LIRe 源代码分析 3：基本接口（ImageSearcher）
=====================================================
LIRe源代码分析系列文章列表：
LIRe 源代码分析 2：基本接口（DocumentBuilder）
LIRe 源代码分析 3：基本接口（ImageSearcher）
LIRe 源代码分析 4：建立索引（DocumentBuilder）[以颜色布局为例]
LIRe 源代码分析 6：检索（ImageSearcher）[以颜色布局为例]
=====================================================
上篇文章介绍了LIRe源代码里的DocumentBuilder的几个基本接口。本文继续研究一下源代码里的ImageSearcher的几个基本接口。
下面来看看与ImageSearcher相关的类的定义：
ImageSearcher：接口，定义了基本的方法。
AbstractImageSearcher：纯虚类，实现了ImageSearcher接口。
ImageSearcherFactory：用于创建ImageSearcher。
ImageSearcher相关的类的继承关系如下图所示。可见，各种算法类都继承了AbstractImageSearcher，而AbstractImageSearcher实现了ImageSearcher接口。
此外还有一个结构体：
ImageSearchHits：用于存储搜索的结果。
详细的源代码如下所示：
ImageSearcher
/*
* This file is part of the LIRe project: http://www.semanticmetadata.net/lire
* LIRe is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* LIRe is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with LIRe; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*
* We kindly ask you to refer the following paper in any publication mentioning Lire:
*
* Lux Mathias, Savvas A. Chatzichristofis. Lire: Lucene Image Retrieval 鈥�
* An Extensible Java CBIR Library. In proceedings of the 16th ACM International
* Conference on Multimedia, pp. 1085-1088, Vancouver, Canada, 2008
*
* http://doi.acm.org/10.1145/1459359.1459577
*
* Copyright statement:
* --------------------
* (c) 2002-2011 by Mathias Lux (mathias@juggle.at)
* http://www.semanticmetadata.net/lire
*/
package net.semanticmetadata.lire;
import org.apache.lucene.document.Document;
import org.apache.lucene.index.IndexReader;
import java.awt.image.BufferedImage;
import java.io.IOException;
import java.io.InputStream;
import java.util.Set;
/**
* <h2>Searching in an Index</h2>
* Use the ImageSearcherFactory for creating an ImageSearcher, which will retrieve the images
* for you from the index.
* <p/>
* <pre>
* IndexReader reader = IndexReader.open(indexPath);
* ImageSearcher searcher = ImageSearcherFactory.createDefaultSearcher();
* FileInputStream imageStream = new FileInputStream("image.jpg");
* BufferedImage bimg = ImageIO.read(imageStream);
* // searching for an image:
* ImageSearchHits hits = null;
* hits = searcher.search(bimg, reader);
* for (int i = 0; i < 5; i++) {
* System.out.println(hits.score(i) + ": " + hits.doc(i).getField(DocumentBuilder.FIELD_NAME_IDENTIFIER).stringValue());
* }
*
* // searching for a document:
* Document document = hits.doc(0);
* hits = searcher.search(document, reader);
* for (int i = 0; i < 5; i++) {
* System.out.println(hits.score(i) + ": " + hits.doc(i).getField(DocumentBuilder.FIELD_NAME_IDENTIFIER).stringValue());
* }
* </pre>
* <p/>
* This file is part of the Caliph and Emir project: http://www.SemanticMetadata.net
* <br>Date: 01.02.2006
* <br>Time: 00:09:42
*
* @author Mathias Lux, mathias@juggle.at
*/
public interface ImageSearcher {
/**
* Searches for images similar to the given image.
*
* @param image the example image to search for.
* @param reader the IndexReader which is used to dsearch through the images.
* @return a sorted list of hits.
* @throws java.io.IOException in case exceptions in the reader occurs
*/
public ImageSearchHits search(BufferedImage image, IndexReader reader) throws IOException;
/**
* Searches for images similar to the given image, defined by the Document from the index.
*
* @param doc the example image to search for.
* @param reader the IndexReader which is used to dsearch through the images.
* @return a sorted list of hits.
* @throws java.io.IOException in case exceptions in the reader occurs
*/
public ImageSearchHits search(Document doc, IndexReader reader) throws IOException;
/**
* Searches for images similar to the given image.
*
* @param image the example image to search for.
* @param reader the IndexReader which is used to dsearch through the images.
* @return a sorted list of hits.
* @throws IOException in case the image could not be read from stream.
*/
public ImageSearchHits search(InputStream image, IndexReader reader) throws IOException;
/**
* Identifies duplicates in the database.
*
* @param reader the IndexReader which is used to dsearch through the images.
* @return a sorted list of hits.
* @throws IOException in case the image could not be read from stream.
*/
public ImageDuplicates findDuplicates(IndexReader reader) throws IOException;
/**
* Modifies the given search by the provided positive and negative examples. This process follows the idea
* of relevance feedback.
*
* @param originalSearch
* @param positives
* @param negatives
* @return
*/
public ImageSearchHits relevanceFeedback(ImageSearchHits originalSearch,
Set<Document> positives, Set<Document> negatives);
}
从接口的源代码可以看出，提供了5个方法，其中有3个名字都叫search()，只是参数不一样。一个是BufferedImage，一个是Document，而另一个是InputStream。
AbstractImageSearcher
/*
* This file is part of the LIRe project: http://www.semanticmetadata.net/lire
* LIRe is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* LIRe is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with LIRe; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*
* We kindly ask you to refer the following paper in any publication mentioning Lire:
*
* Lux Mathias, Savvas A. Chatzichristofis. Lire: Lucene Image Retrieval 鈥�
* An Extensible Java CBIR Library. In proceedings of the 16th ACM International
* Conference on Multimedia, pp. 1085-1088, Vancouver, Canada, 2008
*
* http://doi.acm.org/10.1145/1459359.1459577
*
* Copyright statement:
* --------------------
* (c) 2002-2011 by Mathias Lux (mathias@juggle.at)
* http://www.semanticmetadata.net/lire
*/
package net.semanticmetadata.lire;
import org.apache.lucene.document.Document;
import org.apache.lucene.index.IndexReader;
import javax.imageio.ImageIO;
import java.awt.image.BufferedImage;
import java.io.IOException;
import java.io.InputStream;
import java.util.Set;
/**
* Abstract ImageSearcher, which uses javax.imageio.ImageIO to create a BufferedImage
* from an InputStream.
* <p/>
* This file is part of the Caliph and Emir project: http://www.SemanticMetadata.net
* <br>Date: 01.02.2006
* <br>Time: 00:13:16
*
* @author Mathias Lux, mathias@juggle.at
*/
public abstract class AbstractImageSearcher implements ImageSearcher {
/**
* Searches for images similar to the given image. This simple implementation uses
* {@link ImageSearcher#search(java.awt.image.BufferedImage, org.apache.lucene.index.IndexReader)},
* the image is read using javax.imageio.ImageIO.
*
* @param image the example image to search for.
* @param reader the IndexReader which is used to dsearch through the images.
* @return a sorted list of hits.
* @throws IOException in case the image could not be read from stream.
*/
public ImageSearchHits search(InputStream image, IndexReader reader) throws IOException {
BufferedImage bufferedImage = ImageIO.read(image);
return search(bufferedImage, reader);
}
public ImageSearchHits relevanceFeedback(ImageSearchHits originalSearch, Set<Document> positives, Set<Document> negatives) {
throw new UnsupportedOperationException("Not implemented yet for this kind of searcher!");
}
}
从代码中可以看出AbstractImageSearcher实现了ImageSearcher接口。其中的search(InputStream image, IndexReader reader)方法调用了search(BufferedImage image, IndexReader reader)方法。说白了，就是把2个函数的功能合并为一个函数。
ImageSearcherFactory
/*
* This file is part of the LIRe project: http://www.semanticmetadata.net/lire
* LIRe is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* LIRe is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with LIRe; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*
* We kindly ask you to refer the following paper in any publication mentioning Lire:
*
* Lux Mathias, Savvas A. Chatzichristofis. Lire: Lucene Image Retrieval 鈥�
* An Extensible Java CBIR Library. In proceedings of the 16th ACM International
* Conference on Multimedia, pp. 1085-1088, Vancouver, Canada, 2008
*
* http://doi.acm.org/10.1145/1459359.1459577
*
* Copyright statement:
* ~~~~~~~~~~~~~~~~~~~~
* (c) 2002-2011 by Mathias Lux (mathias@juggle.at)
* http://www.semanticmetadata.net/lire
*/
package net.semanticmetadata.lire;
import net.semanticmetadata.lire.imageanalysis.*;
import net.semanticmetadata.lire.impl.CorrelogramImageSearcher;
import net.semanticmetadata.lire.impl.GenericFastImageSearcher;
import net.semanticmetadata.lire.impl.SimpleImageSearcher;
/**
* <h2>Searching in an Index</h2>
* Use the ImageSearcherFactory for creating an ImageSearcher, which will retrieve the images
* for you from the index.
* <p/>
* <pre>
* IndexReader reader = IndexReader.open(indexPath);
* ImageSearcher searcher = ImageSearcherFactory.createDefaultSearcher();
* FileInputStream imageStream = new FileInputStream("image.jpg");
* BufferedImage bimg = ImageIO.read(imageStream);
* // searching for an image:
* ImageSearchHits hits = null;
* hits = searcher.search(bimg, reader);
* for (int i = 0; i < 5; i++) {
* System.out.println(hits.score(i) + ": " + hits.doc(i).getField(DocumentBuilder.FIELD_NAME_IDENTIFIER).stringValue());
* }
*
* // searching for a document:
* Document document = hits.doc(0);
* hits = searcher.search(document, reader);
* for (int i = 0; i < 5; i++) {
* System.out.println(hits.score(i) + ": " + hits.doc(i).getField(DocumentBuilder.FIELD_NAME_IDENTIFIER).stringValue());
* }
* </pre>
* <p/>
* This file is part of the Caliph and Emir project: http://www.SemanticMetadata.net
* <br>Date: 03.02.2006
* <br>Time: 00:30:07
*
* @author Mathias Lux, mathias@juggle.at
*/
public class ImageSearcherFactory {
/**
* Default number of maximum hits.
*/
public static int NUM_MAX_HITS = 100;
/**
* Creates a new simple image searcher with the desired number of maximum hits.
*
* @param maximumHits
* @return the searcher instance
* @deprecated Use ColorLayout, EdgeHistogram and ScalableColor features instead.
*/
public static ImageSearcher createSimpleSearcher(int maximumHits) {
return ImageSearcherFactory.createColorLayoutImageSearcher(maximumHits);
}
/**
* Returns a new default ImageSearcher with a predefined number of maximum
* hits defined in the {@link ImageSearcherFactory#NUM_MAX_HITS} based on the {@link net.semanticmetadata.lire.imageanalysis.CEDD} feature
*
* @return the searcher instance
*/
public static ImageSearcher createDefaultSearcher() {
return new GenericFastImageSearcher(NUM_MAX_HITS, CEDD.class, DocumentBuilder.FIELD_NAME_CEDD);
}
/**
* Returns a new ImageSearcher with the given number of maximum hits
* which only takes the overall color into account. texture and color
* distribution are ignored.
*
* @param maximumHits defining how many hits are returned in max (e.g. 100 would be ok)
* @return the ImageSearcher
* @see ImageSearcher
* @deprecated Use ColorHistogram or ScalableColor instead
*/
public static ImageSearcher createColorOnlySearcher(int maximumHits) {
return ImageSearcherFactory.createScalableColorImageSearcher(maximumHits);
}
/**
* Returns a new ImageSearcher with the given number of maximum hits and
* the specified weights on the different matching aspects. All weights
* should be in [0,1] whereas a weight of 0 implies that the feature is
* not taken into account for searching. Note that the effect is relative and
* can only be fully applied if the {@link DocumentBuilderFactory#getExtensiveDocumentBuilder() extensive DocumentBuilder}
* is used.
*
* @param maximumHits defining how many hits are returned in max
* @param colorHistogramWeight a weight in [0,1] defining the importance of overall color in the images
* @param colorDistributionWeight a weight in [0,1] defining the importance of color distribution (which color where) in the images
* @param textureWeight defining the importance of texture (which edges where) in the images
* @return the searcher instance or NULL if the weights are not appropriate, eg. all 0 or not in [0,1]
* @see DocumentBuilderFactory
* @deprecated Use ColorLayout, EdgeHistogram and ScalableColor features instead.
*/
public static ImageSearcher createWeightedSearcher(int maximumHits,
float colorHistogramWeight,
float colorDistributionWeight,
float textureWeight) {
if (isAppropriateWeight(colorHistogramWeight)
&& isAppropriateWeight(colorDistributionWeight)
&& isAppropriateWeight(textureWeight)
&& (colorHistogramWeight + colorDistributionWeight + textureWeight > 0f))
return new SimpleImageSearcher(maximumHits, colorHistogramWeight, colorDistributionWeight, textureWeight);
else
return null;
}
/**
* Create and return an ImageSearcher for the {@link net.semanticmetadata.lire.imageanalysis.AutoColorCorrelogram}
* image feature. Be sure to use the same options for the ImageSearcher as you used for the DocumentBuilder.
*
* @param maximumHits number of hits returned.
* @return
*/
public static ImageSearcher createAutoColorCorrelogramImageSearcher(int maximumHits) {
return new GenericFastImageSearcher(maximumHits, AutoColorCorrelogram.class, DocumentBuilder.FIELD_NAME_AUTOCOLORCORRELOGRAM);
// return new CorrelogramImageSearcher(maximumHits, AutoColorCorrelogram.Mode.SuperFast);
}
/**
* Create and return an ImageSearcher for the {@link net.semanticmetadata.lire.imageanalysis.AutoColorCorrelogram}
* image feature. Be sure to use the same options for the ImageSearcher as you used for the DocumentBuilder.
*
* @param maximumHits number of hits returned.
* @return
* @deprecated Use #createAutoColorCorrelogramImageSearcher instead
*/
public static ImageSearcher createFastCorrelogramImageSearcher(int maximumHits) {
return new CorrelogramImageSearcher(maximumHits, AutoColorCorrelogram.Mode.SuperFast);
}
/**
* Create and return an ImageSearcher for the {@link net.semanticmetadata.lire.imageanalysis.CEDD}
* image feature. Be sure to use the same options for the ImageSearcher as you used for the DocumentBuilder.
*
* @param maximumHits
* @return
*/
public static ImageSearcher createCEDDImageSearcher(int maximumHits) {
// return new CEDDImageSearcher(maximumHits);
return new GenericFastImageSearcher(maximumHits, CEDD.class, DocumentBuilder.FIELD_NAME_CEDD);
}
/**
* Create and return an ImageSearcher for the {@link net.semanticmetadata.lire.imageanalysis.FCTH}
* image feature. Be sure to use the same options for the ImageSearcher as you used for the DocumentBuilder.
*
* @param maximumHits
* @return
*/
public static ImageSearcher createFCTHImageSearcher(int maximumHits) {
// return new GenericImageSearcher(maximumHits, FCTH.class, DocumentBuilder.FIELD_NAME_FCTH);
return new GenericFastImageSearcher(maximumHits, FCTH.class, DocumentBuilder.FIELD_NAME_FCTH);
}
/**
* Create and return an ImageSearcher for the {@link net.semanticmetadata.lire.imageanalysis.JCD}
* image feature. Be sure to use the same options for the ImageSearcher as you used for the DocumentBuilder.
*
* @param maximumHits
* @return
*/
public static ImageSearcher createJCDImageSearcher(int maximumHits) {
return new GenericFastImageSearcher(maximumHits, JCD.class, DocumentBuilder.FIELD_NAME_JCD);
}
/**
* Create and return an ImageSearcher for the {@link net.semanticmetadata.lire.imageanalysis.JpegCoefficientHistogram}
* image feature. Be sure to use the same options for the ImageSearcher as you used for the DocumentBuilder.
*
* @param maximumHits
* @return
*/
public static ImageSearcher createJpegCoefficientHistogramImageSearcher(int maximumHits) {
return new GenericFastImageSearcher(maximumHits, JpegCoefficientHistogram.class, DocumentBuilder.FIELD_NAME_JPEGCOEFFS);
}
/**
* Create and return an ImageSearcher for the {@link net.semanticmetadata.lire.imageanalysis.SimpleColorHistogram}
* image feature. Be sure to use the same options for the ImageSearcher as you used for the DocumentBuilder.
*
* @param maximumHits
* @return
*/
public static ImageSearcher createColorHistogramImageSearcher(int maximumHits) {
// return new GenericImageSearcher(maximumHits, SimpleColorHistogram.class, DocumentBuilder.FIELD_NAME_COLORHISTOGRAM);
return new GenericFastImageSearcher(maximumHits, SimpleColorHistogram.class, DocumentBuilder.FIELD_NAME_COLORHISTOGRAM);
}
/**
* Create and return an ImageSearcher for the {@link net.semanticmetadata.lire.imageanalysis.Tamura}
* image feature. Be sure to use the same options for the ImageSearcher as you used for the DocumentBuilder.
*
* @param maximumHits
* @return
*/
public static ImageSearcher createTamuraImageSearcher(int maximumHits) {
return new GenericFastImageSearcher(maximumHits, Tamura.class, DocumentBuilder.FIELD_NAME_TAMURA);
}
/**
* Create and return an ImageSearcher for the {@link net.semanticmetadata.lire.imageanalysis.Gabor}
* image feature. Be sure to use the same options for the ImageSearcher as you used for the DocumentBuilder.
*
* @param maximumHits
* @return
*/
public static ImageSearcher createGaborImageSearcher(int maximumHits) {
return new GenericFastImageSearcher(maximumHits, Gabor.class, DocumentBuilder.FIELD_NAME_GABOR);
}
/**
* Create and return an ImageSearcher for the {@link net.semanticmetadata.lire.imageanalysis.ColorLayout}
* image feature using the byte[] serialization. Be sure to use the same options for the ImageSearcher as
* you used for the DocumentBuilder.
*
* @param maximumHits
* @return
*/
public static ImageSearcher createColorLayoutImageSearcher(int maximumHits) {
return new GenericFastImageSearcher(maximumHits, ColorLayout.class, DocumentBuilder.FIELD_NAME_COLORLAYOUT);
}
/**
* Create and return an ImageSearcher for the {@link net.semanticmetadata.lire.imageanalysis.ScalableColor}
* image feature using the byte[] serialization. Be sure to use the same options for the ImageSearcher as
* you used for the DocumentBuilder.
*
* @param maximumHits
* @return
*/
public static ImageSearcher createScalableColorImageSearcher(int maximumHits) {
return new GenericFastImageSearcher(maximumHits, ScalableColor.class, DocumentBuilder.FIELD_NAME_SCALABLECOLOR);
}
/**
* Create and return an ImageSearcher for the {@link net.semanticmetadata.lire.imageanalysis.EdgeHistogram}
* image feature using the byte[] serialization. Be sure to use the same options for the ImageSearcher as
* you used for the DocumentBuilder.
*
* @param maximumHits
* @return
*/
public static ImageSearcher createEdgeHistogramImageSearcher(int maximumHits) {
return new GenericFastImageSearcher(maximumHits, EdgeHistogram.class, DocumentBuilder.FIELD_NAME_EDGEHISTOGRAM);
}
/**
* Checks if the weight is in [0,1]
*
* @param f the weight to check
* @return true if the weight is in [0,1], false otherwise
*/
private static boolean isAppropriateWeight(float f) {
boolean result = false;
if (f <= 1f && f >= 0) result = true;
return result;
}
}
ImageSearcherFactory是用于创建ImageSearcher的。里面有各种create****ImageSearcher()。每个函数的作用在注释中都有详细的说明。
ImageSearchHits
/*
* This file is part of the LIRe project: http://www.semanticmetadata.net/lire
* LIRe is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* LIRe is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with LIRe; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*
* We kindly ask you to refer the following paper in any publication mentioning Lire:
*
* Lux Mathias, Savvas A. Chatzichristofis. Lire: Lucene Image Retrieval 鈥�
* An Extensible Java CBIR Library. In proceedings of the 16th ACM International
* Conference on Multimedia, pp. 1085-1088, Vancouver, Canada, 2008
*
* http://doi.acm.org/10.1145/1459359.1459577
*
* Copyright statement:
* --------------------
* (c) 2002-2011 by Mathias Lux (mathias@juggle.at)
* http://www.semanticmetadata.net/lire
*/
package net.semanticmetadata.lire;
import org.apache.lucene.document.Document;
/**
* This class simulates the original Lucene Hits object.
* Please note the only a certain number of results are returned.<br>
* <p/>
* This file is part of the Caliph and Emir project: http://www.SemanticMetadata.net
* <br>Date: 02.02.2006
* <br>Time: 23:45:20
*
* @author Mathias Lux, mathias@juggle.at
*/
public interface ImageSearchHits {
/**
* Returns the size of the result list.
*
* @return the size of the result list.
*/
public int length();
/**
* Returns the score of the document at given position.
* Please note that the score in this case is a distance,
* which means a score of 0 denotes the best possible hit.
* The result list starts with position 0 as everything
* in computer science does.
*
* @param position defines the position
* @return the score of the document at given position. The lower the better (its a distance measure).
*/
public float score(int position);
/**
* Returns the document at given position
*
* @param position defines the position.
* @return the document at given position.
*/
public Document doc(int position);
}
该类主要用于存储ImageSearcher类中search()方法返回的结果。
SimpleImageSearchHits是ImageSearcher的实现。该类的源代码如下所示：
/*
* This file is part of the LIRe project: http://www.semanticmetadata.net/lire
* LIRe is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* LIRe is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with LIRe; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*
* We kindly ask you to refer the following paper in any publication mentioning Lire:
*
* Lux Mathias, Savvas A. Chatzichristofis. Lire: Lucene Image Retrieval 鈥�
* An Extensible Java CBIR Library. In proceedings of the 16th ACM International
* Conference on Multimedia, pp. 1085-1088, Vancouver, Canada, 2008
*
* http://doi.acm.org/10.1145/1459359.1459577
*
* Copyright statement:
* --------------------
* (c) 2002-2011 by Mathias Lux (mathias@juggle.at)
* http://www.semanticmetadata.net/lire
*/
package net.semanticmetadata.lire.impl;
import net.semanticmetadata.lire.ImageSearchHits;
import org.apache.lucene.document.Document;
import java.util.ArrayList;
import java.util.Collection;
import java.util.Iterator;
/**
* This file is part of the Caliph and Emir project: http://www.SemanticMetadata.net
* <br>Date: 02.02.2006
* <br>Time: 23:56:15
*
* @author Mathias Lux, mathias@juggle.at
*/
public class SimpleImageSearchHits implements ImageSearchHits {
ArrayList<SimpleResult> results;
public SimpleImageSearchHits(Collection<SimpleResult> results, float maxDistance) {
this.results = new ArrayList<SimpleResult>(results.size());
this.results.addAll(results);
// this step normalizes and inverts the distance ...
// although its now a score or similarity like measure its further called distance
for (Iterator<SimpleResult> iterator = this.results.iterator(); iterator.hasNext(); ) {
SimpleResult result = iterator.next();
result.setDistance(1f - result.getDistance() / maxDistance);
}
}
/**
* Returns the size of the result list.
*
* @return the size of the result list.
*/
public int length() {
return results.size();
}
/**
* Returns the score of the document at given position.
* Please note that the score in this case is a distance,
* which means a score of 0 denotes the best possible hit.
* The result list starts with position 0 as everything
* in computer science does.
*
* @param position defines the position
* @return the score of the document at given position. The lower the better (its a distance measure).
*/
public float score(int position) {
return results.get(position).getDistance();
}
/**
* Returns the document at given position
*
* @param position defines the position.
* @return the document at given position.
*/
public Document doc(int position) {
return results.get(position).getDocument();
}
private float sigmoid(float f) {
double result = 0f;
result = -1d + 2d / (1d + Math.exp(-2d * f / 0.6));
return (float) (1d - result);
}
}
可以看出检索的结果是存在名为results的ArrayList<SimpleResult> 类型的变量中的。
LIRe 源代码分析 3：基本接口（ImageSearcher）的更多相关文章
- LIRe 源代码分析 6：检索（ImageSearcher）[以颜色布局为例]
===================================================== LIRe源代码分析系列文章列表: LIRe 源代码分析 1:整体结构 LIRe 源代码分析 ...
- LIRe 源代码分析 2：基本接口（DocumentBuilder）
===================================================== LIRe源代码分析系列文章列表: LIRe 源代码分析 1:整体结构 LIRe 源代码分析 ...
- LIRe 源代码分析 7：算法类[以颜色布局为例]
===================================================== LIRe源代码分析系列文章列表: LIRe 源代码分析 1:整体结构 LIRe 源代码分析 ...
- LIRe 源代码分析 5：提取特征向量[以颜色布局为例]
===================================================== LIRe源代码分析系列文章列表: LIRe 源代码分析 1:整体结构 LIRe 源代码分析 ...
- LIRe 源代码分析 4：建立索引（DocumentBuilder）[以颜色布局为例]
===================================================== LIRe源代码分析系列文章列表: LIRe 源代码分析 1:整体结构 LIRe 源代码分析 ...
- LIRe 源代码分析 1：整体结构
===================================================== LIRe源代码分析系列文章列表: LIRe 源代码分析 1:整体结构 LIRe 源代码分析 ...
- 转：LIRe 源代码分析
1:整体结构 LIRE(Lucene Image REtrieval)提供一种的简单方式来创建基于图像特性的Lucene索引.利用该索引就能够构建一个基于内容的图像检索(content- based ...
- Android系统进程间通信Binder机制在应用程序框架层的Java接口源代码分析
文章转载至CSDN社区罗升阳的安卓之旅,原文地址:http://blog.csdn.net/luoshengyang/article/details/6642463 在前面几篇文章中,我们详细介绍了A ...
- Android系统进程间通信（IPC）机制Binder中的Client获得Server远程接口过程源代码分析
文章转载至CSDN社区罗升阳的安卓之旅,原文地址:http://blog.csdn.net/luoshengyang/article/details/6633311 在上一篇文章中,我 们分析了And ...
随机推荐
- [boost] build boost with intel compiler 16.0.XXX
Introduction There are few information about how to compile boost with Intel compiler. This article ...
- shell编程--流程控制for,do-while,if-then,break,continue,case等
2.5 流程控制 2.5.1 if语法 1.语法格式 if condition then statements [elif condition then statements. ..] ...
- Android基于JsBridge封装的高效带加载进度的WebView
Tamic http://blog.csdn.net/sk719887916/article/details/52402470 概述 从去年4月项目就一直用起了JsBridge,前面也针对jsBrid ...
- GIF动态图制作
GIF动态图制作 博客写了也有一阵了,一直好奇大牛的博客里demo的动态图是怎么做的,今天抽空研究了一下,找了一个软件,以后再发现有好的工具再继续推荐 GIF制作工具--LICEcap 效果要比下面的 ...
- EBS技术开发之VPD策略
VPD (虚拟专用数据库的简称),主要作用是根据运行环境的上下文,隐式的添加条 件. 好处是在数据库层解决细粒度的角色权限访问,避免在中间层写大量代码:坏处 是数据屏蔽的逻辑太隐蔽了,对于分析查找问题 ...
- [python]mysql数据缓存到redis中 取出时候编码问题
描述: 一个web服务,原先的业务逻辑是把mysql查询的结果缓存在redis中一个小时,加快请求的响应. 现在有个问题就是根据请求的指定的编码返回对应编码的response. 首先是要修改响应的bo ...
- Protobuf-net判断字段是否有值
Protobuf-net判断字段是否有值Unity3d使用Protobuf-net序列化数据与服务器通信,但是发现默认情况下,Protobuf-net生成的cs文件中没有接口判断可选参数是否有值.需有 ...
- Struts 1之DispatchAction
DispatchAction是struts 1 的内置通用分发器 import org.apache.struts.actions.DispatchAction; public class UserA ...
- 【Netty源码分析】ChannelPipeline（二）
在上一篇博客[Netty源码学习]ChannelPipeline(一)中我们只是大体介绍了ChannelPipeline相关的知识,其实介绍的并不详细,接下来我们详细介绍一下ChannelPipeli ...
- 15 ActionProvider代码例子
Menu文件夹下代码: <menu xmlns:android="http://schemas.android.com/apk/res/android" > <! ...