(Page 2 of 2)
As Lincoln is downloaded by more people, and they add pictures and links, the application will become more useful. However, at this early stage, it's unclear whether or not the user experience is good enough to attract people to the software in the first place. "The question is whether it's above or below the threshold of what the user will take," says Luis von Ahn, professor of computer science at Carnegie Mellon University, in Pittsburgh. "It's always possible to do [image recognition] that's better than nothing, but it's hard to do something that's perfect."
Still, he believes that the "application is really cool," and that importantly, the research leverages user-generated information to make the Web links relevant to a given picture. In his own work, von Ahn has developed computer games for people to play that train computers to recognize objects in pictures. The technology is now the basis for the Google Image Labeler, which consists of a game that helps Google serve up more-accurate picture results for keyword searches. Microsoft's approach of having a computer match images in a completely automated way is aiming for the gold standard of computer vision, says von Ahn, but he believes that there are decades of work to be done before the standard will be achieved.
A new approach to face recognition is better at handling low-resolution video.
A modular microscope attachment for cell phones could improve the quality of telemedicine.
If one is interested in downloading "Lincoln", a search of the Microsoft web site using the keyword "Lincoln" will, curiously, not yeild results. Go to Larry Zitnick's home page for a test drive.
http://research.microsoft.com/~larryz/
Thanks for the article, Kate - cool stuff. This works for me when I capture a photo of Mona Lisa, but it failed to identify any of a dozen well-known O'Reilly programming books and Disney DVD movies. It even failed when I photo searched my cover of MIT Technology Review/February - the same magazine cover depicted in this article's graphic. Has anyone had better success?
Damion Hankejh
ingk.com
For the kooaba search engine, the recognition time is far below 1 second for tens of thousands of images (see the video on YouTube), but I can confirm that the whole process over MMS/e-mail takes about 5-8 seconds. Still, with a regular mobile search engine like Google, it takes in average about 40 seconds for a search query. Moreover, it is highly probable that the desired content is not even listed among the search results. With image search, the query is well defined. If for instance the user takes an image of the Moulin Rouge in Paris, he will get only information about the very building like history, etc. If, however, the user takes an image of the Moulin Rouge movie poster, she gets information about the film. In text-based search engines, it's not clear what the user is looking for when typing "Moulin Rouge".
Feel free to test the kooaba search engine on kooaba.com
Manufacturing in the United States is in trouble. That's bad news not just for the country's economy but for the future of innovation.
Our list of the 50 most innovative companies, including the following:
On Twitter
Become a Fan on Facebook
Subscribe to the Feed
deejay
23 Comments
Mobile Web Searches Using Pictures
Very interesting!
But is it only restricted to mobiles? How about if a web-cam is used to take the pix and a search is made using IE?
Reply