The Next Generation of Image Search
|
Written By Drupad Sil | April 28, 2008 | Share This
|
|
A new innovation in image search may soon prevent this picture from showing up for a query of “mcdonalds”. This is a story that’s gotten quite a bit of coverage today, starting with the New York Times. From the Times:
“On Thursday at the International World Wide Web Conference in Beijing, two Google scientists presented a paper describing what the researchers call VisualRank, an algorithm for blending image-recognition software methods with techniques for weighting and ranking images that look most similar.”
How is this different from what is currently done? Danny Sullivan at SearchEngineLand:
“Image search at the major search engines today relies largely on looking at words that are used around images – on the pages that host them, in image file names and in ALT text associated with them. No real image recognition is done by any of the majors. Search for “apples”, and they haven’t actually somehow scanned the images themselves to “see” if they contain pictures of apples.”
In their paper, Yushi Jing and Shumeet Baluja introduce algorithms that can actually “look” at the image itself rather than the associated text, find similarities, and rank the pages in order of similarity to an original image deemed to be the correct result for the query. In the words of VentureBeat’s Anythony Ha:
“The new system proposed in the Google paper ranks images based not on text, but on the common ‘visual themes’ found in each search result. In the McDonald’s example, the VisualRank system would see that the company’s famous golden arches are a common visual theme, and prioritize pictures that feature the arches prominently. Testes of this new system returned 83 percent fewer irrelevant search results than Google Image Search, according to the VisualRank paper.”
This is definitely a cutting-edge development, if it can be implemented successfully. It will help cut back on duplicate images, but more importantly, will reduce image spam, where photos are tagged inappropriately and show up for unrelated searches.
Topics: Google, Search: Image, Search: Innovations, Technology |
[…] Maple Leaf 2.0 - Technology and Web 2.0 News in Canada wrote an interesting post today onHere’s a quick excerpt […]