FeatureVector Class Reference

The FeatureVector class. More...

#include <SIFT/FeatureVector.H>

Collaboration diagram for FeatureVector:

[legend]

List of all members.

Public Member Functions
	FeatureVector (int xSize=4, int ySize=4, int zSize=8, bool wrapZ=true)
	Constructor, xSize,ySize,zSize the size of bins, wrap zBin ?
	~FeatureVector ()
	Destructor.
float	getValue (const uint index) const
	get a value from the feature vector
void	addValue (const float x, const float y, const float orientation, const float value)
	add value
void	normalize ()
	normalize Vector
void	threshold (const float limit)
	threshold Vector and renormalize
void	toByteKey (std::vector< byte > &dest, float thresh=0.2, bool norm=true)
	Convert to byte, to get ready for injection into a Keypoint.
int	size ()
int	getXSize ()
	Return the x Bin Size.
int	getYSize ()
	Return the y Bin Size.
int	getZSize ()
	Return the z Bin Size.
Image< PixRGB< byte > >	getFeatureVectorImage (std::vector< byte > &fv)
	Get the feature vector image.
double	getMag ()
	Get the magnitude of the feature vector.
std::vector< float >	getFeatureVector () const
	Get the raw feature vector.

---

Detailed Description

The FeatureVector class.

Stores temporary SIFT Keypoint features. The final feature vector in a SIFT Keypoint object is composed of bytes while here we use floats and are an intermediary onto which we can do normalizations and such before injecting into a Keypoint. A FeatureVector has 128 bins, corresponding to 4 bins for x * 4 bins for y * 4 bins for orientation.

Definition at line 56 of file FeatureVector.H.

---

Constructor & Destructor Documentation

FeatureVector::FeatureVector	(	int	xSize = 4,
		int	ySize = 4,
		int	zSize = 8,
		bool	wrapZ = true
	)

Constructor, xSize,ySize,zSize the size of bins, wrap zBin ?

Definition at line 45 of file FeatureVector.C.

FeatureVector::~FeatureVector

(

)

Destructor.

Definition at line 51 of file FeatureVector.C.

---

Member Function Documentation

void FeatureVector::addValue	(	const float	x,
		const float	y,
		const float	orientation,
		const float	value
	)

add value

Linearly add a value. Indices are as follows: 0.0 <= x <= 4.0 : [0 .. 3] x=2.0 falls equally between bins 1 and 2; 0.0 <= y <= 4.0 : [0 .. 3] y=2.0 falls equally between bins 1 and 2; 0.0 <= o <= 8.0 : [0 .. 7] o=4.0 falls equally between bins 3 and 4

Definition at line 55 of file FeatureVector.C.

Referenced by HistogramOfGradients::addToBin(), and ColorHist::createFeatureVector().

std::vector<float> FeatureVector::getFeatureVector

(

)

const [inline]

Get the raw feature vector.

Definition at line 108 of file FeatureVector.H.

Referenced by JunctionHOG::calculateJunctionHistogram(), ColorHist::createFeatureVector(), and HistogramOfGradients::normalizeFeatureVector().

Image< PixRGB< byte > > FeatureVector::getFeatureVectorImage

(

std::vector< byte > &

fv

)

Get the feature vector image.

Definition at line 214 of file FeatureVector.C.

References Image< T >::clear(), drawDisk(), Point2D< T >::i, and NO_INIT.

double FeatureVector::getMag

(

)

Get the magnitude of the feature vector.

Definition at line 251 of file FeatureVector.C.

References sqrt().

float FeatureVector::getValue

(

const uint

index

)

const

get a value from the feature vector

Definition at line 190 of file FeatureVector.C.

References ASSERT.

int FeatureVector::getXSize

(

)

[inline]

Return the x Bin Size.

Definition at line 93 of file FeatureVector.H.

Referenced by HistogramOfGradients::normalizeFeatureVector().

int FeatureVector::getYSize

(

)

[inline]

Return the y Bin Size.

Definition at line 96 of file FeatureVector.H.

Referenced by HistogramOfGradients::normalizeFeatureVector().

int FeatureVector::getZSize

(

)

[inline]

Return the z Bin Size.

Definition at line 99 of file FeatureVector.H.

Referenced by HistogramOfGradients::addToBin(), and HistogramOfGradients::normalizeFeatureVector().

void FeatureVector::normalize

(

)

normalize Vector

Definition at line 154 of file FeatureVector.C.

Referenced by threshold(), and toByteKey().

void FeatureVector::threshold

(

const float

limit

)

threshold Vector and renormalize

Definition at line 174 of file FeatureVector.C.

References normalize().

Referenced by toByteKey().

void FeatureVector::toByteKey	(	std::vector< byte > &	dest,
		float	thresh = 0.2,
		bool	norm = true
	)

Convert to byte, to get ready for injection into a Keypoint.

CAUTION: this calls normalize(), then threshold(0.2f), then normalize() again (all of which modify the FeatureVector), then converts all values multiplied by 512 to byte with clamping. Call this only when you are ready to discard the FeatureVector and only keep the final SIFT Keypoint byte data.

Definition at line 197 of file FeatureVector.C.

References normalize(), and threshold().

---

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

• src/SIFT/FeatureVector.H
• src/SIFT/FeatureVector.C