<?xml version="1.0" encoding="UTF-8"?><!-- generator="wordpress.com" -->
<rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:wfw="http://wellformedweb.org/CommentAPI/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	>

<channel>
	<title>visualrank &amp;laquo; WordPress.com Tag Feed</title>
	<link>http://en.wordpress.com/tag/visualrank/</link>
	<description>Feed of posts on WordPress.com tagged "visualrank"</description>
	<pubDate>Sun, 29 Nov 2009 02:47:10 +0000</pubDate>

	<generator>http://en.wordpress.com/tags/</generator>
	<language>en</language>

<item>
<title><![CDATA[Google VisualRank : recherche optimisée des images pour Google]]></title>
<link>http://guilmain.wordpress.com/2008/05/07/googl-visualrank-recherche-optimisee-des-images-pour-google/</link>
<pubDate>Wed, 07 May 2008 13:20:00 +0000</pubDate>
<dc:creator>Olivier Guilmain</dc:creator>
<guid>http://guilmain.wordpress.com/2008/05/07/googl-visualrank-recherche-optimisee-des-images-pour-google/</guid>
<description><![CDATA[Le moteur a annoncé la création d&#8217;une nouvelle technologie de recherche d&#8217;images. L]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><div style="text-align:justify;">Le moteur a annoncé la création d&#8217;une nouvelle technologie de recherche d&#8217;images. L&#8217;acteur a adapté son algorithme du PageRank aux besoins des images. Nommée <span style="font-weight:bold;">VisualRank</span>, cette fonction va chercher au-delà du nom d&#8217;une image. Elle se base notamment sur la reconnaissance du contenu et sur la pertinence. Pour l&#8217;instant encore en phase de test, le VisualRank n&#8217;est déployé que sur certaines expressions et marques clés sur son moteur anglophone.</div>
<p>Source : <a href="http://www.journaldunet.com/solutions/breve/26277/visualrank-google-ne-veut-plus-etre-sourd-aux-images.shtml">JDN</a></p>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[Enquanto a computação quântica não vem...]]></title>
<link>http://suzanacohen.wordpress.com/2008/04/30/enquanto-a-computacao-quantica-nao-vem/</link>
<pubDate>Wed, 30 Apr 2008 15:21:13 +0000</pubDate>
<dc:creator>suzanacohen</dc:creator>
<guid>http://suzanacohen.wordpress.com/2008/04/30/enquanto-a-computacao-quantica-nao-vem/</guid>
<description><![CDATA[O Google apresentou na semana passada na International World Wide Web Conference, em Pequim, o protó]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p>O Google apresentou na semana passada na <a href="http://www2008.org/" target="_blank">International World Wide Web Conference</a>, em Pequim, o protótipo para um sistema de busca de imagens de precisão (Precision Image Search), que incluiria um sistema de &#8220;rank visual&#8221; (VisualRank). O sistema seria mais ou menos uma mistura de um algorítmo para reconhecimento de imagens, com técnicas para avaliação de importância (peso) e ranking de imagens que apresentam semelhanças.</p>
<p>Atualmente os resultados das buscas por imagens geralmente se pautam em palavras presentes nos textos que acompanham ou estão associados às mesmas. Em um futuro próximo, portanto, a busca por imagens certamente passará por um filtro qualitativo, descartando muitas imagens não-relacionadas ao termo de busca em questão.</p>
<p>O assunto me lembrou o <a href="http://fisicafacil.wordpress.com/2008/04/25/qubits-computacao-quantica-qubits-quantum-computation/" target="_blank">exemplo</a> dado pelo Aba &#8211; no <a href="http://fisicafacil.wordpress.com" target="_blank">Fisica Fácil</a> &#8211; a uma possível aplicação da computação quântica no reconhecimento de imagens, por exemplo, de uma mesma pessoa hoje e há 30 anos.</p>
<p>Fonte: <a href="http://www.nytimes.com/2008/04/28/technology/28google.html?_r=1&#38;ref=technology&#38;oref=slogin" target="_blank">&#8220;A Google Prototype for a Precision Image Search&#8221;, New York Times, 28/04/2008</a>.</p>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[Google revolutionerar bildsökning med VisualRank]]></title>
<link>http://edenstrom.wordpress.com/2008/04/30/google-revolutionerar-bildsokning-med-visualrank/</link>
<pubDate>Wed, 30 Apr 2008 10:39:22 +0000</pubDate>
<dc:creator>edenstrom</dc:creator>
<guid>http://edenstrom.wordpress.com/2008/04/30/google-revolutionerar-bildsokning-med-visualrank/</guid>
<description><![CDATA[Jag har tidigare nämnt hur det på flera håll i världen jobbas på förbättrad bildsökning. En utveckli]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p>Jag har tidigare nämnt hur det på flera håll i världen jobbas på förbättrad bildsökning. En utveckling som inte minst CMS plattformar kommer gynnas av. <a href="http://www.google.com"></p>
<p>Google</a> och <a href="http://images.google.se">Google Image Search</a> ska nu tagit första stegen med ett VisualRank system som ska fungera som <a href="http://www.google.com">Google</a>s PageRank.<br />
<img src="http://technology.timesonline.co.uk/multimedia/archive/00325/google185_325122a.jpg" alt="" width="185" height="185" /></p>
<p>Nuvarande <a href="http://images.google.se">Google Image Search</a> <a href="http://en.wikipedia.org/wiki/Search_engine_results_page">SERP</a> baseras på alt-taggar och text som omgärdar bilden. I framtiden ska <a href="http://images.google.se">Google Image Search</a> VisualRank baseras på snarlika objekt identifierade i en bildfil. En katt ska kunna kännas igen i vilken bild som helst, baserat på djurets form, färg och sammansättning.</p>
<p>Tester av det nya systemet ska ha gett 83 procent färre irrelevanta sökresultat jämfört med gamla <a href="http://images.google.se">Google Image Search</a>, enligt ett dokument om VisualRank.</p>
<p>Svenska <a href="http://www.polarrose.com">Polar Rose</a> är ett annat av gängen som jobbar på framtidens bildsök.</p>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[VisualRank για μια πιο αποτελεσματική αναζήτηση εικόνων]]></title>
<link>http://altervedo.wordpress.com/2008/04/29/visualrank-%ce%b3%ce%b9%ce%b1-%ce%bc%ce%b9%ce%b1-%cf%80%ce%b9%ce%bf-%ce%b1%cf%80%ce%bf%cf%84%ce%b5%ce%bb%ce%b5%cf%83%ce%bc%ce%b1%cf%84%ce%b9%ce%ba%ce%ae-%ce%b1%ce%bd%ce%b1%ce%b6%ce%ae%cf%84%ce%b7/</link>
<pubDate>Tue, 29 Apr 2008 17:38:15 +0000</pubDate>
<dc:creator>altervedo</dc:creator>
<guid>http://altervedo.wordpress.com/2008/04/29/visualrank-%ce%b3%ce%b9%ce%b1-%ce%bc%ce%b9%ce%b1-%cf%80%ce%b9%ce%bf-%ce%b1%cf%80%ce%bf%cf%84%ce%b5%ce%bb%ce%b5%cf%83%ce%bc%ce%b1%cf%84%ce%b9%ce%ba%ce%ae-%ce%b1%ce%bd%ce%b1%ce%b6%ce%ae%cf%84%ce%b7/</guid>
<description><![CDATA[Όλοι λίγο πολύ γνωρίζουμε το PageRank του Google. Φαίνεται ότι οι άνθρωποι του Google σκοπεύουν να ε]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p><img src="http://img107.imageshack.us/img107/1064/googlevisualrankws9.jpg" alt="" />Όλοι λίγο πολύ γνωρίζουμε το PageRank του Google. Φαίνεται ότι οι άνθρωποι του Google σκοπεύουν να επεκτείνουν την έννοια του PageRank και στις εικόνες για μια αποτελεσματική αναζήτηση εικόνων στο web. Η νέα τεχνολογία που ονομάζεται VisualRank, υπόσχεται παροχή αποτελεσμάτων που θα περιέχουν πολύ λιγότερες μη σχετικές εικόνες με τους όρους αναζήτησής μας, αλλά και εμφάνιση των αποτελεσμάτων με βάση το rank των εικόνων, δηλαδή το πόσο ανταποκρινόνται στους όρους που υποβάλλαμε προς αναζήτηση.</p>
<p>Ως σήμερα, με εξαίρεση ελάχιστων περιπτώσεων, η αναζήτηση εικόνων γίνεται με βάση το κείμενο που τις συνοδεύει και όχι με τα πραγματικά περιεχόμενα της εικόνας, όπως τα αντιλαμβάνονται τα ανθρώπινα μάτια. Η τεχνολογία που απαιτείται για την αναγνώριση και συσχέτιση σχημάτων που αναπαριστόνται σε φωτογραφίες υπάρχει ή τουλάχιστον έχει επιτευχθεί μερικώς, όπως στην αναγνώριση προσώπων, ωστόσο η χρήση της σε projects τεράστιων διαστάσεων όπως το index εικόνων του Google θεωρείται από πολλούς αδύνατη.</p>
<p>Σκοπός της νέας μεθόδου-τεχνολογίας, που προτείνουν οι Yushi Jing και Shumeet Baluja της Google, δεν είναι τόσο η ακριβής αναγνώριση σχετικών εικόνων, όσο η αναγνώριση ενός κοινού visual theme. Πιο συγκεκριμένα, γίνεται η αναγνώριση των πολλαπλών οπτικών θεμάτων και η κατάταξή τους με βάση το βάρος τους ( rank ) σε σχέση με τους όρους αναζήτησης.</p>
<p>Οι άνθρωποι της Google στην εργασία τους, που παρουσίασαν κατά τη διάρκεια του συνεδρίου WWW2008 Beijing 2008, στο Πεκίνο, και που ολοκληρώθηκε την 25η Απριλίου 2008 ( 21-25 Απριλίου 2008 ), ανέφεραν ότι κατάφεραν με τη μέθοδό τους την εξάλειψη έως και του 83% των μη σχετικών εικόνων που προέκυπταν κατά τις αναζητήσεις.</p>
<p>Μπορείτε να διαβάσετε ολόκληρο το κείμενο της εργασίας από το αρχείο <a title="PageRank for Product Image Search" href="http://www.www2008.org/papers/pdf/p307-jingA.pdf" target="_blank">PDF</a>.</p>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[Lets Talk VisualRank]]></title>
<link>http://shaila11.wordpress.com/2008/04/29/lets-talk-visualrank/</link>
<pubDate>Tue, 29 Apr 2008 16:16:03 +0000</pubDate>
<dc:creator>shaila11</dc:creator>
<guid>http://shaila11.wordpress.com/2008/04/29/lets-talk-visualrank/</guid>
<description><![CDATA[&#8220;VisualRank&#8221; is nothing but an algorithm an algorithm for blending image-recognition sof]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p>&#8220;VisualRank&#8221; is nothing but an algorithm an algorithm for blending image-recognition software methods with techniques for weighting and ranking images that look most similar.</p>
<p>VisualRank differs from existent image search &#8211; the most popular tool for searching images. Although image search has become popular, results are usually generated today by using cues from the text associated with each image. On Thursday at the International World Wide Web Conference in Beijing, two Google scientists presented a paper describing VisualRank claiming that now they have a software technology intended to do for digital images on the Web what the company’s original PageRank software did for searches of Web pages.</p>
<p>Historically, Google is not the first into the visual product search category. Riya, a Silicon Valley start-up, introduced Like.com in 2006. The service, which refers users to shopping sites, makes it possible for a Web shopper to select a particular visual attribute, such as a certain style of brown shoes or a buckle, and then be presented with similar products available from other Web merchants.</p>
<p style="text-align:center;"><img class="alignnone size-medium wp-image-63 aligncenter" src="http://shaila11.wordpress.com/files/2008/04/visualrank.jpg?w=300" alt="visualrank" width="326" height="210" /></p>
<p>Riya and like.com had become first true visual search engines, where the content of photos are used to search and reterive similar items. Riya was founded in August 2004 and has assembled one of the largest visual computing research teams in the world and it has raised $19.5 million from venture and private equity investors, including Bay Partners, BlueRun Ventures, and Leapfrog Ventures.</p>
<p><a title="Holiday Inn Flagstaff Hotel" href="http://www.hiflagstaff.com" target="_blank">flagstaff hotel</a> &#124; <a href="http://www.amcheck.com/">Payroll Services</a> &#124; <a href="http://www.calibrus.com">TPV</a> &#124; <a href="http://www.backupsolutions.com">Online Backup</a> &#124; <a href="http://www.hunt-africa.com">African Hunting Safaris</a> &#124; <a href="http://www.bestwesternpaysoninn.com/">Payson Hotel</a> &#124; <a href="http://www.hielax.com/">lax airport hotel</a> &#124; <a href="http://www.fairfieldinnflagstaff.com/">flagstaff accommodation</a></p>
</div>]]></content:encoded>
</item>
<item>
<title><![CDATA[VisualRank: Google's new image search algorithm]]></title>
<link>http://biswaroop.wordpress.com/2008/04/29/visualrank-googles-new-image-search-algorithm/</link>
<pubDate>Tue, 29 Apr 2008 09:39:14 +0000</pubDate>
<dc:creator>biswaroop</dc:creator>
<guid>http://biswaroop.wordpress.com/2008/04/29/visualrank-googles-new-image-search-algorithm/</guid>
<description><![CDATA[A recent paper from Google describes a new image search algorithm that ranks images based on their v]]></description>
<content:encoded><![CDATA[<div class='snap_preview'><p>A recent <a href="http://www.www2008.org/papers/pdf/p307-jingA.pdf">paper</a> from Google describes a new image search algorithm that ranks images based on their visual similarity. This <a href="http://www.nytimes.com/2008/04/28/technology/28google.html?ref=business">NYTimes article</a> and <a href="http://searchengineland.com/080428-052458.php">this</a> give a good introduction to the algorithm. I will try to give an overview of their approach:</p>
<ul>
<li>They extract local descriptors (SIFT descriptors) on the images.</li>
<li>Measure of similarity between two images is defined as the number of interest points (descriptor vectors) shared between the two images divided by their average number of interest points.</li>
<li>The similarity between images is considered as probabilistic <em>visual hyperlinks</em> (this is necessary as there are no actual links between the images) and this leads to using the PageRank algorithm for ranking.</li>
</ul>
<p>The above ranking method can be interpreted as finding multiple visual themes and their strengths in a large set of images and using this for ranking them. An example from the paper is shown below. There are many comic representations of the painting MonaLisa and all of them are based on the original painting. The original painting will contain more matched local features than others (and hence will be rated as having a stronger <em>visual hyperlink</em>). As seen in the image below, the center of the graph contains images corresponding to the original version of the painting.</p>
<p><img class="aligncenter" src="http://farm4.static.flickr.com/3179/2448788038_3fb85f9e5a_o_d.jpg" alt="" /></p>
<p>The authors of the paper above have posted some clarifications about the paper <a href="http://googleresearch.blogspot.com/2008/05/visualrank.html">here</a>.</p>
<p>On a similar not, came across a good talk on Image retrieval, especially semantic image retrieval.</p>
<p><span style='text-align:center;display:block;'><object width='400' height='330' type='application/x-shockwave-flash' data='http://video.google.com/googleplayer.swf?docId=2225647906131550844'><param name='allowScriptAccess' value='never' /><param name='movie' value='http://video.google.com/googleplayer.swf?docId=2225647906131550844'/><param name='quality' value='best'/><param name='bgcolor' value='#ffffff' /><param name='scale' value='noScale' /><param name='wmode' value='window'/></object></span></p>
<p>Using Statistics to Search and Annotate Pictures -&#62; Gives a brief introduction to image retrieval (query by sketch, query by example) followed by the concept of semantic image retrieval.</p>
<p><iframe src='http://digg.com/api/diggthis.php?u=http%3A%2F%2Fdigg.com%2Fsoftware%2FVisualRank_Google_s_new_image_search_algorithm' height='82' width='55' frameborder='0' scrolling='no' style='float: right; margin-left: 10px; margin-bottom: 5px; padding: 4px 0 2px 4px; background: #fff;'></iframe></p>
</div>]]></content:encoded>
</item>

</channel>
</rss>
