itk::Statistics::Algorithm Namespace Reference

Functions
template<typename TSubsample >
void	DownHeap (TSubsample *sample, unsigned int activeDimension, int beginIndex, int endIndex, int node)
template<typename TSample >
void	FindSampleBound (const TSample *sample, const typename TSample::ConstIterator &begin, const typename TSample::ConstIterator &end, typename TSample::MeasurementVectorType &min, typename TSample::MeasurementVectorType &max)
template<typename TSubsample >
void	FindSampleBoundAndMean (const TSubsample *sample, int beginIndex, int endIndex, typename TSubsample::MeasurementVectorType &min, typename TSubsample::MeasurementVectorType &max, typename TSubsample::MeasurementVectorType &mean)
template<typename TSize >
TSize	FloorLog (TSize size)
template<typename TSubsample >
void	HeapSort (TSubsample *sample, unsigned int activeDimension, int beginIndex, int endIndex)
template<typename TSubsample >
void	InsertSort (TSubsample *sample, unsigned int activeDimension, int beginIndex, int endIndex)
template<typename TSubsample >
void	IntrospectiveSort (TSubsample *sample, unsigned int activeDimension, int beginIndex, int endIndex, int sizeThreshold)
template<typename TSubsample >
void	IntrospectiveSortLoop (TSubsample *sample, unsigned int activeDimension, int beginIndex, int endIndex, int depthLimit, int sizeThreshold)
template<typename TValue >
TValue	MedianOfThree (const TValue a, const TValue b, const TValue c)
template<typename TSubsample >
TSubsample::MeasurementType	NthElement (TSubsample *sample, unsigned int activeDimension, int beginIndex, int endIndex, int nth)
template<typename TSubsample >
int	Partition (TSubsample *sample, unsigned int activeDimension, int beginIndex, int endIndex, const typename TSubsample::MeasurementType partitionValue)
template<typename TSubsample >
TSubsample::MeasurementType	QuickSelect (TSubsample *sample, unsigned int activeDimension, int beginIndex, int endIndex, int kth)
template<typename TSubsample >
TSubsample::MeasurementType	QuickSelect (TSubsample *sample, unsigned int activeDimension, int beginIndex, int endIndex, int kth, typename TSubsample::MeasurementType medianGuess)

Function Documentation

DownHeap()

template<typename TSubsample >

void itk::Statistics::Algorithm::DownHeap	(	TSubsample *	sample,
		unsigned int	activeDimension,
		int	beginIndex,
		int	endIndex,
		int	node
	)

FindSampleBound()

template<typename TSample >

void itk::Statistics::Algorithm::FindSampleBound	(	const TSample *	sample,
		const typename TSample::ConstIterator &	begin,
		const typename TSample::ConstIterator &	end,
		typename TSample::MeasurementVectorType &	min,
		typename TSample::MeasurementVectorType &	max
	)

Finds the minimum and maximum of each measurement value, over an interval within a sample.

Samples are considered independently of each other.

Parameters

[in]	sample	An instance of TSample.
[in]	begin	An iterator to the first measurement vector in sample to be considered.
[in]	end	An iterator to the last measurement vector in sample to be considered.
[out]	min	A new measurement vector, with the minimum values.
[out]	max	A new measurement vector, with the maximum values.

Template Parameters

TSample

A subtype of Statistics::Sample.

FindSampleBoundAndMean()

template<typename TSubsample >

void itk::Statistics::Algorithm::FindSampleBoundAndMean	(	const TSubsample *	sample,
		int	beginIndex,
		int	endIndex,
		typename TSubsample::MeasurementVectorType &	min,
		typename TSubsample::MeasurementVectorType &	max,
		typename TSubsample::MeasurementVectorType &	mean
	)

The endIndex should points one point after the last elements.

FloorLog()

template<typename TSize >

TSize itk::Statistics::Algorithm::FloorLog

(

TSize

size

)

HeapSort()

template<typename TSubsample >

void itk::Statistics::Algorithm::HeapSort	(	TSubsample *	sample,
		unsigned int	activeDimension,
		int	beginIndex,
		int	endIndex
	)

InsertSort()

template<typename TSubsample >

void itk::Statistics::Algorithm::InsertSort	(	TSubsample *	sample,
		unsigned int	activeDimension,
		int	beginIndex,
		int	endIndex
	)

IntrospectiveSort()

template<typename TSubsample >

void itk::Statistics::Algorithm::IntrospectiveSort	(	TSubsample *	sample,
		unsigned int	activeDimension,
		int	beginIndex,
		int	endIndex,
		int	sizeThreshold
	)

IntrospectiveSortLoop()

template<typename TSubsample >

void itk::Statistics::Algorithm::IntrospectiveSortLoop	(	TSubsample *	sample,
		unsigned int	activeDimension,
		int	beginIndex,
		int	endIndex,
		int	depthLimit,
		int	sizeThreshold
	)

MedianOfThree()

template<typename TValue >

TValue itk::Statistics::Algorithm::MedianOfThree	(	const TValue	a,
		const TValue	b,
		const TValue	c
	)

NthElement()

template<typename TSubsample >

TSubsample::MeasurementType itk::Statistics::Algorithm::NthElement	(	TSubsample *	sample,
		unsigned int	activeDimension,
		int	beginIndex,
		int	endIndex,
		int	nth
	)

NthElement is an algorithm for finding the n-th largest element of a list.

In this case, only one of the components of the measurement vectors is considered. This component is defined by the argument activeDimension. The search is restricted to the range between the index begin and end, also passed as arguments. This algorithm was based on the procedure used in the STL nth_element method.

Partition()

template<typename TSubsample >

int itk::Statistics::Algorithm::Partition	(	TSubsample *	sample,
		unsigned int	activeDimension,
		int	beginIndex,
		int	endIndex,
		const typename TSubsample::MeasurementType	partitionValue
	)

The Partition algorithm performs partial sorting in a sample.

Given a partitionValue, the algorithm moves to the beginning of the sample all MeasurementVectors whose component activeDimension is smaller than the partitionValue. In this way, the sample is partially sorted in two groups. First the group with activeDimension component smaller than the partitionValue, then the group of MeasurementVectors with activeDimension component larger than the partitionValue. The Partition algorithm takes as input a sample, and a range in that sample defined by [beginIndex,endIndex]. Only the activeDimension components of the MeasurementVectors in the sample will be considered by the algorithm. The Algorithm return an index in the range of [beginIndex,endIndex] pointing to the element with activeDimension component closest to the partitionValue.

The endIndex should points one point after the last elements if multiple partitionValue exist in the sample the return index will points the middle of such values. Implemented following the description of the partition algorithm in the QuickSelect entry of the Wikipedia.

See also

https://en.wikipedia.org/wiki/Selection_algorithm.

QuickSelect() [1/2]

template<typename TSubsample >

TSubsample::MeasurementType itk::Statistics::Algorithm::QuickSelect	(	TSubsample *	sample,
		unsigned int	activeDimension,
		int	beginIndex,
		int	endIndex,
		int	kth
	)

QuickSelect is an algorithm for finding the k-th largest element of a list.

In this case, only one of the components of the measurement vectors is considered. This component is defined by the argument activeDimension. The search is restricted to the range between the index begin and end, also passed as arguments.

The endIndex should point one point after the last elements. Note that kth is an index in a different scale than [beginIndex,endIndex]. For example, it is possible to feed this function with beginIndex=15, endIndex=23, and kth=3, since we can ask for the element 3rd in the range [15,23].

See also

https://en.wikipedia.org/wiki/Selection_algorithm.

QuickSelect() [2/2]

template<typename TSubsample >

TSubsample::MeasurementType itk::Statistics::Algorithm::QuickSelect	(	TSubsample *	sample,
		unsigned int	activeDimension,
		int	beginIndex,
		int	endIndex,
		int	kth,
		typename TSubsample::MeasurementType	medianGuess
	)

QuickSelect is an algorithm for finding the k-th largest element of a list.

In this case, only one of the components of the measurement vectors is considered. This component is defined by the argument activeDimension. The search is restricted to the range between the index begin and end, also passed as arguments.

The endIndex should point one point after the last elements. Note that kth is an index in a different scale than [beginIndex,endIndex]. For example, it is possible to feed this function with beginIndex=15, endIndex=23, and kth=3, since we can ask for the element 3rd in the range [15,23]. In this version, a guess value for the median index is provided in the argument medianGuess. The algorithm returns the value of the activeDimension component in the MeasurementVector located in the kth position.

See also

https://en.wikipedia.org/wiki/Selection_algorithm