Sunday, October 11, 2009

Where is machine vision headed?

If you’re reading this you’re probably a believer in the value of machine vision. You see the benefits of using computers to analyze images and make decisions about what they “see.” But if you lift your head up from the here-and-now to contemplate a world with machine vision embedded in every device, appliance, road sign and billboard, you might start to wonder where we’re headed.

Vanilla Sky” was one of the first movies to postulate a world of total surveillance – remember how the advertising hoardings recognized Tom Cruise as he walked past? More recently, “Eagle Eye” has raised awareness of what intelligent surveillance might look like. So should we be worried?

I believe that we should start to discuss to what extent we want to be recognized, identified and tracked as we go about our daily lives. There are some real civil liberties issues at stake and as scientists and engineers we have a responsibility to inform and alert the general populace as to what will soon be possible. But I don’t believe the machines are about to take over. In fact I seriously doubt whether we will ever see computers that can really think.

In this regard I’m on the same page as Noel Sharkey, Professor of Artificial intelligence and Robotics at the University of Sheffield. Interviewed in New Scientist, (September 1st, 2009) he has some very smart things to say about AI and where it’s taking us.

Read and reflect.

1 comment:

Fred Hapgood said...

Perhaps just as big a change might follow on developing the ability to move competent machine vision across applications cheaply. Most of what humans do for a living is defined by what machines can't do cheaply, and a huge portion of that has to do with using vision to do generalized sorting, assembly, path navigation,
and manipulation. Once machines get as good as pigeons at general pattern recognition a very large number of those jobs will be at risk. I recall reading the estimates of one analyst, whose name I have alas forgotten, at about 30% of all jobs. Replies if any to hapgood@pobox.com.