boofcv.alg.feature.orientation

Class OrientationHistogramSift

java.lang.Object

boofcv.alg.feature.orientation.OrientationHistogramSift

public class OrientationHistogramSift
extends Object

Computes the orientation of a region around a point in scale-space as specified in the SIFT [1] paper. A histogram of gradients is computed and the largest peaks are returned as the region's direction. The gradient is weighted using a Gaussian distribution.

INTERPOLATION: Instead of fitting a curve to adjacent bins, the solution is computed by summing up dx and dy independently in each bin. Then when a bin is selected the angle is set to the atan() of dx,dy sums.

[1] Lowe, D. "Distinctive image features from scale-invariant keypoints". International Journal of Computer Vision, 60, 2 (2004), pp.91--110.

Author:

Peter Abeles

Constructor Summary

Constructors

Constructor and Description
OrientationHistogramSift(int histogramSize, double sigmaToRadius, double sigmaEnlarge) Configures orientation estimation

Method Summary

Methods

Modifier and Type	Method and Description
int	getImageIndex() Which image in the scale space is being used.
org.ddogleg.struct.GrowQueue_F64	getOrientations() A list of found orientations
double	getPeakOrientation() Orientation of the largest peak
double	getPixelScale() The ratio of pixels in the octave to original image
void	process(double c_x, double c_y, double scale) Estimates the orientation(s) of a region at the specified location and scale
void	setScaleSpace(SiftImageScaleSpace ss) Specify the input

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

OrientationHistogramSift

public OrientationHistogramSift(int histogramSize,
                        double sigmaToRadius,
                        double sigmaEnlarge)

Configures orientation estimation

Parameters:

histogramSize - Number of elements in the histogram. Standard is 36

sigmaToRadius - Convert a sigma to region radius. Try 2.5

sigmaEnlarge - How much the scale is enlarged by. Standard is 1.5

Method Detail

setScaleSpace

public void setScaleSpace(SiftImageScaleSpace ss)

Specify the input

Parameters:

ss - Scale space representation of input image

process

public void process(double c_x,
           double c_y,
           double scale)

Estimates the orientation(s) of a region at the specified location and scale

Parameters:

c_x - Location x-axis

c_y - Location y-axis

scale - Scale (blur standard deviations)

getImageIndex

public int getImageIndex()

Which image in the scale space is being used.

getPixelScale

public double getPixelScale()

The ratio of pixels in the octave to original image

getOrientations

public org.ddogleg.struct.GrowQueue_F64 getOrientations()

A list of found orientations

Returns:

orientations

getPeakOrientation

public double getPeakOrientation()

Orientation of the largest peak