Module imagingbook.common

Package imagingbook.common.sift

Class SiftDetector

java.lang.Object

imagingbook.common.sift.SiftDetector

public class SiftDetector
extends Object

This class implements the detection of SIFT features from images, as described in [1]. See Ch. 25 of [2] for more details.

Currently only images of type FloatProcessor are supported. The input image is normalized to values in [0,1] before SIFT detection starts. A large set of parameters can be specified (see SiftParameters).

[1] D. G. Lowe. Distinctive image features from scale-invariant keypoints. International Journal of Computer Vision 60, 91–110 (2004).
[2] W. Burger, M.J. Burge, Digital Image Processing – An Algorithmic Introduction, 3rd ed, Springer (2022).

Version:

2022/11/20

See Also:

• SiftParameters

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static enum	SiftDetector.NeighborhoodType3D	Types of 3D neighborhoods used in min/max detection

Constructor Summary

Constructors

Constructor	Description
SiftDetector(FloatProcessor fp)	Constructor using default parameters.
SiftDetector(FloatProcessor fp, SiftParameters params)	Constructor using specific parameters.

Method Summary

Modifier and Type	Method	Description
DogScaleSpace	getDogScaleSpace()	Returns the DogScaleSpace for this SIFT detector instance.
GaussianScaleSpace	getGaussianScaleSpace()	Returns the GaussianScaleSpace for this SIFT detector instance.
List<SiftDescriptor>	getSiftFeatures()	Calculates and returns a list of SIFT descriptors.

Methods inherited from class java.lang.Object

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

Constructor Details

SiftDetector

public SiftDetector(FloatProcessor fp)

Constructor using default parameters.

Parameters:

fp - the input image

See Also:

• SiftDetector(FloatProcessor, SiftParameters)

SiftDetector

public SiftDetector(FloatProcessor fp,
 SiftParameters params)

Constructor using specific parameters. The input image is normalized to values in [0,1], the minimum pixel value being mapped to 0 and the maximum value to 1. An exception is thrown if the supplied image is "flat", i.e., contains only a single pixel value. A large set of parameters can be specified (see SiftParameters). The constructor sets up the complete Gaussian and DoG scale spaces but does not perform feature detection itself, which is done by calling getSiftFeatures().

Parameters:

fp - the input image

params - an instance of SiftParameters

See Also:

• SiftParameters

Method Details

getSiftFeatures

public List<SiftDescriptor> getSiftFeatures()

Calculates and returns a list of SIFT descriptors.

Returns:

a list of extracted SIFT descriptors

getGaussianScaleSpace

public GaussianScaleSpace getGaussianScaleSpace()

Returns the GaussianScaleSpace for this SIFT detector instance. Mainly intended for debugging and testing.

Returns:

the GaussianScaleSpace for this SIFT detector

getDogScaleSpace

public DogScaleSpace getDogScaleSpace()

Returns the DogScaleSpace for this SIFT detector instance. Mainly intended for debugging and testing.

Returns:

the DogScaleSpace for this SIFT detector