasurementVectorType &    query, MeasurementVectorType &    lowerBound, MeasurementVectorType &    upperBound, double    radius )

Returns true if the ball defined by the distance between the query point and the farthest neighbor overlaps with the bounding box defined by the lower and the upper bounds.

template<class TSample> void itk::Statistics::KdTree< TSample >::DeleteNode (  KdTreeNodeType *    node )

Deletes the node recursively

template<class TSample> Iterator itk::Statistics::KdTree< TSample >::End (  void    )  [inline]

Definition at line 531 of file itkKdTree.h.

template<class TSample> virtual const char* itk::Statistics::KdTree< TSample >::GetClassName (    )  const [virtual]

Run-time type information (and related methods) Reimplemented from itk::Object.

template<class TSample> DistanceMetricType* itk::Statistics::KdTree< TSample >::GetDistanceMetric (    )  [inline]

Get the pointer to the distance metric. Definition at line 480 of file itkKdTree.h.

template<class TSample> KdTreeNodeType* itk::Statistics::KdTree< TSample >::GetEmptyTerminalNode (    )  [inline]

Returns the pointer to the empty terminal node. A KdTree object has a single empty terminal node in memory. when the split process has to create an empty terminal node, the single instance is reused for this case Definition at line 457 of file itkKdTree.h.

template<class TSample> FrequencyType itk::Statistics::KdTree< TSample >::GetFrequency (  InstanceIdentifier    id )  [inline]

Returns the frequency of the measurement vector identified by the instance identifier Definition at line 476 of file itkKdTree.h.

template<class TSample> MeasurementVectorType itk::Statistics::KdTree< TSample >::GetMeasurementVector (  InstanceIdentifier    id )  [inline]

Returns the measurement vector identified by the instance identifier that is an identifier defiend for the input sample Definition at line 471 of file itkKdTree.h.

template<class TSample> int itk::Statistics::KdTree< TSample >::GetNumberOfVisits (    )  [inline]

Returns the number of measurement vectors that have been visited to find the k-nearest neighbors. Definition at line 495 of file itkKdTree.h.

template<class TSample> KdTreeNodeType* itk::Statistics::KdTree< TSample >::GetRoot (    )  [inline]

Returns the pointer to the root node. Definition at line 466 of file itkKdTree.h.

template<class TSample> TSample* itk::Statistics::KdTree< TSample >::GetSample (    )  [inline]

Returns the pointer to the input sample Definition at line 447 of file itkKdTree.h.

template<class TSample> itk::Statistics::KdTree< TSample >::itkStaticConstMacro (  MeasurementVectorSize   , unsigned    int, TSample::MeasurementVectorSize    )

Length of the measurement vector. k in the k-d tree

template<class TSample> int itk::Statistics::KdTree< TSample >::NearestNeighborSearchLoop (  KdTreeNodeType *    node, MeasurementVectorType &    query, MeasurementVectorType &    lowerBound, MeasurementVectorType &    upperBound )  [protected]

search loop

template<class TSample> Pointer itk::Statistics::KdTree< TSample >::New (    )  [static]

Method for creation through the object factory. Reimplemented from itk::Object.

template<class TSample> void itk::Statistics::KdTree< TSample >::PrintSelf (  std::ostream &    os, Indent    indent )  const [protected, virtual]

Methods invoked by Print() to print information about the object including superclasses. Typically not called by the user (use Print() instead) but used in the hierarchical print process to combine the output of several classes. Reimplemented from itk::Object.

template<class TSample> void itk::Statistics::KdTree< TSample >::PrintTree (  KdTreeNodeType *    node, int    level, unsigned int    activeDimension )

Prints out the tree information

template<class TSample> void itk::Statistics::KdTree< TSample >::Search (  MeasurementVectorType &    query, double    radius, InstanceIdentifierVectorType &    result )

Searches the neighbors fallen into a hypersphere

template<class TSample> void itk::Statistics::KdTree< TSample >::Search (  MeasurementVectorType &    query, unsigned int    k, InstanceIdentifierVectorType &    result )

Searches the k-nearest neighbors

template<class TSample> int itk::Statistics::KdTree< TSample >::SearchLoop (  KdTreeNodeType *    node, MeasurementVectorType &    query, MeasurementVectorType &    lowerBound, MeasurementVectorType &    upperBound )  [protected]

search loop

template<class TSample> void itk::Statistics::KdTree< TSample >::SetBucketSize (  unsigned int    size )

Sets the number of measurement vectors that can be stored in a terminal node

template<class TSample> void itk::Statistics::KdTree< TSample >::SetRoot (  KdTreeNodeType *    root )  [inline]

Sets the root node of the KdTree that is a result of KdTreeGenerator or WeightedCentroidKdTreeGenerator. Definition at line 462 of file itkKdTree.h.

template<class TSample> void itk::Statistics::KdTree< TSample >::SetSample (  TSample *    sample )

Sets the input sample that provides the measurement vectors to the k-d tree

template<class TSample> unsigned long itk::Statistics::KdTree< TSample >::Size (  void    )  [inline]

Definition at line 450 of file itkKdTree.h.

---

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

• itkKdTree.h

---