Now, computers that come with common sense

November 14th, 2007 - 2:27 am ICT by admin  
It allows the automated labelling system to identify objects in photographs.

Whereas a conventional automated object identifier would label a person, a tennis racket, a tennis court and a lemon in a photo, the new post-processing context check will re-label the lemon as a tennis ball.

Normally, computers with the latest image labelling algorithms don’t have the contextual wits to know that a lemon in a lawn set up is out of sync.

The new research aims to change that.

“We think our paper is the first to bring external semantic context to the problem of object recognition,” said computer science professor Serge Belongie from UC San Diego.

In the research, the scientists found that Google Labs tool called Google Sets could be used to provide external contextual information to automated object identifiers.

Google Sets generates lists of related items or objects from just a few examples.

If one types in John, Paul and George, it will return the words Ringo, Beatles and John Lennon. If one types “neon” and “argon”, it will give the list of the rest of the noble gasses.

In some ways, Google Sets is a proxy for common sense. In our paper, we showed that you can use this common sense to provide contextual information that improves the accuracy of automated image labeling systems,” said Belongie.

Belongie said the image labelling system involved a three-step process.

First, an automated system split the image up into different regions through the process of image segmentation.

In the tennis set up photograph, the Image segmentation separated the person, the court, the racket and the yellow sphere.

Next, an automated system provided a ranked list of probable labels for each of these image regions.

Finally, the system added a dose of context by processing all the different possible combinations of labels within the image and maximizing the contextual agreement among the labelled objects within each picture.

To make the system foolproof, the researchers also injected context into an automated image labelling system through a post-processing context check.

According to the scientists, their approach strives to maximize the contextual agreement among the labelled objects within each picture.

A paper describing the research will be presented on October 18 2007 at ICCV 2007 - the 11th IEEE International Conference on Computer Vision in Rio de Janeiro, Brazil. (ANI)

Related Stories

Tags: , , , , , , , , , , , , , , , , , , ,

Posted in Sci-Tech |