boofcv.abst.feature.describe

Class DescribeRegionPoint_SIFT<T extends boofcv.struct.image.ImageGray>

java.lang.Object

boofcv.abst.feature.describe.DescribeRegionPoint_SIFT<T>

All Implemented Interfaces:

DescribeRegionPoint<T,TupleDesc_F64>, DescriptorInfo<TupleDesc_F64>

public class DescribeRegionPoint_SIFT<T extends boofcv.struct.image.ImageGray>
extends java.lang.Object
implements DescribeRegionPoint<T,TupleDesc_F64>

Allows you to use SIFT features independent of the SIFT detector. A SIFT scale-space is computed with all octaves and most of the scales saved. When a few feature is requested it looks up the closest scale image and uses that as the input image.

Constructor Summary

Constructors

Constructor and Description
DescribeRegionPoint_SIFT(SiftScaleSpace scaleSpace, DescribePointSift<boofcv.struct.image.GrayF32> describe, java.lang.Class<T> imageType)

Method Summary

Modifier and Type	Method and Description
TupleDesc_F64	createDescription() Creates new description instance which can be processed by this class
double	getCanonicalWidth() Returns the width of the square (or approximation of) sample region at a scale of one.
java.lang.Class<TupleDesc_F64>	getDescriptionType() The type of region descriptor generated
boofcv.struct.image.ImageType<T>	getImageType() Description of the type of image it can process
boolean	process(double x, double y, double orientation, double radius, TupleDesc_F64 description) Extract a description of the local image at the given point, scale, and orientation.
boolean	requiresOrientation() True if the descriptor uses orientation information.
boolean	requiresRadius() If size information is used when computing the descriptor.
void	setImage(T image) Specified the image which is to be processed.

Methods inherited from class java.lang.Object

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

Constructor Detail

DescribeRegionPoint_SIFT

public DescribeRegionPoint_SIFT(SiftScaleSpace scaleSpace,
                                DescribePointSift<boofcv.struct.image.GrayF32> describe,
                                java.lang.Class<T> imageType)

Method Detail

setImage

public void setImage(T image)

Description copied from interface: DescribeRegionPoint

Specified the image which is to be processed.

Specified by:

setImage in interface DescribeRegionPoint<T extends boofcv.struct.image.ImageGray,TupleDesc_F64>

Parameters:

image - The image which contains the features.

process

public boolean process(double x,
                       double y,
                       double orientation,
                       double radius,
                       TupleDesc_F64 description)

Description copied from interface: DescribeRegionPoint

Extract a description of the local image at the given point, scale, and orientation. WARNING: Check the returned value to make sure a description was actually computed. Some implementations might now allow features to extend outside the image border and will return false.

Specified by:

process in interface DescribeRegionPoint<T extends boofcv.struct.image.ImageGray,TupleDesc_F64>

Parameters:

x - Coordinate of the point.

y - Coordinate of the point.

orientation - Direction the feature is pointing at in radians. 0 = x-axis PI/2 = y-axis

radius - Radius of the detected object in pixels.

description - (output) Storage for extracted feature. Use DescriptorInfo.createDescription() to create descriptor.

Returns:

true if a descriptor can computed or false if not.

requiresRadius

public boolean requiresRadius()

Description copied from interface: DescribeRegionPoint

If size information is used when computing the descriptor.

Specified by:

requiresRadius in interface DescribeRegionPoint<T extends boofcv.struct.image.ImageGray,TupleDesc_F64>

Returns:

true is the radius is used when computing the descriptor or false if not

requiresOrientation

public boolean requiresOrientation()

Description copied from interface: DescribeRegionPoint

True if the descriptor uses orientation information.

Specified by:

requiresOrientation in interface DescribeRegionPoint<T extends boofcv.struct.image.ImageGray,TupleDesc_F64>

Returns:

if orientation needs to be provided or not

getImageType

public boofcv.struct.image.ImageType<T> getImageType()

Description copied from interface: DescribeRegionPoint

Description of the type of image it can process

Specified by:

getImageType in interface DescribeRegionPoint<T extends boofcv.struct.image.ImageGray,TupleDesc_F64>

Returns:

ImageDataType

getCanonicalWidth

public double getCanonicalWidth()

Description copied from interface: DescribeRegionPoint

Returns the width of the square (or approximation of) sample region at a scale of one. When multiplied by the scale, pixels outside of the square region should not influence the descriptor's value.

Specified by:

getCanonicalWidth in interface DescribeRegionPoint<T extends boofcv.struct.image.ImageGray,TupleDesc_F64>

Returns:

width of descriptor at a scale of one

createDescription

public TupleDesc_F64 createDescription()

Description copied from interface: DescriptorInfo

Creates new description instance which can be processed by this class

Specified by:

createDescription in interface DescriptorInfo<TupleDesc_F64>

Returns:

New descriptor

getDescriptionType

public java.lang.Class<TupleDesc_F64> getDescriptionType()

Description copied from interface: DescriptorInfo

The type of region descriptor generated

Specified by:

getDescriptionType in interface DescriptorInfo<TupleDesc_F64>

Returns:

Returns the descriptor type.