EU puts the web just a photo away

 

Gallery

Posted on 5 Jan 2009 at 11:57

An EU funded project has developed technology that can recognise features in images and automatically find related information on the internet.

Researchers working on MOBVIS—Vision Technologies and Intelligent Maps for Mobile Attentive Interfaces in Urban Scenarios— say that the system can accurately identify features, such as buildings monuments or even bus stops and other street furniture. It then adds an icon to the image that points to the related information

They believe that the potential range of applications is vast, particularly when the technology is available on mobile phones. For instance the user could get details about bus routes are served by a particular stop or instantly compare shopping prices.

Tourism is another potential market. Instead of getting a guidebook that people read to discover where to go, take a picture and find out the history and culture of where you are. Similarly when buying property, house hunters could take a picture of the streetscape to find what is available locally.

“Really, we do not know all the applications that could potentially be developed from a technology like this. We can easily imagine some of the obvious ones, but after that, who knows?” said Lucas Paletta, coordinator of the MOBVIS project.

The system works from a database of geo-referenced panoramas, that contain details of the buildings, monuments, banners and logos that appear in the panoramas. Information relating to individual buildings or monuments is then added to the database manually. It even allows for the fact that in many cities, many of the buildings can be very similar. A feature-matching algorithm developed by the University of Ljubljana in Slovenia, can accurately detect minute but telling differences between similar objects, both by the appearance of the buildings themselves and their context in the streetscape.

The result is a technology that has proved highly accurate in real world tests. Users were given a five-minute instruction by an outside contractor, and then sent around to explore the Austrian city of Graz with their mobile phones.

The system reliably detected the right building 80% of the time, a figure that Aleš Leonardis, head of the Ljubljana team is convinced can be improved.

“But that’s not the most remarkable result of the prototype test,” Leonardis said. “It was remarkable that there were no false positives. Sometimes the system couldn’t identify a building, but it never put the incorrect link on a building.”

Author: Simon Aughton