net.semanticmetadata.lire.searchers.custom

类 SingleNddCeddImageSearcher

java.lang.Object

net.semanticmetadata.lire.searchers.AbstractImageSearcher

net.semanticmetadata.lire.searchers.custom.SingleNddCeddImageSearcher

所有已实现的接口:

ImageSearcher

public class SingleNddCeddImageSearcher
extends AbstractImageSearcher

A ImageSearcher that retrieves just the first result, caches the whole index and optimizes search time by bundling searches. Please note that as soon as the instance is created, changes in the index are not reflected.

作者:

Mathias Lux, mathias@juggle.at

字段概要

字段

限定符和类型	字段和说明
protected GlobalFeature	cachedInstance
protected java.util.TreeSet<SimpleResult>	docs
protected java.util.ArrayList<double[]>	featureCache
protected boolean	isCaching
protected java.util.logging.Logger	logger
protected double	maxDistance
protected org.apache.lucene.index.IndexReader	reader
protected boolean	useSimilarityScore

构造器概要

构造器

构造器和说明
SingleNddCeddImageSearcher(org.apache.lucene.index.IndexReader reader) Creates a new ImageSearcher for searching just one single image based on CEDD from a RAM cached data set.
SingleNddCeddImageSearcher(org.apache.lucene.index.IndexReader reader, boolean approximate) Creates a new ImageSearcher for searching just one single image based on CEDD from a RAM cached data set.
SingleNddCeddImageSearcher(org.apache.lucene.index.IndexReader reader, boolean approximate, java.lang.Class descriptorClass, java.lang.String fieldName) Eventually to be used with other LireFeature classes.

方法概要

限定符和类型	方法和说明
ImageDuplicates	findDuplicates(org.apache.lucene.index.IndexReader reader) Identifies duplicates in the database.
SimpleResult	findMostSimilar(GlobalFeature globalFeature)
SimpleResult[]	findMostSimilar(GlobalFeature[] globalFeatures)
SimpleResult[]	findMostSimilar(GlobalFeature[] globalFeatures, int offset, int length)
protected double	findSimilar(org.apache.lucene.index.IndexReader reader, GlobalFeature globalFeature)
protected double	getDistance(org.apache.lucene.document.Document document, GlobalFeature globalFeature) Main similarity method called for each and every document in the index.
protected void	init(org.apache.lucene.index.IndexReader reader)
ImageSearchHits	search(java.awt.image.BufferedImage image, org.apache.lucene.index.IndexReader reader) Searches for images similar to the given image.
ImageSearchHits	search(org.apache.lucene.document.Document doc, org.apache.lucene.index.IndexReader reader) Searches for images similar to the given image, defined by the Document from the index.
java.lang.String	toString()

从类继承的方法 net.semanticmetadata.lire.searchers.AbstractImageSearcher

relevanceFeedback, search

从类继承的方法 java.lang.Object

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

字段详细资料

logger

protected java.util.logging.Logger logger

cachedInstance

protected GlobalFeature cachedInstance

isCaching

protected boolean isCaching

featureCache

protected java.util.ArrayList<double[]> featureCache

reader

protected org.apache.lucene.index.IndexReader reader

docs

protected java.util.TreeSet<SimpleResult> docs

maxDistance

protected double maxDistance

useSimilarityScore

protected boolean useSimilarityScore

构造器详细资料

SingleNddCeddImageSearcher

public SingleNddCeddImageSearcher(org.apache.lucene.index.IndexReader reader)

Creates a new ImageSearcher for searching just one single image based on CEDD from a RAM cached data set.

参数:

reader - the index reader pointing to the index. It will be cached first, so changes will not be reflected in this instance.

SingleNddCeddImageSearcher

public SingleNddCeddImageSearcher(org.apache.lucene.index.IndexReader reader,
                                  boolean approximate)

Creates a new ImageSearcher for searching just one single image based on CEDD from a RAM cached data set. Set approximate to true if you want to speed up search and loose accuracy.

参数:

reader - the index reader pointing to the index. It will be cached first, so changes will not be reflected in this instance.

approximate - set to true if you want to trade accuracy to speed, setting to true is faster (~ double speed), but less accurate

SingleNddCeddImageSearcher

public SingleNddCeddImageSearcher(org.apache.lucene.index.IndexReader reader,
                                  boolean approximate,
                                  java.lang.Class descriptorClass,
                                  java.lang.String fieldName)

Eventually to be used with other LireFeature classes.

参数:

reader -

approximate -

descriptorClass -

方法详细资料

init

protected void init(org.apache.lucene.index.IndexReader reader)

search

public ImageSearchHits search(java.awt.image.BufferedImage image,
                              org.apache.lucene.index.IndexReader reader)
                       throws java.io.IOException

从接口复制的说明: ImageSearcher

Searches for images similar to the given image.

参数:

image - the example image to search for.

reader - the IndexReader which is used to search through the images.

返回:

a sorted list of hits.

抛出:

java.io.IOException - in case exceptions in the reader occurs

findSimilar

protected double findSimilar(org.apache.lucene.index.IndexReader reader,
                             GlobalFeature globalFeature)
                      throws java.io.IOException

参数:

reader -

globalFeature -

返回:

the maximum distance found for normalizing.

抛出:

java.io.IOException

findMostSimilar

public SimpleResult findMostSimilar(GlobalFeature globalFeature)
                             throws java.io.IOException

抛出:

java.io.IOException

findMostSimilar

public SimpleResult[] findMostSimilar(GlobalFeature[] globalFeatures)
                               throws java.io.IOException

抛出:

java.io.IOException

findMostSimilar

public SimpleResult[] findMostSimilar(GlobalFeature[] globalFeatures,
                                      int offset,
                                      int length)
                               throws java.io.IOException

抛出:

java.io.IOException

getDistance

protected double getDistance(org.apache.lucene.document.Document document,
                             GlobalFeature globalFeature)

Main similarity method called for each and every document in the index.

参数:

document -

globalFeature -

返回:

the distance between the given feature and the feature stored in the document.

search

public ImageSearchHits search(org.apache.lucene.document.Document doc,
                              org.apache.lucene.index.IndexReader reader)
                       throws java.io.IOException

从接口复制的说明: ImageSearcher

Searches for images similar to the given image, defined by the Document from the index.

参数:

doc - the example image to search for.

reader - the IndexReader which is used to dsearch through the images.

返回:

a sorted list of hits.

抛出:

java.io.IOException - in case exceptions in the reader occurs

findDuplicates

public ImageDuplicates findDuplicates(org.apache.lucene.index.IndexReader reader)
                               throws java.io.IOException

从接口复制的说明: ImageSearcher

Identifies duplicates in the database.

参数:

reader - the IndexReader which is used to dsearch through the images.

返回:

a sorted list of hits.

抛出:

java.io.IOException - in case the image could not be read from stream.

toString

public java.lang.String toString()

覆盖:

toString 在类中 java.lang.Object