boofcv.alg.feature.describe

Class DescribePointSurfMod<II extends boofcv.struct.image.ImageGray>

java.lang.Object

boofcv.alg.feature.describe.DescribePointSurf<II>

boofcv.alg.feature.describe.DescribePointSurfMod<II>

public class DescribePointSurfMod<II extends boofcv.struct.image.ImageGray>
extends DescribePointSurf<II>

Modified SURF descriptor which attempts to smooth out edge conditions. Based upon MU-SURF described in [1] it computes features in over lapping sub-regions and has a separate set of weights for the large grid and for sub-regions. Due to these improvements it will in general produce better results than DescribePointSurf at the cost of additional computations.

[1] M. Agrawal, K. Konolige, and M. Blas, "CenSurE: Center Surround Extremas for Realtime Feature Detection and Matching," Computer Vision – ECCV 2008

Field Summary

Fields inherited from class boofcv.alg.feature.describe.DescribePointSurf

featureDOF, gradient, gradientSafe, ii, kerXX, kerYY, radiusDescriptor, weight, widthLargeGrid, widthSample, widthSubRegion

Constructor Summary

Constructors

Constructor and Description
DescribePointSurfMod(java.lang.Class<II> imageType) Create a SURF-64 descriptor.
DescribePointSurfMod(int widthLargeGrid, int widthSubRegion, int widthSample, int overLap, double sigmaLargeGrid, double sigmaSubRegion, boolean useHaar, java.lang.Class<II> imageType) Creates a SURF descriptor of arbitrary dimension by changing how the local region is sampled.

Method Summary

Modifier and Type	Method and Description
void	features(double c_x, double c_y, double c, double s, double scale, boofcv.struct.sparse.SparseImageGradient gradient, double[] features) An improved SURF descriptor as presented in CenSurE paper.
int	getCanonicalWidth() Width of sampled region when sampling is aligned with image pixels

Methods inherited from class boofcv.alg.feature.describe.DescribePointSurf

computeLaplaceSign, createDescription, describe, describe, getDescriptionLength, setImage

Methods inherited from class java.lang.Object

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

Constructor Detail

DescribePointSurfMod

public DescribePointSurfMod(int widthLargeGrid,
                            int widthSubRegion,
                            int widthSample,
                            int overLap,
                            double sigmaLargeGrid,
                            double sigmaSubRegion,
                            boolean useHaar,
                            java.lang.Class<II> imageType)

Creates a SURF descriptor of arbitrary dimension by changing how the local region is sampled.

Parameters:

widthLargeGrid - Number of sub-regions wide the large grid is. Typically 4.

widthSubRegion - Number of sample points wide a sub-region is. Typically 5.

widthSample - The size of a sample point. Typically 3.

overLap - Number of sample points sub-regions overlap, Typically 2.

sigmaLargeGrid - Sigma used to weight points in the large grid. Typically 2.5

sigmaSubRegion - Sigma used to weight points in the sub-region grid. Typically 2.5

useHaar - If true the Haar wavelet will be used (what was used in [1]), false means an image gradient approximation will be used. True is recommended.

DescribePointSurfMod

public DescribePointSurfMod(java.lang.Class<II> imageType)

Create a SURF-64 descriptor. See [1] for details.

Method Detail

features

public void features(double c_x,
                     double c_y,
                     double c,
                     double s,
                     double scale,
                     boofcv.struct.sparse.SparseImageGradient gradient,
                     double[] features)

An improved SURF descriptor as presented in CenSurE paper. The sub-regions now overlap and more points are sampled in the sub-region to allow overlap.

Overrides:

features in class DescribePointSurf<II extends boofcv.struct.image.ImageGray>

Parameters:

c_x - Center of the feature x-coordinate.

c_y - Center of the feature y-coordinate.

c - cosine of the orientation

s - sine of the orientation

scale - The scale of the wavelets.

features - Where the features are written to. Must be 4*(widthLargeGrid*widthSubRegion)^2 large.

getCanonicalWidth

public int getCanonicalWidth()

Description copied from class: DescribePointSurf

Width of sampled region when sampling is aligned with image pixels

Overrides:

getCanonicalWidth in class DescribePointSurf<II extends boofcv.struct.image.ImageGray>

Returns:

width of descriptor sample