<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	>
<channel>
	<title>Comments on: Search for Faces with Google Image Search</title>
	<atom:link href="http://cybernetnews.com/2007/05/29/search-for-faces-with-google-image-search/feed/" rel="self" type="application/rss+xml" />
	<link>http://cybernetnews.com/2007/05/29/search-for-faces-with-google-image-search/</link>
	<description>Technology News</description>
	<pubDate>Sun, 07 Sep 2008 05:21:20 +0000</pubDate>
	<generator>http://wordpress.org/?v=2.6.1</generator>
		<item>
		<title>By: Alexei Yavlinsky</title>
		<link>http://cybernetnews.com/2007/05/29/search-for-faces-with-google-image-search/#comment-96838</link>
		<dc:creator>Alexei Yavlinsky</dc:creator>
		<pubDate>Tue, 29 May 2007 22:02:45 +0000</pubDate>
		<guid isPermaLink="false">http://tech.cybernetnews.com/2007/05/29/search-for-faces-with-google-image-search/#comment-96838</guid>
		<description>Hi, we have now been working for a while on a prototype, proof-of-concept image search engine called Behold that combines visual concept analysis with traditional text search. It does what the article says but in a more extended manner.

http://www.beholdsearch.com

It currently indexes over 1 million images from university websites (obviously much smaller scale than Google). So far it handles 57 visual concepts which can be helpful when extracted html metadata is poor. It can do what this article says, i.e. restrict search results obtained through traditional search by the probability of the image containing a particular visual concept. This feature is available here for you to try:

&lt;a href="http://go.beholdsearch.com/searchc.jsp" rel="nofollow"&gt;http://go.beholdsearch.com/searchc.jsp&lt;/a&gt;

More about this feature on this blog post:

&lt;a href="http://beholdsearch.blogspot.com/2006/11/new-feature.html" rel="nofollow"&gt;http://beholdsearch.blogspot.com/2006/11/new-feature.html&lt;/a&gt;</description>
		<content:encoded><![CDATA[<p>Hi, we have now been working for a while on a prototype, proof-of-concept image search engine called Behold that combines visual concept analysis with traditional text search. It does what the article says but in a more extended manner.</p>
<p><a href="http://www.beholdsearch.com" rel="nofollow">http://www.beholdsearch.com</a></p>
<p>It currently indexes over 1 million images from university websites (obviously much smaller scale than Google). So far it handles 57 visual concepts which can be helpful when extracted html metadata is poor. It can do what this article says, i.e. restrict search results obtained through traditional search by the probability of the image containing a particular visual concept. This feature is available here for you to try:</p>
<p><a href="http://go.beholdsearch.com/searchc.jsp" rel="nofollow">http://go.beholdsearch.com/searchc.jsp</a></p>
<p>More about this feature on this blog post:</p>
<p><a href="http://beholdsearch.blogspot.com/2006/11/new-feature.html" rel="nofollow">http://beholdsearch.blogspot.c.....ature.html</a></p>
]]></content:encoded>
	</item>
	<item>
		<title>By: &#187; Visual concept : Google Images Search Categories Ramenos Blog</title>
		<link>http://cybernetnews.com/2007/05/29/search-for-faces-with-google-image-search/#comment-96788</link>
		<dc:creator>&#187; Visual concept : Google Images Search Categories Ramenos Blog</dc:creator>
		<pubDate>Tue, 29 May 2007 16:10:59 +0000</pubDate>
		<guid isPermaLink="false">http://tech.cybernetnews.com/2007/05/29/search-for-faces-with-google-image-search/#comment-96788</guid>
		<description>[...] have read the news on this blog. You can test a search with Google Images and you&#8217;ll see that the Search Engine knows how to [...]</description>
		<content:encoded><![CDATA[<p>[...] have read the news on this blog. You can test a search with Google Images and you&#8217;ll see that the Search Engine knows how to [...]</p>
]]></content:encoded>
	</item>
</channel>
</rss>
