Video: Man vs machine in football fan battle

Can this computer understand the offside rule?

Published: 17 April 2007 13:00 GMT by Steve Ranger

Tags: bt

• Latest Videos
• Most Popular Videos

Which is a better football fan - man or machine?

That's a puzzle BT researchers are working on.

The researchers, working at the Adastral Park labs near Ipswich, have developed an algorithm that tries to spot the most interesting bits of a football game - the goals, corners and rows with the ref.

The idea is that in future the algorithm - in development for about five years so far - could be used to generate football highlights packages automatically. In the longer term the technology could be used to help computers spot emotions, so a machine could recognise funny or sad parts of a video, for example.

To make sure the algorithm performs correctly, it has to be tested against the actual responses of a living, breathing human football fan. Both the computer and the human test subject watch the same game of football. The responses of the human guinea pig are monitored by their skin conductance, then that data is compared to what the computer picked out as interesting.

BT broadband applications researcher Dave Chatting, part of the team that conducts the experiments, said: "You sweat when you find things exciting."

silicon.com wanted to see what all the fuss was about so news editor Steve Ranger offered himself up as a guinea pig.

How did his responses to the Blackburn-Arsenal FA Cup replay measure up to what the algorithm thought were the most interesting parts of the game?