net.semanticmetadata.lire.indexers.parallel

类 ParallelIndexer

java.lang.Object

net.semanticmetadata.lire.indexers.parallel.ParallelIndexer

所有已实现的接口:

java.lang.Runnable

public class ParallelIndexer
extends java.lang.Object
implements java.lang.Runnable

This class allows for creating indexes in a parallel manner. The class at hand reads files from the disk and acts as producer, while several consumer threads extract the features from the given files.

Use the methods addExtractor(java.lang.Class<? extends net.semanticmetadata.lire.imageanalysis.features.Extractor>) to add your own features. Check the source of this class -- the main method -- to get an idea.

Created by mlux on 15/04/2013.

作者:

Mathias Lux, mathias@juggle.at, Nektarios Anagnostopoulos, nek.anag@gmail.com

构造器概要

构造器

构造器和说明
ParallelIndexer(int numOfThreads, java.lang.String indexPath, java.io.File imageList)
ParallelIndexer(int numOfThreads, java.lang.String indexPath, java.io.File imageList, boolean overWrite)
ParallelIndexer(int numOfThreads, java.lang.String indexPath, java.io.File imageList, GlobalDocumentBuilder.HashingMode hashingMode) Constructor for use with hashing.
ParallelIndexer(int numOfThreads, java.lang.String indexPath, java.io.File imageList, GlobalDocumentBuilder.HashingMode hashingMode, boolean useDocValues) Constructor for use with hashing and optional storage in DocValues instead of Lucene fields.
ParallelIndexer(int numOfThreads, java.lang.String indexPath, java.io.File imageList, GlobalDocumentBuilder.HashingMode hashingMode, boolean useDocValues, int queueSize) Constructor for use with hashing and optional storage in DocValues instead of Lucene fields.
ParallelIndexer(int numOfThreads, java.lang.String indexPath, java.io.File imageList, int[] numOfClusters, int numOfDocsForCodebooks)
ParallelIndexer(int numOfThreads, java.lang.String indexPath, java.io.File imageList, int[] numOfClusters, int numOfDocsForCodebooks, java.lang.Class<? extends AbstractAggregator> aggregator)
ParallelIndexer(int numOfThreads, java.lang.String indexPath, java.io.File imageList, int numOfClusters, int numOfDocsForCodebooks)
ParallelIndexer(int numOfThreads, java.lang.String indexPath, java.io.File imageList, int numOfClusters, int numOfDocsForCodebooks, java.lang.Class<? extends AbstractAggregator> aggregator)
ParallelIndexer(int numOfThreads, java.lang.String indexPath, java.io.File imageList, java.lang.String fromIndexPath)
ParallelIndexer(int numOfThreads, java.lang.String indexPath, java.lang.String imageDirectory)
ParallelIndexer(int numOfThreads, java.lang.String indexPath, java.lang.String imageDirectory, boolean overWrite)
ParallelIndexer(int numOfThreads, java.lang.String indexPath, java.lang.String imageDirectory, int[] numOfClusters, int numOfDocsForCodebooks)
ParallelIndexer(int numOfThreads, java.lang.String indexPath, java.lang.String imageDirectory, int[] numOfClusters, int numOfDocsForCodebooks, java.lang.Class<? extends AbstractAggregator> aggregator)
ParallelIndexer(int numOfThreads, java.lang.String indexPath, java.lang.String imageDirectory, int numOfClusters, int numOfDocsForCodebooks)
ParallelIndexer(int numOfThreads, java.lang.String indexPath, java.lang.String imageDirectory, int numOfClusters, int numOfDocsForCodebooks, java.lang.Class<? extends AbstractAggregator> aggregator)
ParallelIndexer(int numOfThreads, java.lang.String indexPath, java.lang.String imageDirectory, java.lang.String fromIndexPath)

方法概要

限定符和类型	方法和说明
void	addExtractor(java.lang.Class<? extends Extractor> extractorClass)
void	addExtractor(java.lang.Class<? extends GlobalFeature> globalFeatureClass, SimpleExtractor.KeypointDetector detector)
void	addExtractor(java.lang.Class<? extends GlobalFeature> globalFeatureClass, SimpleExtractor.KeypointDetector detector, Cluster[] codebook)
void	addExtractor(java.lang.Class<? extends GlobalFeature> globalFeatureClass, SimpleExtractor.KeypointDetector detector, int numKeyPoints)
void	addExtractor(java.lang.Class<? extends GlobalFeature> globalFeatureClass, SimpleExtractor.KeypointDetector detector, java.util.LinkedList<Cluster[]> codebooks)
void	addExtractor(java.lang.Class<? extends LocalFeatureExtractor> localFeatureExtractorClass, Cluster[] codebook)
void	addExtractor(java.lang.Class<? extends LocalFeatureExtractor> localFeatureExtractorClass, java.util.LinkedList<Cluster[]> codebooks)
ImagePreprocessor	getImagePreprocessor()
double	getPercentageDone()
boolean	hasEnded()
static void	main(java.lang.String[] args)
void	run()
void	setCustomDocumentBuilder(java.lang.Class<? extends DocumentBuilder> customDocumentBuilder) WARNING!!
void	setImagePreprocessor(ImagePreprocessor imagePreprocessor)

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

ParallelIndexer

public ParallelIndexer(int numOfThreads,
                       java.lang.String indexPath,
                       java.lang.String imageDirectory)

ParallelIndexer

public ParallelIndexer(int numOfThreads,
                       java.lang.String indexPath,
                       java.lang.String imageDirectory,
                       int numOfClusters,
                       int numOfDocsForCodebooks)

ParallelIndexer

public ParallelIndexer(int numOfThreads,
                       java.lang.String indexPath,
                       java.lang.String imageDirectory,
                       int numOfClusters,
                       int numOfDocsForCodebooks,
                       java.lang.Class<? extends AbstractAggregator> aggregator)

ParallelIndexer

public ParallelIndexer(int numOfThreads,
                       java.lang.String indexPath,
                       java.lang.String imageDirectory,
                       int[] numOfClusters,
                       int numOfDocsForCodebooks)

ParallelIndexer

public ParallelIndexer(int numOfThreads,
                       java.lang.String indexPath,
                       java.lang.String imageDirectory,
                       int[] numOfClusters,
                       int numOfDocsForCodebooks,
                       java.lang.Class<? extends AbstractAggregator> aggregator)

ParallelIndexer

public ParallelIndexer(int numOfThreads,
                       java.lang.String indexPath,
                       java.lang.String imageDirectory,
                       boolean overWrite)

ParallelIndexer

public ParallelIndexer(int numOfThreads,
                       java.lang.String indexPath,
                       java.lang.String imageDirectory,
                       java.lang.String fromIndexPath)

ParallelIndexer

public ParallelIndexer(int numOfThreads,
                       java.lang.String indexPath,
                       java.io.File imageList)

ParallelIndexer

public ParallelIndexer(int numOfThreads,
                       java.lang.String indexPath,
                       java.io.File imageList,
                       GlobalDocumentBuilder.HashingMode hashingMode)

Constructor for use with hashing.

参数:

numOfThreads - number of threads used for processing.

indexPath - the directory the index witll be written to.

imageList - the list of images, one path per line.

hashingMode - the mode used for Hashing, use HashingMode.None if you don't want hashing.

ParallelIndexer

public ParallelIndexer(int numOfThreads,
                       java.lang.String indexPath,
                       java.io.File imageList,
                       GlobalDocumentBuilder.HashingMode hashingMode,
                       boolean useDocValues)

Constructor for use with hashing and optional storage in DocValues instead of Lucene fields.

参数:

numOfThreads - number of threads used for processing.

indexPath - the directory the index witll be written to.

imageList - the list of images, one path per line.

hashingMode - the mode used for Hashing, use HashingMode.None if you don't want hashing.

useDocValues - set to true if you want to use DocValues instead of Fields.

ParallelIndexer

public ParallelIndexer(int numOfThreads,
                       java.lang.String indexPath,
                       java.io.File imageList,
                       GlobalDocumentBuilder.HashingMode hashingMode,
                       boolean useDocValues,
                       int queueSize)

Constructor for use with hashing and optional storage in DocValues instead of Lucene fields.

参数:

numOfThreads - number of threads used for processing.

indexPath - the directory the index witll be written to.

imageList - the list of images, one path per line.

hashingMode - the mode used for Hashing, use HashingMode.None if you don't want hashing.

useDocValues - set to true if you want to use DocValues instead of Fields.

queueSize - the size of the reading queue to minimize disk usage.

ParallelIndexer

public ParallelIndexer(int numOfThreads,
                       java.lang.String indexPath,
                       java.io.File imageList,
                       int numOfClusters,
                       int numOfDocsForCodebooks)

ParallelIndexer

public ParallelIndexer(int numOfThreads,
                       java.lang.String indexPath,
                       java.io.File imageList,
                       int numOfClusters,
                       int numOfDocsForCodebooks,
                       java.lang.Class<? extends AbstractAggregator> aggregator)

ParallelIndexer

public ParallelIndexer(int numOfThreads,
                       java.lang.String indexPath,
                       java.io.File imageList,
                       int[] numOfClusters,
                       int numOfDocsForCodebooks)

ParallelIndexer

public ParallelIndexer(int numOfThreads,
                       java.lang.String indexPath,
                       java.io.File imageList,
                       int[] numOfClusters,
                       int numOfDocsForCodebooks,
                       java.lang.Class<? extends AbstractAggregator> aggregator)

ParallelIndexer

public ParallelIndexer(int numOfThreads,
                       java.lang.String indexPath,
                       java.io.File imageList,
                       boolean overWrite)

ParallelIndexer

public ParallelIndexer(int numOfThreads,
                       java.lang.String indexPath,
                       java.io.File imageList,
                       java.lang.String fromIndexPath)

方法详细资料

main

public static void main(java.lang.String[] args)

addExtractor

public void addExtractor(java.lang.Class<? extends Extractor> extractorClass)

addExtractor

public void addExtractor(java.lang.Class<? extends GlobalFeature> globalFeatureClass,
                         SimpleExtractor.KeypointDetector detector)

addExtractor

public void addExtractor(java.lang.Class<? extends GlobalFeature> globalFeatureClass,
                         SimpleExtractor.KeypointDetector detector,
                         int numKeyPoints)

addExtractor

public void addExtractor(java.lang.Class<? extends LocalFeatureExtractor> localFeatureExtractorClass,
                         Cluster[] codebook)

addExtractor

public void addExtractor(java.lang.Class<? extends GlobalFeature> globalFeatureClass,
                         SimpleExtractor.KeypointDetector detector,
                         Cluster[] codebook)

addExtractor

public void addExtractor(java.lang.Class<? extends LocalFeatureExtractor> localFeatureExtractorClass,
                         java.util.LinkedList<Cluster[]> codebooks)

addExtractor

public void addExtractor(java.lang.Class<? extends GlobalFeature> globalFeatureClass,
                         SimpleExtractor.KeypointDetector detector,
                         java.util.LinkedList<Cluster[]> codebooks)

setCustomDocumentBuilder

public void setCustomDocumentBuilder(java.lang.Class<? extends DocumentBuilder> customDocumentBuilder)

WARNING!! This can be used in order to use a different DocumentBuilder than the GlobalDocumentBuilder, LocalDocumentBuilder or the SimpleDocumentBuilder. Every time only one custom DocumentBuilder can be used. At the same time, using the addExtractor methods one can add other builders to be used at the same time. BUT when using a custom DocumentBuilder, sampling can be used for Local Features. This means that, if you want to use a LocalDocumentBuilder or SimpleDocumentBuilder, you can use them, only combined with pre-computed codebooks!!

参数:

customDocumentBuilder -

run

public void run()

指定者:

run 在接口中 java.lang.Runnable

getPercentageDone

public double getPercentageDone()

getImagePreprocessor

public ImagePreprocessor getImagePreprocessor()

setImagePreprocessor

public void setImagePreprocessor(ImagePreprocessor imagePreprocessor)

hasEnded

public boolean hasEnded()