Computer science from A to Z

Y is for eYes

© INRIA Sophie Auvin - Y comme Yeux

Lions have great strength, but it would be of no use to them if nature had not given them eyes”. [Montesquieu]

Machines do not have eyes, but why shouldn't they be capable of seeing?

The first computer scientists who, as far back as the 1950s, were programming their computers to process images, sought above all to improve their quality: increase the contrasts, strengthen the contours, etc.

Then, with digital images becoming more and more numerous, their classification began to be automated: do these aerial photos show urban areas or forests? Which identity photos most resemble the photofit of the suspect?

In addition to classification, we are now working on the detailed interpretation of images: are those shapes on the road two cars or two parts of the same truck? Is that mark on the X-ray a tumour? Image processing software is enriched by general knowledge about the objects around us, their behaviour and their interactions: we can finally say that computers can see!

Did you know...?

© INRIA / Sophie Chauvin - Yeux

An autonomous robot must see its environment in three dimensions and in motion. To achieve this, some researchers are drawing inspiration from the visual apparatus of a frog which, while rudimentary, allows it to spot and capture moving flies.

Keywords: Computer science from A to Z Science awareness

Top