New visual search technology from Google
New software to enable image searching is being described by Google. According to a paper (pdf) given this week at the World Wide Web conference in Beijing, VisualRank addresses the problem of searching images for content. Unlike text, "human recognizable objects are usually not automatically detectable in images."
Shumeet Baluja and Yushi Jingy applied Google's Web search tool, PageRank, to create large matrixes of related images, not based on attached text tags but on their proximity in users' searches.
Instead of modelling the relationship between objects and image features, we model the expected user behavior given the visual similarities of the images to be ranked. By treating images as web documents and their similarities as probabilistic visual hyperlinks, we estimate the likelihood of images visited by a user traversing through these visual-hyperlinks. Those with more estimated “visits” will be ranked higher than others.
This project is not available to the public yet, but has apparently resulted in improved image search results.
Comments:
No Comments for this post yet...
Leave a comment
Allowed XHTML tags: <p, ul, ol, li, dl, dt, dd, address, blockquote, ins, del, span, bdo, br, em, strong, dfn, code, samp, kdb, var, cite, abbr, acronym, q, sub, sup, tt, i, b, big, small>
