What if computers could see?


Richard Wildes, left, and VP Research & Innovation Stan Shapson with Wildes’ CFI certificate

If computers could see the way Richard Wildes imagines they could, they might be friendlier and more personable. How? They could automatically interpret the movements and gestures that their users make. Interaction with computers would thus become more natural as they began to take advantage of cues that humans use in their day-to-day interactions.

Gesture recognition is just one of the projects Wildes, a computer scientist and vision researcher, is pursuing in a new laboratory at York, set up with funds from the Canada Foundation for Innovation New Opportunities program.

Another problem Wildes is studying in the CFI-funded lab is how information about the physical and geometric structure of the world can be distilled from camera images of a scene.

This research could lead to more sophisticated computer interpretation of aerial images from binocular imagery or a video sequence, and could be used in geomatics – computerization applied to geography – and space robotics.

Left: an example of a robot that could be used in space or planetary exploration

Wildes, who obtained his PhD from the Massachusetts Institute of Technology in 1989, is interested in machine and biological vision, and allied aspects of image processing, robotics and artificial intelligence. He is particularly interested in the analysis and understanding of binocular and motion imagery.

The new lab will be devoted to analytic and empirical studies of visual information processing, with an emphasis on binocular and motion imagery. The work may enhance our understanding of how biological systems see. And it could pave the way to endowing machines with sight.