net.semanticmetadata.lire.imageanalysis.features.global

类 AutoColorCorrelogram

java.lang.Object

net.semanticmetadata.lire.imageanalysis.features.global.AutoColorCorrelogram

所有已实现的接口:

Extractor, FeatureVector, GlobalFeature, LireFeature

public class AutoColorCorrelogram
extends java.lang.Object
implements GlobalFeature

Feature for the AutoCorrelogram based on color as described in Huang, J.; Kumar, S. R.; Mitra, M.; Zhu, W. & Zabih, R. (2007) "Image Indexing Using Color Correlograms", IEEE Computer Society

see also DOI 10.1109/CVPR.1997.609412

Todo: Change the 2-dim array to a one dim array, as this is much faster in Java.

嵌套类概要

嵌套类

限定符和类型	类和说明
static class	AutoColorCorrelogram.ExtractionMethod Defines which algorithm to use to extract the features vector
static class	AutoColorCorrelogram.Mode Defines the available analysis modes: Superfast uses the approach described in the paper, Quarterneighbourhood investigates the pixels in down and to the right of the respective pixel and FullNeighbourhood investigates the whole lot of pixels within maximumDistance of the respective pixel.

构造器概要

构造器

构造器和说明
AutoColorCorrelogram()
AutoColorCorrelogram(IAutoCorrelogramFeatureExtractor extractionAlgorith) Creates a new AutoCorrelogram with specified algorithm of extraction Uses distance set {1,2,3,4} which is chosen to be compatible with legacy code
AutoColorCorrelogram(int maxDistance) Creates a new AutoColorCorrelogram, where the distance k is limited to a maximum of maxDistance (see publication mentioned above)
AutoColorCorrelogram(int[] distanceSet) Creates a new AutoColorCorrelogram, where the distance k is in distance set
AutoColorCorrelogram(int[] distanceSet, IAutoCorrelogramFeatureExtractor extractionAlgorith) Creates a new AutoCorrelogram with specified algorithm of extraction and distance set
AutoColorCorrelogram(int maxDistance, AutoColorCorrelogram.Mode mode) Creates a new AutoColorCorrelogram using a maximum L_inf pixel distance for analysis and given mode
AutoColorCorrelogram(int numBins, int[] distanceSet, IAutoCorrelogramFeatureExtractor extractionAlgorith) Creates a new AutoColorCorrelogram using a maximum L_inf pixel distance for analysis and given mode

方法概要

限定符和类型	方法和说明
void	extract(java.awt.image.BufferedImage bi) Extracts the feature vector from a BufferedImage.
void	extract(int[][][] img)
byte[]	getByteArrayRepresentation() Returns a compact byte[] based representation of the feature vector.
double	getDistance(LireFeature vd) The distance function for this type of feature
java.lang.String	getFeatureName() Gives a descriptive name of the feature, i.e. a name to show up in benchmarks, menus, UIs, etc.
double[]	getFeatureVector() Convenience method to get the feature vector as double[] array.
java.lang.String	getFieldName() Returns the preferred field name for indexing.
void	setByteArrayRepresentation(byte[] in) Sets the feature vector values based on the byte[] data.
void	setByteArrayRepresentation(byte[] in, int offset, int length) Sets the feature vector values based on the byte[] data.

从类继承的方法 java.lang.Object

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

构造器详细资料

AutoColorCorrelogram

public AutoColorCorrelogram()

AutoColorCorrelogram

public AutoColorCorrelogram(int maxDistance)

Creates a new AutoColorCorrelogram, where the distance k is limited to a maximum of maxDistance (see publication mentioned above)

参数:

maxDistance - upper limit of k

AutoColorCorrelogram

public AutoColorCorrelogram(int maxDistance,
                            AutoColorCorrelogram.Mode mode)

Creates a new AutoColorCorrelogram using a maximum L_inf pixel distance for analysis and given mode

参数:

maxDistance - maximum L_inf pixel distance for analysis

mode - the mode of calculation (determines the speed of extraction)

AutoColorCorrelogram

public AutoColorCorrelogram(int[] distanceSet)

Creates a new AutoColorCorrelogram, where the distance k is in distance set

参数:

distanceSet - distance set

AutoColorCorrelogram

public AutoColorCorrelogram(int[] distanceSet,
                            IAutoCorrelogramFeatureExtractor extractionAlgorith)

Creates a new AutoCorrelogram with specified algorithm of extraction and distance set

参数:

distanceSet - distance set

extractionAlgorith - the algorithm to extract

AutoColorCorrelogram

public AutoColorCorrelogram(IAutoCorrelogramFeatureExtractor extractionAlgorith)

Creates a new AutoCorrelogram with specified algorithm of extraction Uses distance set {1,2,3,4} which is chosen to be compatible with legacy code

参数:

extractionAlgorith - the algorithm to extract

AutoColorCorrelogram

public AutoColorCorrelogram(int numBins,
                            int[] distanceSet,
                            IAutoCorrelogramFeatureExtractor extractionAlgorith)

Creates a new AutoColorCorrelogram using a maximum L_inf pixel distance for analysis and given mode

方法详细资料

extract

public void extract(java.awt.image.BufferedImage bi)

从接口复制的说明: Extractor

Extracts the feature vector from a BufferedImage.

指定者:

extract 在接口中 Extractor

参数:

bi - the source image

getByteArrayRepresentation

public byte[] getByteArrayRepresentation()

从接口复制的说明: LireFeature

Returns a compact byte[] based representation of the feature vector.

指定者:

getByteArrayRepresentation 在接口中 LireFeature

返回:

a compact byte[] array containing the feature vector.

另请参阅:

LireFeature.setByteArrayRepresentation(byte[])

setByteArrayRepresentation

public void setByteArrayRepresentation(byte[] in)

从接口复制的说明: LireFeature

Sets the feature vector values based on the byte[] data. Use LireFeature.getByteArrayRepresentation() to generate a compatible byte[] array.

指定者:

setByteArrayRepresentation 在接口中 LireFeature

参数:

in - the byte[] data.

另请参阅:

LireFeature.getByteArrayRepresentation()

setByteArrayRepresentation

public void setByteArrayRepresentation(byte[] in,
                                       int offset,
                                       int length)

从接口复制的说明: LireFeature

Sets the feature vector values based on the byte[] data. Use LireFeature.getByteArrayRepresentation() to generate a compatible byte[] array.

指定者:

setByteArrayRepresentation 在接口中 LireFeature

参数:

in - the byte[] array containing the data.

offset - the offset, i.e. where the feature vector starts.

length - the length of the data representing the feature vector.

另请参阅:

LireFeature.getByteArrayRepresentation()

getFeatureVector

public double[] getFeatureVector()

从接口复制的说明: FeatureVector

Convenience method to get the feature vector as double[] array. This method is used by classifiers, etc.

指定者:

getFeatureVector 在接口中 FeatureVector

返回:

the feature vector as a double[] array.

extract

public void extract(int[][][] img)

getDistance

public double getDistance(LireFeature vd)

从接口复制的说明: LireFeature

The distance function for this type of feature

指定者:

getDistance 在接口中 LireFeature

参数:

vd - the feature vector to compare the current instance to.

返回:

the distance (or dissimilarity) between the instance and the parameter.

getFeatureName

public java.lang.String getFeatureName()

从接口复制的说明: LireFeature

Gives a descriptive name of the feature, i.e. a name to show up in benchmarks, menus, UIs, etc.

指定者:

getFeatureName 在接口中 LireFeature

返回:

the name of the feature.

getFieldName

public java.lang.String getFieldName()

从接口复制的说明: LireFeature

Returns the preferred field name for indexing.

指定者:

getFieldName 在接口中 LireFeature

返回:

the field name preferred for indexing in a Lucene index.