Scientists Use Google Widget to Improve Image Labeling


	Scientists Use Google Widget to Improve Image Labeling

mugofcoffee

10-22-2007 8:35 AM

541 views

tags:

google, scientists, research, computer vision, pictures, semantic web.

mugofcoffee says:

The Omni-potent & Omni-present, The single Universal Force above us all, The God Almighty GOOGLE!

Add a Comment

Login to Comment. Not a member yet? Sign up

Today's Top Clips

visit the Top Clips page

View the Top Clips from October 22, 2007

Embed This Clip In Your Site...

<div style="margin: 12px 0px; font-family: arial; color: #333333; background: #ffffff; border: solid 4px #e5e5e5; width: 100%; clear: left;"><div class="CM_CTB_Content_Wrap" style="margin: 0px; padding: 0px;background-color: #ffffff;"><div style="border-bottom: solid 1px #dcdcdc; white-space: nowrap; margin-bottom: 8px; background-color: #eeeeee ;background-image: url(http://clipmarks.com/images/source-bg.gif); background-repeat: repeat-x; height: 24px; line-height: 24px; vertical-align: middle; padding-bottom: 4px; color: #666666; font-size: 10px;" ><a href="http://clipmarks.com/clip-to-blog/" title="see clips that are hot right now"><img src="http://content.clipmarks.com/blog_embed/328c0eb1-a460-42d8-b2e4-ed20d1b83bd9/8DE8D39F-650E-455C-9582-F4D13A4F85CD/" alt="" width="19" height="19" border="0" style="vertical-align: middle; margin: 0px 4px; display: inline; border: none; float:none;" /></a>clipped from <a title="http://blog.wired.com/wiredscience/2007/10/scientists-use-.html" href="http://blog.wired.com/wiredscience/2007/10/scientists-use-.html" style="font-size: 11px;">blog.wired.com</a></div><blockquote style="text-align: left; padding: 0px 8px; margin: 4px 0px 8px 0px; background: transparent; border: none;" cite="http://blog.wired.com/wiredscience/2007/10/scientists-use-.html"><P><A href="http://blog.wired.com/.shared/image.html?/photos/uncategorized/2007/10/18/contextimages.jpg"><IMG title="Contextimages" height="173" alt="Credit: UC San Diego" hspace="0" src="http://blog.wired.com/wiredscience/images/2007/10/18/contextimages.jpg" width="250" vspace="0" border="0" /></A> University of California researchers <A href="http://blog.wired.com/wiredscience/2007/10/”http://www.sciencedaily.com/releases/2007/10/071017174328.htm">are claiming</A> they’ve added “common sense” to computers’ ability to recognize objects in photographs.  </P></blockquote><div style="height: 2px; font-size: 2px; background: #dcdcdc; border-bottom: solid 1px #f5f5f5; margin: 2px 4px;"></div><blockquote style="text-align: left; padding: 0px 8px; margin: 4px 0px 8px 0px; background: transparent; border: none;" cite="http://blog.wired.com/wiredscience/2007/10/scientists-use-.html"><P>The team will present their findings today at the IEEE’s <A href="http://blog.wired.com/wiredscience/2007/10/”http://iccv2007.rutgers.edu/">International Conference on Computer Vision</A> in a paper called “Objects in Context” (<A href="http://blog.wired.com/wiredscience/2007/10/”http://www.google.com/url?sa=t&ct=res&cd=2&url=http:%2F%2Fwww.cs.ucsd.edu%2F~sjb%2Ficcv2007a.pdf&ei=spMXR4imJYaUgQPYv6X4Bw&usg=AFQjCNFHsrdvvyTn1_3vK5rwpawUOGA-Iw&sig2=q5G70aWGHO1LGYKnIs5Lbw">pdf</A>). Their technique could immediately improve the automatic cataloging of images in large collections (like those at the <A href="http://www.loc.gov/">Library of Congress</A>), albeit marginally in the early going. Down the road, it could provide computer object recognition systems with <A href="http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=33948">the context</A> that allows humans to fill in our visual fields with usually reliable information. Really good systems, combined with metadata about when the photo was taken, for example, could provide the information for a <A href="http://en.wikipedia.org/wiki/Semantic_Web">semantic web</A> for pictures, allowing automatic geographical and event tagging. <BR /></P></blockquote></div><div style="margin: 0px 6px 6px 4px;"><table style="font-size: 11px;border-spacing: 0px;padding: 0px;" cellpadding="0" cellspacing="0" width="100%"><tr><td style="background:transparent;border-width:0px;padding:0px;">&nbsp;</td><td align="right" style="background:transparent;border-width:0px;padding:0px;width:107px" width="107"><a href="http://clipmarks.com/share/8DE8D39F-650E-455C-9582-F4D13A4F85CD/blog/" title="blog or email this clip"><img src="http://content9.clipmarks.com/images/c2b-foot.png" border="0" alt="blog it" width="107" height="17" style="border-width:0px;padding:0px;margin:0px;" /></a></td></tr></table></div></div>

New from the makers of Clipmarks: Amplify.com - Don't just share the news...Amplify it!

Clipmarks: Home; New Clips; Top Clips; Dashboard

Popular Topics: News; Life; Science; Technology; Entertainment

Get Started: Sign Up; Install Clipping Tool; How Clipping Works; Clip-to-Blog™; ClipSearch

Tools and Resources: FAQ; ClipWeek; Top Clippers; Top Tags; Site Map

About Clipmarks: About Us; Contact; Copyright; Privacy; EULA

Search Options

Today's Top Clips