/*
* 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 any or one of the following publications in
* any publication mentioning or employing Lire:
*
* Lux Mathias, Savvas A. Chatzichristofis. Lire: Lucene Image Retrieval –
* An Extensible Java CBIR Library. In proceedings of the 16th ACM International
* Conference on Multimedia, pp. 1085-1088, Vancouver, Canada, 2008
* URL: http://doi.acm.org/10.1145/1459359.1459577
*
* Lux Mathias. Content Based Image Retrieval with LIRE. In proceedings of the
* 19th ACM International Conference on Multimedia, pp. 735-738, Scottsdale,
* Arizona, USA, 2011
* URL: http://dl.acm.org/citation.cfm?id=2072432
*
* Mathias Lux, Oge Marques. Visual Information Retrieval using Java and LIRE
* Morgan & Claypool, 2013
* URL: http://www.morganclaypool.com/doi/abs/10.2200/S00468ED1V01Y201301ICR025
*
* Copyright statement:
* ====================
* (c) 2002-2013 by Mathias Lux (mathias@juggle.at)
* http://www.semanticmetadata.net/lire, http://www.lire-project.net
*
* Updated: 18.01.15 08:07
*/
package net.semanticmetadata.lire;
import junit.framework.TestCase;
import net.semanticmetadata.lire.imageanalysis.CEDD;
import net.semanticmetadata.lire.imageanalysis.LireFeature;
import net.semanticmetadata.lire.imageanalysis.ScalableColor;
import net.semanticmetadata.lire.impl.ChainedDocumentBuilder;
import net.semanticmetadata.lire.impl.GenericFastImageSearcher;
import net.semanticmetadata.lire.impl.SimpleResult;
import net.semanticmetadata.lire.impl.VisualWordsImageSearcher;
import net.semanticmetadata.lire.impl.custom.SingleNddCeddImageSearcher;
import org.apache.lucene.document.Document;
import org.apache.lucene.index.DirectoryReader;
import org.apache.lucene.index.IndexReader;
import org.apache.lucene.store.FSDirectory;
import org.apache.lucene.util.BytesRef;
import javax.imageio.ImageIO;
import java.awt.image.BufferedImage;
import java.io.*;
import java.util.Iterator;
import java.util.LinkedList;
import java.util.concurrent.LinkedBlockingQueue;
/**
* This file is part of Caliph & Emir
* Date: 03.02.2006
* Time: 00:29:56
*
* @author Mathias Lux, mathias@juggle.at
*/
public class TestImageSearcher extends TestCase {
private String[] testFiles = new String[]{"img01.JPG", "img02.JPG", "img03.JPG", "img04.JPG", "img05.JPG",
"img06.JPG", "img07.JPG", "img08.JPG", "img08a.JPG"};
private String testFilesPath = "src/test/resources/images/";
private String indexPath = "test-index-small";
private int numsearches = 25;
private DocumentBuilder getDocumentBuilder() {
ChainedDocumentBuilder result = new ChainedDocumentBuilder();
result.addBuilder(DocumentBuilderFactory.getAutoColorCorrelogramDocumentBuilder());
result.addBuilder(DocumentBuilderFactory.getScalableColorBuilder());
result.addBuilder(DocumentBuilderFactory.getCEDDDocumentBuilder());
result.addBuilder(DocumentBuilderFactory.getColorHistogramDocumentBuilder());
result.addBuilder(DocumentBuilderFactory.getColorLayoutBuilder());
result.addBuilder(DocumentBuilderFactory.getTamuraDocumentBuilder());
result.addBuilder(DocumentBuilderFactory.getEdgeHistogramBuilder());
result.addBuilder(DocumentBuilderFactory.getFCTHDocumentBuilder());
result.addBuilder(DocumentBuilderFactory.getGaborDocumentBuilder());
result.addBuilder(DocumentBuilderFactory.getJCDDocumentBuilder());
result.addBuilder(DocumentBuilderFactory.getJpegCoefficientHistogramDocumentBuilder());
return result;
}
public void testSearch() throws IOException {
IndexReader reader = DirectoryReader.open(FSDirectory.open(new File(indexPath)));
int numDocs = reader.numDocs();
System.out.println("numDocs = " + numDocs);
ImageSearcher searcher = ImageSearcherFactory.createScalableColorImageSearcher(50);
FileInputStream imageStream = new FileInputStream(testFilesPath + testFiles[0]);
BufferedImage bimg = ImageIO.read(imageStream);
ImageSearchHits hits = null;
long time = System.currentTimeMillis();
for (int i = 0; i < numsearches; i++) {
hits = searcher.search(bimg, reader);
}
time = System.currentTimeMillis() - time;
System.out.println(((float) time / (float) numsearches) + " ms per search with image, averaged on " + numsearches);
for (int i = 0; i < 5; i++) {
System.out.println(hits.score(i) + ": " + hits.doc(i).getField(DocumentBuilder.FIELD_NAME_IDENTIFIER).stringValue());
}
Document document = hits.doc(0);
time = System.currentTimeMillis();
for (int i = 0; i < numsearches; i++) {
hits = searcher.search(document, reader);
}
time = System.currentTimeMillis() - time;
System.out.println(((float) time / (float) numsearches) + " ms per search with document, averaged on " + numsearches);
for (int i = 0; i < 5; i++) {
System.out.println(hits.score(i) + ": " + hits.doc(i).getField(DocumentBuilder.FIELD_NAME_IDENTIFIER).stringValue());
}
document = getDocumentBuilder().createDocument(bimg, testFilesPath + testFiles[0]);
time = System.currentTimeMillis();
for (int i = 0; i < numsearches; i++) {
hits = searcher.search(document, reader);
}
time = System.currentTimeMillis() - time;
System.out.println(((float) time / (float) numsearches) + " ms per search with document, averaged on " + numsearches);
for (int i = 0; i < 5; i++) {
System.out.println(hits.score(i) + ": " + hits.doc(i).getField(DocumentBuilder.FIELD_NAME_IDENTIFIER).stringValue());
}
}
public void testFindDuplicates() throws Exception {
IndexReader reader = DirectoryReader.open(FSDirectory.open(new File(indexPath)));
int numDocs = reader.numDocs();
System.out.println("numDocs = " + numDocs);
ImageSearcher searcher = ImageSearcherFactory.createCEDDImageSearcher(50);
ImageDuplicates imageDuplicates = searcher.findDuplicates(reader);
if (imageDuplicates == null) {
System.out.println("No duplicates found");
return;
}
for (int i = 0; i < imageDuplicates.length(); i++) {
System.out.println(imageDuplicates.getDuplicate(i).toString());
}
}
public void testCorrelationSearch() throws IOException {
IndexReader reader = DirectoryReader.open(FSDirectory.open(new File(indexPath)));
int numDocs = reader.numDocs();
System.out.println("numDocs = " + numDocs);
ImageSearcher searcher = ImageSearcherFactory.createAutoColorCorrelogramImageSearcher(10);
FileInputStream imageStream = new FileInputStream(testFilesPath + testFiles[0]);
BufferedImage bimg = ImageIO.read(imageStream);
ImageSearchHits hits = null;
long time = System.currentTimeMillis();
// for (int i = 0; i < numsearches; i++) {
hits = searcher.search(bimg, reader);
// }
time = System.currentTimeMillis() - time;
System.out.println(((float) time / (float) numsearches) + " ms per search with image, averaged on " + numsearches);
for (int i = 0; i < hits.length(); i++) {
System.out.println(hits.score(i) + ": " + hits.doc(i).getField(DocumentBuilder.FIELD_NAME_IDENTIFIER).stringValue());
}
Document document = hits.doc(4);
time = System.currentTimeMillis();
for (int i = 0; i < numsearches; i++) {
hits = searcher.search(document, reader);
}
time = System.currentTimeMillis() - time;
System.out.println(((float) time / (float) numsearches) + " ms per search with document, averaged on " + numsearches);
for (int i = 0; i < hits.length(); i++) {
System.out.println(hits.score(i) + ": " + hits.doc(i).getField(DocumentBuilder.FIELD_NAME_IDENTIFIER).stringValue());
}
}
public void testCEDDSearch() throws IOException {
IndexReader reader = DirectoryReader.open(FSDirectory.open(new File(indexPath)));
int numDocs = reader.numDocs();
System.out.println("numDocs = " + numDocs);
ImageSearcher searcher = ImageSearcherFactory.createCEDDImageSearcher(30);
FileInputStream imageStream = new FileInputStream(testFilesPath + testFiles[0]);
BufferedImage bimg = ImageIO.read(imageStream);
ImageSearchHits hits = null;
long time = System.currentTimeMillis();
// for (int i = 0; i < numsearches; i++) {
hits = searcher.search(bimg, reader);
// }
time = System.currentTimeMillis() - time;
System.out.println(((float) time / (float) numsearches) + " ms per search with image, averaged on " + numsearches);
for (int i = 0; i < hits.length(); i++) {
System.out.println(hits.score(i) + ": " + hits.doc(i).getField(DocumentBuilder.FIELD_NAME_IDENTIFIER).stringValue());
}
Document document = hits.doc(4);
time = System.currentTimeMillis();
for (int i = 0; i < numsearches; i++) {
hits = searcher.search(document, reader);
}
time = System.currentTimeMillis() - time;
System.out.println(((float) time / (float) numsearches) + " ms per search with document, averaged on " + numsearches);
for (int i = 0; i < hits.length(); i++) {
System.out.println(hits.score(i) + ": " + hits.doc(i).getField(DocumentBuilder.FIELD_NAME_IDENTIFIER).stringValue());
}
}
public void testSimpleColorHistogramSearch() throws IOException {
IndexReader reader = DirectoryReader.open(FSDirectory.open(new File(indexPath)));
int numDocs = reader.numDocs();
System.out.println("numDocs = " + numDocs);
ImageSearcher searcher = ImageSearcherFactory.createColorHistogramImageSearcher(30);
FileInputStream imageStream = new FileInputStream(testFilesPath + testFiles[0]);
BufferedImage bimg = ImageIO.read(imageStream);
ImageSearchHits hits = null;
long time = System.currentTimeMillis();
// for (int i = 0; i < numsearches; i++) {
hits = searcher.search(bimg, reader);
// }
time = System.currentTimeMillis() - time;
System.out.println(((float) time / (float) numsearches) + " ms per search with image, averaged on " + numsearches);
for (int i = 0; i < hits.length(); i++) {
System.out.println(hits.score(i) + ": " + hits.doc(i).getField(DocumentBuilder.FIELD_NAME_IDENTIFIER).stringValue());
}
Document document = hits.doc(4);
time = System.currentTimeMillis();
for (int i = 0; i < numsearches; i++) {
hits = searcher.search(document, reader);
}
time = System.currentTimeMillis() - time;
System.out.println(((float) time / (float) numsearches) + " ms per search with document, averaged on " + numsearches);
for (int i = 0; i < hits.length(); i++) {
System.out.println(hits.score(i) + ": " + hits.doc(i).getField(DocumentBuilder.FIELD_NAME_IDENTIFIER).stringValue());
}
}
public void testMSERHistogramSearcher() throws IOException {
String query = "312.jpg";
VisualWordsImageSearcher searcher = new VisualWordsImageSearcher(25, DocumentBuilder.FIELD_NAME_MSER + DocumentBuilder.FIELD_NAME_BOVW);
IndexReader reader = DirectoryReader.open(FSDirectory.open(new File("wang-index")));
ImageSearchHits hits = searcher.search(findDoc(reader, query), reader);
saveToHtml("mser", hits, query);
}
private Document findDoc(IndexReader reader, String file) throws IOException {
for (int i = 0; i < reader.numDocs(); i++) {
Document document = reader.document(i);
String s = document.getValues(DocumentBuilder.FIELD_NAME_IDENTIFIER)[0];
if (s.endsWith(File.separator + file)) {
// System.out.println("s = " + s);
return document;
}
}
return null;
}
private void saveToHtml(String prefix, ImageSearchHits hits, String queryImage) throws IOException {
BufferedWriter bw = new BufferedWriter(new FileWriter("results-" + prefix + ".html"));
bw.write("<html>\n" +
"<head><title>Search Results</title></head>\n" +
"<body bgcolor=\"#FFFFFF\">\n");
bw.write("<h3>query</h3>\n");
bw.write("<a href=\"file://" + queryImage + "\"><img src=\"file://" + queryImage + "\"></a><p>\n");
bw.write("<h3>results</h3>\n");
for (int i = 0; i < hits.length(); i++) {
bw.write(hits.score(i) + " - <a href=\"file://" + hits.doc(i).get("descriptorImageIdentifier") + "\"><img src=\"file://" + hits.doc(i).get("descriptorImageIdentifier") + "\"></a><p>\n");
}
bw.write("</body>\n" +
"</html>");
bw.close();
}
public void testCachingSearcher() throws IOException {
IndexReader ir = DirectoryReader.open(FSDirectory.open(new File("C:\\Temp\\test-100k-cedd-idx")));
GenericFastImageSearcher is = new GenericFastImageSearcher(1, CEDD.class, true, ir);
SingleNddCeddImageSearcher nis = new SingleNddCeddImageSearcher(ir);
LinkedList<Document> q = new LinkedList<Document>();
for (int i = 0; i < Math.min(1000, ir.maxDoc()); i++) {
q.add(ir.document(i));
}
long time = System.currentTimeMillis();
int count = 0;
for (Iterator<Document> iterator = q.iterator(); iterator.hasNext(); ) {
Document next = iterator.next();
String id = is.search(next, ir).doc(0).getValues(DocumentBuilder.FIELD_NAME_IDENTIFIER)[0];
CEDD cedd = new CEDD();
BytesRef binaryValue = next.getBinaryValue(cedd.getFieldName());
cedd.setByteArrayRepresentation(binaryValue.bytes, binaryValue.offset, binaryValue.length);
String s = nis.findMostSimilar(cedd).getDocument().getValues(DocumentBuilder.FIELD_NAME_IDENTIFIER)[0];
String qID = next.getValues(DocumentBuilder.FIELD_NAME_IDENTIFIER)[0];
System.out.println(s.equals(id) + " " + id.equals(qID) + " " + qID.equals(s));
count++;
if (count > 100) break;
}
long l = System.currentTimeMillis() - time;
System.out.printf("Tested %d search requests on %d documents: overall time of %d:%02d, %.2f ms per search", count, ir.maxDoc(), l / (1000 * 60), (l / 1000) % 60, ((float) l / (float) count));
}
public void testCustomCachingSearcher() throws IOException {
IndexReader ir = DirectoryReader.open(FSDirectory.open(new File("C:\\Temp\\test-100k-cedd-idx")));
SingleNddCeddImageSearcher is = new SingleNddCeddImageSearcher(ir);
LinkedList<LireFeature> q = new LinkedList<LireFeature>();
for (int i = 0; i < ir.maxDoc(); i++) {
Document d = ir.document(i);
CEDD cedd = new CEDD();
BytesRef binaryValue = d.getBinaryValue(cedd.getFieldName());
cedd.setByteArrayRepresentation(binaryValue.bytes, binaryValue.offset, binaryValue.length);
q.add(cedd);
}
long time = System.currentTimeMillis();
int count = 0;
for (Iterator<LireFeature> iterator = q.iterator(); iterator.hasNext(); ) {
LireFeature next = iterator.next();
is.findMostSimilar(next);
count++;
if (count > 100) break;
}
long l = System.currentTimeMillis() - time;
System.out.printf("Tested %d search requests on %d documents: overall time of %d:%02d, %.2f ms per search", count, ir.maxDoc(), l / (1000 * 60), (l / 1000) % 60, ((float) l / (float) count));
}
public void testCachingSearcherParallel() throws IOException, InterruptedException {
final IndexReader ir = DirectoryReader.open(FSDirectory.open(new File("C:\\Temp\\test-100k-cedd-idx")));
SingleNddCeddImageSearcher is = new SingleNddCeddImageSearcher(ir);
LinkedList<LireFeature> q = new LinkedList<LireFeature>();
for (int i = 0; i < ir.maxDoc(); i++) {
Document d = ir.document(i);
CEDD cedd = new CEDD();
BytesRef binaryValue = d.getBinaryValue(cedd.getFieldName());
cedd.setByteArrayRepresentation(binaryValue.bytes, binaryValue.offset, binaryValue.length);
q.add(cedd);
}
int count = 0;
Thread[] searchers = new Thread[3];
final LinkedBlockingQueue<LireFeature> queryQueue = new LinkedBlockingQueue<LireFeature>(1000);
for (int i = 0; i < searchers.length; i++) {
searchers[i] = new Thread(new Runnable() {
@Override
public void run() {
SingleNddCeddImageSearcher is = new SingleNddCeddImageSearcher(ir);
LireFeature remove;
while ((remove = queryQueue.remove()) instanceof CEDD) {
try {
is.findMostSimilar(remove);
} catch (IOException e) {
e.printStackTrace();
}
}
}
});
searchers[i].start();
}
long time = System.currentTimeMillis();
for (Iterator<LireFeature> iterator = q.iterator(); iterator.hasNext() && count < 1000; ) {
LireFeature next = iterator.next();
try {
queryQueue.put(next);
count++;
} catch (InterruptedException e) {
e.printStackTrace();
}
}
for (int i = 0; i < 8; i++) {
queryQueue.put(new ScalableColor());
}
for (int i = 0; i < searchers.length; i++) {
searchers[i].join();
}
long l = System.currentTimeMillis() - time;
System.out.printf("Tested %d search requests on %d documents: overall time of %d:%02d, %.2f ms per search", count, ir.maxDoc(), l / (1000 * 60), (l / 1000) % 60, ((float) l / (float) count));
}
public void testCachingSearcherParallelWithBundling() throws IOException, InterruptedException {
final IndexReader ir = DirectoryReader.open(FSDirectory.open(new File("C:\\Temp\\test-100k-cedd-idx")));
LinkedList<LireFeature> q = new LinkedList<LireFeature>();
for (int i = 0; i < ir.maxDoc(); i++) {
Document d = ir.document(i);
CEDD cedd = new CEDD();
BytesRef binaryValue = d.getBinaryValue(cedd.getFieldName());
cedd.setByteArrayRepresentation(binaryValue.bytes, binaryValue.offset, binaryValue.length);
q.add(cedd);
}
int count = 0;
Thread[] searchers = new Thread[4];
final LinkedBlockingQueue<WorkItem> queryQueue = new LinkedBlockingQueue<WorkItem>(100);
for (int i = 0; i < searchers.length; i++) {
searchers[i] = new Thread(new Runnable() {
@Override
public void run() {
SingleNddCeddImageSearcher is = new SingleNddCeddImageSearcher(ir);
WorkItem remove;
while ((remove = queryQueue.remove()).features != null) {
try {
SimpleResult[] hits = is.findMostSimilar(remove.features);
for (int j = 0; j < hits.length; j++) {
if (hits[j].getIndexNumber() != remove.id[j]) System.err.println("oops");
}
} catch (IOException e) {
e.printStackTrace();
}
}
}
});
searchers[i].start();
}
long time = System.currentTimeMillis();
LireFeature[] qarr = new LireFeature[10];
int[] iarr = new int[10];
int currentIndex = 0;
int bundleCount = 0;
Iterator<LireFeature> iterator = q.iterator();
while (iterator.hasNext() && bundleCount < 200) {
LireFeature next = iterator.next();
try {
iarr[currentIndex] = count;
qarr[currentIndex++] = next;
if (currentIndex >= qarr.length) { // do bundled search
currentIndex = 0;
queryQueue.put(new WorkItem(qarr.clone(), iarr.clone()));
bundleCount++;
}
count++;
} catch (InterruptedException e) {
e.printStackTrace();
}
}
for (int i = 0; i < 8; i++) {
queryQueue.put(new WorkItem(null, null));
}
for (int i = 0; i < searchers.length; i++) {
searchers[i].join();
}
long l = System.currentTimeMillis() - time;
System.out.printf("Tested %d search requests on %d documents: overall time of %d:%02d, %.2f ms per search", count, ir.maxDoc(), l / (1000 * 60), (l / 1000) % 60, ((float) l / (float) count));
}
public void testCachingSearcherBundling() throws IOException {
IndexReader ir = DirectoryReader.open(FSDirectory.open(new File("C:\\Temp\\test-100k-cedd-idx")));
SingleNddCeddImageSearcher is = new SingleNddCeddImageSearcher(ir);
LinkedList<LireFeature> q = new LinkedList<LireFeature>();
for (int i = 0; i < ir.maxDoc(); i++) {
Document d = ir.document(i);
CEDD cedd = new CEDD();
BytesRef binaryValue = d.getBinaryValue(cedd.getFieldName());
cedd.setByteArrayRepresentation(binaryValue.bytes, binaryValue.offset, binaryValue.length);
q.add(cedd);
}
long time = System.currentTimeMillis();
int count = 0;
LireFeature[] qarr = new LireFeature[10];
int currentIndex = 0;
for (Iterator<LireFeature> iterator = q.iterator(); iterator.hasNext(); ) {
LireFeature next = iterator.next();
qarr[currentIndex++] = next;
if (currentIndex >= qarr.length) { // do bundled search
currentIndex = 0;
is.findMostSimilar(qarr);
}
count++;
if (count > 999 & currentIndex == 0) break;
}
long l = System.currentTimeMillis() - time;
System.out.printf("Tested %d search requests on %d documents: overall time of %d:%02d, %.2f ms per search", count, ir.maxDoc(), l / (1000 * 60), (l / 1000) % 60, ((float) l / (float) count));
}
}
class WorkItem {
LireFeature[] features;
int[] id;
public WorkItem(LireFeature[] features, int[] id) {
this.features = features;
this.id = id;
}
}