Seeing numbers - can computing hope to recreate human vision?


Seeing numbers - can computing hope to recreate human vision?

Justin Richards

This year's Roger Needham lecture - Computer Vision and the Geometry of Nature - will examine the search for mathematical models and algorithms that can explain and emulate the human visual abilities most of us take for granted.

We can easily recognise thousands of objects, follow complicated movements, and almost subconsciously build a three-dimensional view of the world through stereo vision.

When a camera captures a scene, its rich visual complexity is not lost and we can still enjoy the images and recognise the picture's contents.

However, the visual patterns are translated into complex numerical arrangements which current mathematics and statistics strive to represent and understand.

Although science is far from having a complete understanding of the processes of vision, the past decade has seen applications of artificial vision move out of the lab into the real world.

In the lecture, Andrew Fitzgibbon will be talking about the use of computer vision in obtaining a 3D representation of the world, and the application of these techniques with regard to cinematic special effects in films such as the Harry Potter and the Lord of the Rings series.

It will begin by looking at various aspects of the human visual system, and try to answer the question of whether it is reasonable to expect to emulate human vision without first building an artificial intelligence.

Fitzgibbon will describe several classic experiments which suggest that not all tasks require artificial intelligence and will consider applications where these tasks arise, for example robot navigation and special effects.

He will also show how a combination of engineering and geometry can provide reliable systems, after looking at how a man-made world simplifies the solutions to these problems.

"We will be able to see how modern models of the statistics of the irregular natural world can be brought to bear, in combination with the paradigm of Bayesian inference, yielding powerful algorithms, which work on trees and clouds as well as on buildings and roads," said Fitzgibbon.

The final section of the lecture will speculate on the tractability of some harder problems, such as object recognition and unsupervised learning of visual competences.

The Roger Needham Lecture will take place on 23 October at the Royal Society in London.

More information


Email Alerts

Register now to receive IT-related news, guides and more, delivered to your inbox.
By submitting your personal information, you agree to receive emails regarding relevant products and special offers from TechTarget and its partners. You also agree that your personal information may be transferred and processed in the United States, and that you have read and agree to the Terms of Use and the Privacy Policy.

COMMENTS powered by Disqus  //  Commenting policy