• 如果您觉得本站非常有看点，那么赶紧使用Ctrl+D 收藏吧
> java > Java ImageVectorization类的典型用法和代码示例

Java ImageVectorization类的典型用法和代码示例

java 2次浏览

本文整理汇总了Java中gr.iti.mklab.visual.vectorization.ImageVectorization类的典型用法代码示例。如果您正苦于以下问题：Java ImageVectorization类的具体用法？Java ImageVectorization怎么用？Java ImageVectorization使用的例子？那么恭喜您, 这里精选的类代码示例或许可以为您提供帮助。

ImageVectorization类属于gr.iti.mklab.visual.vectorization包，在下文中一共展示了ImageVectorization类的17个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于我们的系统推荐出更棒的Java代码示例。

示例1: VisualIndexerBolt

​ 点赞 3 ​ 

import gr.iti.mklab.visual.vectorization.ImageVectorization; //导入依赖的package包/类
public VisualIndexerBolt(String webServiceHost, String indexCollection, String[] codebookFiles, String pcaFile) throws Exception {
	
	_webServiceHost = webServiceHost;
	_indexCollection = indexCollection;
	
	ImageVectorization.setFeatureExtractor(new SURFExtractor());
	
	VladAggregatorMultipleVocabularies vladAggregator = new VladAggregatorMultipleVocabularies(codebookFiles, numCentroids, 
			AbstractFeatureExtractor.SURFLength);
	
	ImageVectorization.setVladAggregator(vladAggregator);
	
	int initialLength = numCentroids.length * numCentroids[0] * AbstractFeatureExtractor.SURFLength;
	if(initialLength > targetLengthMax) {
		PCA pca = new PCA(targetLengthMax, 1, initialLength, true);
		pca.loadPCAFromFile(pcaFile);
		ImageVectorization.setPcaProjector(pca);
	}
	
}

开发者ID:socialsensor，
项目名称:storm-focused-crawler，
代码行数:21，
代码来源:VisualIndexerBolt.java

示例2: FeatureExtractorBolt

​ 点赞 3 ​ 

import gr.iti.mklab.visual.vectorization.ImageVectorization; //导入依赖的package包/类
public FeatureExtractorBolt(String webServiceHost, String indexCollection, String[] codebookFiles, String pcaFile) throws Exception {

		ImageVectorization.setFeatureExtractor(new SURFExtractor());
		
		VladAggregatorMultipleVocabularies vladAggregator = new VladAggregatorMultipleVocabularies(codebookFiles, numCentroids, 
				AbstractFeatureExtractor.SURFLength);
		
		ImageVectorization.setVladAggregator(vladAggregator);
		
		int initialLength = numCentroids.length * numCentroids[0] * AbstractFeatureExtractor.SURFLength;
		if(initialLength > targetLengthMax) {
			PCA pca = new PCA(targetLengthMax, 1, initialLength, true);
			pca.loadPCAFromFile(pcaFile);
			ImageVectorization.setPcaProjector(pca);
		}
	}

开发者ID:socialsensor，
项目名称:storm-focused-crawler，
代码行数:17，
代码来源:FeatureExtractorBolt.java

示例3: VisualIndexer

​ 点赞 3 ​ 

import gr.iti.mklab.visual.vectorization.ImageVectorization; //导入依赖的package包/类
public VisualIndexer(String webServiceHost, String indexCollection, String[] codebookFiles, String pcaFile) throws Exception {
	
	this.webServiceHost = webServiceHost;
	this.indexCollection = indexCollection;
	

	
	ImageVectorization.setFeatureExtractor(new SURFExtractor());
	VladAggregatorMultipleVocabularies vladAggregator = new VladAggregatorMultipleVocabularies(codebookFiles, numCentroids, 
			AbstractFeatureExtractor.SURFLength);
	
	ImageVectorization.setVladAggregator(vladAggregator);
	int initialLength = numCentroids.length * numCentroids[0] * AbstractFeatureExtractor.SURFLength;
	PCA pca = new PCA(targetLengthMax, 1, initialLength, true);
	pca.loadPCAFromFile(pcaFile);
	ImageVectorization.setPcaProjector(pca);
}

开发者ID:socialsensor，
项目名称:socialsensor-multimedia-analysis，
代码行数:18，
代码来源:VisualIndexer.java

示例4: MediaItemsVisualIndexer

​ 点赞 3 ​ 

import gr.iti.mklab.visual.vectorization.ImageVectorization; //导入依赖的package包/类
MediaItemsVisualIndexer(String host, String dbname, String collectionName, 
		String webServiceHost, String indexCollection, String[] codebookFiles, String pcaFile) 
		throws Exception {
	
	this.mongo = new MongoClient(host);
	this.db = mongo.getDB(dbname);
	this.input_collection = db.getCollection(collectionName);
	
	input_collection.addOption(Bytes.QUERYOPTION_NOTIMEOUT);
	
	ImageVectorization.setFeatureExtractor(new SURFExtractor());
	VladAggregatorMultipleVocabularies vladAggregator = new VladAggregatorMultipleVocabularies(codebookFiles, numCentroids, 
			AbstractFeatureExtractor.SURFLength);
	
	ImageVectorization.setVladAggregator(vladAggregator);
	int initialLength = numCentroids.length * numCentroids[0] * AbstractFeatureExtractor.SURFLength;
	PCA pca = new PCA(targetLengthMax, 1, initialLength, true);
	pca.loadPCAFromFile(pcaFile);
	ImageVectorization.setPcaProjector(pca);
	
	this.visualIndexHandler = new VisualIndexHandler(webServiceHost, indexCollection);
}

开发者ID:socialsensor，
项目名称:socialsensor-multimedia-analysis，
代码行数:23，
代码来源:MediaItemsVisualIndexer.java

示例5: VisualIndexerBolt

​ 点赞 3 ​ 

import gr.iti.mklab.visual.vectorization.ImageVectorization; //导入依赖的package包/类
public VisualIndexerBolt(String webServiceHost, String indexCollection, String[] codebookFiles, String pcaFile) throws Exception {
	
	this.webServiceHost = webServiceHost;
	this.indexCollection = indexCollection;
	
	ImageVectorization.setFeatureExtractor(new SURFExtractor());
	
	VladAggregatorMultipleVocabularies vladAggregator = new VladAggregatorMultipleVocabularies(codebookFiles, numCentroids, 
			AbstractFeatureExtractor.SURFLength);
	
	ImageVectorization.setVladAggregator(vladAggregator);
	
	int initialLength = numCentroids.length * numCentroids[0] * AbstractFeatureExtractor.SURFLength;
	if(initialLength > targetLengthMax) {
		PCA pca = new PCA(targetLengthMax, 1, initialLength, true);
		pca.loadPCAFromFile(pcaFile);
		ImageVectorization.setPcaProjector(pca);
	}
}

开发者ID:socialsensor，
项目名称:socialsensor-multimedia-analysis，
代码行数:20，
代码来源:VisualIndexerBolt.java

示例6: VisualIndex

​ 点赞 2 ​ 

import gr.iti.mklab.visual.vectorization.ImageVectorization; //导入依赖的package包/类
public VisualIndex(String learningFolder, String indexFolder) throws Exception {
					
	int maximumNumVectors = 20000;

	int m = 128, k_c = 256, numCoarseCentroids = 8192;

	String[] codebookFiles = { 
		learningFolder + "surf_l2_128c_0.csv",
		learningFolder + "surf_l2_128c_1.csv", 
		learningFolder + "surf_l2_128c_2.csv",
		learningFolder + "surf_l2_128c_3.csv" 
	};
					
	String pcaFile = learningFolder + "pca_surf_4x128_32768to1024.txt";

	String coarseQuantizerFile = learningFolder + "qcoarse_1024d_8192k.csv";
	String productQuantizerFile = learningFolder + "pq_1024_128x8_rp_ivf_8192k.csv";

	ImageVectorization.setFeatureExtractor(new SURFExtractor());
	ImageVectorization.setVladAggregator(new VladAggregatorMultipleVocabularies(codebookFiles, numCentroids, AbstractFeatureExtractor.SURFLength));

	if (targetLengthMax < initialLength) {
		PCA pca = new PCA(targetLengthMax, 1, initialLength, true);
		pca.loadPCAFromFile(pcaFile);
		ImageVectorization.setPcaProjector(pca);
	}

	String linearIndexFolder = indexFolder + "/linear"; 
	String ivfpqIndexFolder = indexFolder + "/ivfpq"; 
	
	this.linearIndex = new Linear(targetLengthMax, maximumNumVectors, false, linearIndexFolder, true, true, 0);
	
	this.ivfpqIndex = new IVFPQ(targetLength, maximumNumVectors, false, ivfpqIndexFolder, m, k_c, PQ.TransformationType.RandomPermutation, numCoarseCentroids, true, 0);
	ivfpqIndex.loadCoarseQuantizer(coarseQuantizerFile);
	ivfpqIndex.loadProductQuantizer(productQuantizerFile);
	ivfpqIndex.setW(128); // how many (out of 8192) lists should be visited during search.
}

开发者ID:MKLab-ITI，
项目名称:mgraph-summarization，
代码行数:38，
代码来源:VisualIndex.java

示例7: index

​ 点赞 2 ​ 

import gr.iti.mklab.visual.vectorization.ImageVectorization; //导入依赖的package包/类
public void index(List<Item> items) throws IOException {
	for(Item item : items) {
		
		Map<String, String> mediaItems = item.getMediaItems();

		for(Entry<String, String> mi : mediaItems.entrySet()) {
			
			String mediaId = mi.getKey();
			String mediaUrl = mi.getValue();
		
			try {
				URL url = new URL(mediaUrl);
				byte[] content = IOUtils.toByteArray(url);
				
				BufferedImage img = ImageIO.read(new ByteArrayInputStream(content));
				
				ImageVectorization imvec = new ImageVectorization(mediaId, img , targetLengthMax, maxNumPixels);
				
				ImageVectorizationResult imvr = imvec.call();
				double[] vector = imvr.getImageVector();
				
				// the full vector is indexed in the disk-based index
				linearIndex.indexVector(mediaId, vector);
				
				// the vector is truncated to the correct dimension and renormalized before sending to the ram-based index
				double[] newVector = Arrays.copyOf(vector, targetLength);
				if (newVector.length < vector.length) {
					Normalization.normalizeL2(newVector);
				}
				ivfpqIndex.indexVector(mediaId, newVector);
				
			} catch (Exception e) {
				e.printStackTrace();
			}
		}
	}
}

开发者ID:MKLab-ITI，
项目名称:mgraph-summarization，
代码行数:38，
代码来源:VisualIndex.java

示例8: IndexingManager

​ 点赞 2 ​ 

import gr.iti.mklab.visual.vectorization.ImageVectorization; //导入依赖的package包/类
private IndexingManager() {
    try {
        int[] numCentroids = {128, 128, 128, 128};
        int initialLength = numCentroids.length * numCentroids[0] * AbstractFeatureExtractor.SURFLength;

        String[] codebookFiles = {
                learningFolder + "surf_l2_128c_0.csv",
                learningFolder + "surf_l2_128c_1.csv",
                learningFolder + "surf_l2_128c_2.csv",
                learningFolder + "surf_l2_128c_3.csv"
        };

        String pcaFile = learningFolder + "pca_surf_4x128_32768to1024.txt";


        //visualIndex = new Linear(targetLengthMax, 10000000, false, BDBEnvHome, false, false, 0);
        //int existingVectors = visualIndex.getLoadCounter();
        SURFExtractor extractor = new SURFExtractor();
        ImageVectorization.setFeatureExtractor(extractor);
        double[][][] codebooks = AbstractFeatureAggregator.readQuantizers(codebookFiles, numCentroids,
                AbstractFeatureExtractor.SURFLength);
        ImageVectorization.setVladAggregator(new VladAggregatorMultipleVocabularies(codebooks));
        if (targetLengthMax < initialLength) {
            System.out.println("targetLengthMax : " + targetLengthMax + " initialLengh " + initialLength);
            pca = new PCA(targetLengthMax, 1, initialLength, true);
            pca.loadPCAFromFile(pcaFile);
            ImageVectorization.setPcaProjector(pca);
        }
    } catch (Exception ex) {
        //TODO: do something
    }
}

开发者ID:kandreadou，
项目名称:reveal-web-service，
代码行数:33，
代码来源:IndexingManager.java

示例9: indexImage

​ 点赞 2 ​ 

import gr.iti.mklab.visual.vectorization.ImageVectorization; //导入依赖的package包/类
public boolean indexImage(String imageFolder, String imageFilename, String collection) throws Exception {

        if (Strings.isNullOrEmpty(collection)) {
            collection = DEFAULT_COLLECTION_NAME;
        }
        AbstractSearchStructure index = indices.get(collection);
        if (index == null) {
            createIndex(collection);
            index = indices.get(collection);
        }
        ImageVectorization imvec = new ImageVectorization(imageFolder, imageFilename, targetLengthMax, maxNumPixels);
        ImageVectorizationResult imvr = imvec.call();
        double[] vector = imvr.getImageVector();
        return index.indexVector(imageFilename, vector);
    }

开发者ID:kandreadou，
项目名称:reveal-web-service，
代码行数:16，
代码来源:IndexingManager.java

示例10: findSimilar

​ 点赞 2 ​ 

import gr.iti.mklab.visual.vectorization.ImageVectorization; //导入依赖的package包/类
public Answer findSimilar(String url, String collection, int neighbours) throws Exception {
    if (Strings.isNullOrEmpty(collection)) {
        collection = DEFAULT_COLLECTION_NAME;
    }
    AbstractSearchStructure index = indices.get(collection);
    if (index == null) {
        createIndex(collection);
        index = indices.get(collection);
    }
    BufferedImage img = ImageIO.read(new URL(url));
    ImageVectorization imvec = new ImageVectorization(url, img, targetLengthMax, maxNumPixels);
    ImageVectorizationResult imvr = imvec.call();
    double[] vector = imvr.getImageVector();
    return index.computeNearestNeighbors(neighbours, vector);
}

开发者ID:kandreadou，
项目名称:reveal-web-service，
代码行数:16，
代码来源:IndexingManager.java

示例11: initialize

​ 点赞 2 ​ 

import gr.iti.mklab.visual.vectorization.ImageVectorization; //导入依赖的package包/类
private void initialize(Context context) throws Exception {

        int[] numCentroids = {128, 128, 128, 128};
        int initialLength = numCentroids.length * numCentroids[0] * AbstractFeatureExtractor.SURFLength;

        /*if (!VisualJob.IS_LOCAL) {

            Path files[] = DistributedCache.getLocalCacheFiles(context.getConfiguration());
            for (Path path : files) {
                context.write(new Text(path.toString() + " " + path.getName()), new DoubleArrayWritable(new double[0]));
            }
        }*/

        String[] codebookFiles = {
                VisualJob.LEARNING_FILES_PATH + "surf_l2_128c_0.csv",
                VisualJob.LEARNING_FILES_PATH + "surf_l2_128c_1.csv",
                VisualJob.LEARNING_FILES_PATH + "surf_l2_128c_2.csv",
                VisualJob.LEARNING_FILES_PATH + "surf_l2_128c_3.csv"
        };
        String pcaFile = VisualJob.LEARNING_FILES_PATH + "pca_surf_4x128_32768to1024.txt";

        ImageVectorization.setFeatureExtractor(new SURFExtractor());
        double[][][] codebooks = AbstractFeatureAggregator.readQuantizers(codebookFiles, numCentroids,
                AbstractFeatureExtractor.SURFLength);
        ImageVectorization.setVladAggregator(new VladAggregatorMultipleVocabularies(codebooks));
        if (targetLengthMax < initialLength) {
            PCA pca = new PCA(targetLengthMax, 1, initialLength, true);
            pca.loadPCAFromFile(pcaFile);
            ImageVectorization.setPcaProjector(pca);
        }

        // Initialize the downloader, the vectorizer and the indexer
        downloader = new ImageDownloader("", numDownloadThreads);
        downloader.setSaveOriginal(false);
        downloader.setSaveThumb(false);
        downloader.setFollowRedirects(false);
        vectorizer = new ImageVectorizer("surf", codebookFiles, numCentroids,
                targetLengthMax, pcaFile, true, numVectorizationThreads);

    }

开发者ID:MKLab-ITI，
项目名称:multimedia-indexing，
代码行数:41，
代码来源:VisualThreadedMapper.java

示例12: index

​ 点赞 2 ​ 

import gr.iti.mklab.visual.vectorization.ImageVectorization; //导入依赖的package包/类
public static void index(String directory, String linearIndexFolder, String ivfpqIndexFolder, String learningFolder) throws Exception {
	
	File dir = new File(directory);
	if(!dir.isDirectory())
		return;
	
	if(!learningFolder.endsWith("/"))
		learningFolder += "/";
	
	// parameters
	int maxNumPixels = 768 * 512;
	int[] numCentroids = { 128, 128, 128, 128 };
			
	int initialLength = numCentroids.length * numCentroids[0] * AbstractFeatureExtractor.SURFLength;
			
	int targetLengthMax = 1024;
	int targetLength = 1024;
			
	int maximumNumVectors = 20000;

	int m = 128, k_c = 256, numCoarseCentroids = 8192;

	String[] codebookFiles = { 
		learningFolder + "surf_l2_128c_0.csv",
		learningFolder + "surf_l2_128c_1.csv", 
		learningFolder + "surf_l2_128c_2.csv",
		learningFolder + "surf_l2_128c_3.csv" 
	};
			
	String pcaFile = learningFolder + "pca_surf_4x128_32768to1024.txt";

	String coarseQuantizerFile = learningFolder + "qcoarse_1024d_8192k.csv";
	String productQuantizerFile = learningFolder + "pq_1024_128x8_rp_ivf_8192k.csv";

	ImageVectorization.setFeatureExtractor(new SURFExtractor());
	ImageVectorization.setVladAggregator(new VladAggregatorMultipleVocabularies(codebookFiles,
					numCentroids, AbstractFeatureExtractor.SURFLength));

	if (targetLengthMax < initialLength) {
		PCA pca = new PCA(targetLengthMax, 1, initialLength, true);
		pca.loadPCAFromFile(pcaFile);
		ImageVectorization.setPcaProjector(pca);
	}

	Linear linear = new Linear(targetLengthMax, maximumNumVectors, false, linearIndexFolder, true, true, 0);
	IVFPQ ivfpq_1 = new IVFPQ(targetLength, maximumNumVectors, false, ivfpqIndexFolder, m, k_c, PQ.TransformationType.RandomPermutation, numCoarseCentroids, true, 0);
	ivfpq_1.loadCoarseQuantizer(coarseQuantizerFile);
	ivfpq_1.loadProductQuantizer(productQuantizerFile);
	ivfpq_1.setW(128); // how many (out of 8192) lists should be visited during search.
	
	for(String id : dir.list()) {
		System.out.println(id);
		
		ImageVectorization imvec = new ImageVectorization(dir.toString()+"/", id, targetLengthMax, maxNumPixels);
		
		ImageVectorizationResult imvr = imvec.call();
		double[] vector = imvr.getImageVector();
		
		// the full vector is indexed in the disk-based index
		linear.indexVector(id, vector);
		
		// the vector is truncated to the correct dimension and renormalized before sending to the ram-based index
		double[] newVector = Arrays.copyOf(vector, targetLength);
		if (newVector.length < vector.length) {
			Normalization.normalizeL2(newVector);
		}
		ivfpq_1.indexVector(id, newVector);
		
	}
	
}

开发者ID:MKLab-ITI，
项目名称:mgraph-summarization，
代码行数:72，
代码来源:MultimediaHandler.java

示例13: execute

​ 点赞 2 ​ 

import gr.iti.mklab.visual.vectorization.ImageVectorization; //导入依赖的package包/类
public void execute(Tuple tuple) {
	MediaItem mediaItem = (MediaItem) tuple.getValueByField("MediaItem");
	if(mediaItem == null)
		return;
	
	HttpGet httpget = null;
	ImageVector imageVector = null;
	try {
		String id = mediaItem.getId();
		String type = mediaItem.getType();
		
		String url = type.equals("image") ? mediaItem.getUrl() : mediaItem.getThumbnail();
		
		httpget = new HttpGet(url.replaceAll(" ", "%20"));
		httpget.setConfig(_requestConfig);
		HttpResponse response = _httpclient.execute(httpget);
		
		StatusLine status = response.getStatusLine();
		int code = status.getStatusCode();
		
		if(code<200 || code>=300) {
			_logger.error("Failed fetch media item " + id + ". URL=" + url +  
					". Http code: " + code + " Error: " + status.getReasonPhrase());
			
			mediaItem.setVisualIndexed(false);
			_collector.emit(tuple(mediaItem, imageVector));
			
			return;
		}
		
		HttpEntity entity = response.getEntity();
		if(entity == null) {
			_logger.error("Entity is null for " + id + ". URL=" + url +  
					". Http code: " + code + " Error: " + status.getReasonPhrase());
			
			mediaItem.setVisualIndexed(false);
			_collector.emit(tuple(mediaItem, imageVector));
			
			return;
		}
		
		InputStream input = entity.getContent();
		byte[] imageContent = IOUtils.toByteArray(input);
		
		BufferedImage image = ImageIO.read(new ByteArrayInputStream(imageContent));
	
		if(image != null) {
			ImageVectorization imvec = new ImageVectorization(id, image, targetLengthMax, maxNumPixels);
			
			if(mediaItem.getWidth()==null && mediaItem.getHeight()==null) {
				mediaItem.setSize(image.getWidth(), image.getHeight());
			}

			ImageVectorizationResult imvr = imvec.call();
			double[] vector = imvr.getImageVector();
			
			if(vector==null || vector.length==0) {
				_logger.error("Error in feature extraction for " + id);
			}
			
			imageVector = new ImageVector(id, url, vector);				
		}
		_collector.emit(tuple(mediaItem, imageVector));
		
	} 
	catch (Exception e) {
		_logger.error(e);
		_collector.emit(tuple(mediaItem, imageVector));
	}
	finally {
		if(httpget != null)
			httpget.abort();
	}
}

开发者ID:socialsensor，
项目名称:storm-focused-crawler，
代码行数:75，
代码来源:FeatureExtractorBolt.java

示例14: execute

​ 点赞 2 ​ 

import gr.iti.mklab.visual.vectorization.ImageVectorization; //导入依赖的package包/类
public void execute(Tuple tuple) {
	
	String id = tuple.getStringByField("id");
	String url = tuple.getStringByField("url");
	Double score = tuple.getDoubleByField("score");
	
	boolean size = tuple.getBooleanByField("size");
	
	System.out.println("Fetch and extract feature vector for " + id + " with score " + score);
	try {
		
		
		BufferedImage image = ImageIO.read(new URL(url));
		
		Integer width=-1, height=-1;
		boolean indexed = false;
		if(image != null) {
			
			ImageVectorization imvec = new ImageVectorization(id, image, targetLengthMax, maxNumPixels);
			
			if(!size) {
				width = image.getWidth();
				height = image.getHeight();
			}

			ImageVectorizationResult imvr = imvec.call();
			double[] vector = imvr.getImageVector();

		
			indexed = visualIndex.index(id, vector);

		}
		
		if(indexed) {
			_collector.emit(tuple(id, Boolean.TRUE, width, height));
		}
		else {
			_collector.emit(tuple(id, Boolean.FALSE, width, height));
		}
	} 
	catch (Exception e) {
		System.out.println("Exception: " + e.getMessage());
		_collector.emit(tuple(id, Boolean.FALSE, -1, -1));
		return;
	}
}

开发者ID:socialsensor，
项目名称:socialsensor-multimedia-analysis，
代码行数:47，
代码来源:VisualIndexer.java

示例15: run

​ 点赞 2 ​ 

import gr.iti.mklab.visual.vectorization.ImageVectorization; //导入依赖的package包/类
@Override
public void run() {
	while(true) {
		DBObject query = new BasicDBObject();
		query.put("indexed", false);
		query.put("type", "image");
		
	
		DBCursor items = input_collection.find(query);
		if(items.size()==0)
			break;
		
		int k=0;
		for(DBObject item : items) {
			String id = (String) item.get("id");
			String url = (String) item.get("url");
			
			
			try {
				InputStream is = new URL(url).openStream();
				byte[] content = IOUtils.toByteArray(is);
				
				if(++k%100==0) {
					System.out.println(k + " media items indexed.");
					break;
				}
				
				BufferedImage image = ImageIO.read(new ByteArrayInputStream(content));
				
				ImageVectorization imvec = new ImageVectorization(id, image, targetLengthMax, maxNumPixels);
				ImageVectorizationResult imvr = imvec.call();
				double[] vector = imvr.getImageVector();
				
				if(visualIndexHandler.index(id, vector)) {
					DBObject q = new BasicDBObject("id", id);
					DBObject o = new BasicDBObject("$set", 
							new BasicDBObject("indexed",true));		
					input_collection.update(q , o , false, true);
				}
			} catch (Exception e) {
				continue;
			}
			
			
		}
	}
	System.out.println("DONE: All");
}

开发者ID:socialsensor，
项目名称:socialsensor-multimedia-analysis，
代码行数:49，
代码来源:MediaItemsVisualIndexer.java

示例16: execute

​ 点赞 2 ​ 

import gr.iti.mklab.visual.vectorization.ImageVectorization; //导入依赖的package包/类
public void execute(Tuple tuple) {
	
	String id = tuple.getStringByField("id");
	String url = tuple.getStringByField("url");
	//Double score = tuple.getDoubleByField("score");
	
	boolean size = tuple.getBooleanByField("size");
	
	//System.out.println("Fetch and extract feature vector for " + id + " with score " + score);
	try {
		
		byte[] imageContent = IOUtils.toByteArray(new URL(url));
		BufferedImage image = ImageIO.read(new ByteArrayInputStream(imageContent));
		
		
		Integer width=-1, height=-1;
		boolean indexed = false;
		
		if(image != null) {
			
			ImageVectorization imvec = new ImageVectorization(id, image, targetLengthMax, maxNumPixels);
			
			if(!size) {
				width = image.getWidth();
				height = image.getHeight();
			}

			ImageVectorizationResult imvr = imvec.call();
			double[] vector = imvr.getImageVector();

			
			indexed = visualIndex.index(id, vector);

		}
		
		if(indexed) {
			_collector.emit(tuple(id, Boolean.TRUE, width, height));
		}
		else {
			_collector.emit(tuple(id, Boolean.FALSE, width, height));
		}
	} 
	catch (Exception e) {
		System.out.println("Exception: " + e.getMessage() + " url: "+url);
		_collector.emit(tuple(id, Boolean.FALSE, -1, -1));
		return;
	}
}

开发者ID:socialsensor，
项目名称:socialsensor-multimedia-analysis，
代码行数:49，
代码来源:VisualIndexerBolt.java

示例17: cluster

​ 点赞 2 ​ 

import gr.iti.mklab.visual.vectorization.ImageVectorization; //导入依赖的package包/类
public static void cluster() throws Exception {
	
	String learningFiles = "/disk2_data/VisualIndex/learning_files/";
	
	String[] codebookFiles = { 
			learningFiles + "surf_l2_128c_0.csv",
			learningFiles + "surf_l2_128c_1.csv", 
			learningFiles + "surf_l2_128c_2.csv",
			learningFiles + "surf_l2_128c_3.csv" };
	
	String pcaFile = learningFiles + "pca_surf_4x128_32768to1024.txt";

	
	int[] numCentroids = { 128, 128, 128, 128 };
	int targetLengthMax = 1024;
	
	int maxNumPixels = 768 * 512; // use 1024*768 for better/slower extraction
	
	ImageVectorization.setFeatureExtractor(new SURFExtractor());
	
	VladAggregatorMultipleVocabularies vladAggregator = new VladAggregatorMultipleVocabularies(codebookFiles, numCentroids, 
			AbstractFeatureExtractor.SURFLength);
	
	ImageVectorization.setVladAggregator(vladAggregator);
	
	int initialLength = numCentroids.length * numCentroids[0] * AbstractFeatureExtractor.SURFLength;
	if(initialLength > targetLengthMax) {
		PCA pca = new PCA(targetLengthMax, 1, initialLength, true);
		pca.loadPCAFromFile(pcaFile);
		ImageVectorization.setPcaProjector(pca);
	}
	
	String mongoHost = "160.40.51.18";
	String mongoDb = "FeteBerlin";
	String mongoCollection = "MediaItems";
	
	MongoClient client = new MongoClient(mongoHost);
	DB db = client.getDB(mongoDb);
	DBCollection coll = db.getCollection(mongoCollection);
	DBCollection clustersCollection = db.getCollection("MediaItemClusters3");
			
	VisualIndexHandler vIndex = new VisualIndexHandler("http://160.40.51.18:8080/VisualIndexService", "fete3");
	
	DBCursor cursor = coll.find(new BasicDBObject("type","image"));
	
	ArrayBlockingQueue<Vec> queue1 = new ArrayBlockingQueue<Vec>(5000);
	ArrayBlockingQueue<Vec> queue2 = new ArrayBlockingQueue<Vec>(5000);
	
	Thread clusterer = new Thread(new Clusterer(queue2, vIndex, coll, clustersCollection));
	List<Thread> extractors = new ArrayList<Thread>();
	for(int i=0;i<48;i++) {
		extractors.add(new Thread(new Extractor(queue1, queue2)));
	}
	
	clusterer.start();
	for(Thread t : extractors) {
		t.start();
	}
	while(cursor.hasNext()) {
		
		DBObject mItem = cursor.next();
		String id = (String) mItem.get("id");
		String url = (String) mItem.get("url");
			
		try {
			Vec vec = new Vec(id, url, null);
			queue1.put(vec);
			Thread.sleep(5);
		}
		catch(Exception e) {
			e.printStackTrace();
		}
	}
	System.out.println("Wait for clusterer to finish!!!");
	clusterer.join();
}

开发者ID:socialsensor，
项目名称:socialsensor-multimedia-analysis，
代码行数:77，
代码来源:VisualIndexerBolt.java

版权声明：本文转自网络文章，转载此文章仅为分享知识，如有侵权，请联系管理员进行删除。
喜欢 (0)