5 Computer Vision (1970-2000)

As briefly mentioned in the preface, computer vision is the problem of recognizing and locating the objects in a scene. The input to a computer vision program usually is the light intensities at a (typically large) number of points in the image, called pixels. In the early days of AI, the ease with which humans interpret visual scenes led researchers to vastly underestimate the difficulty of the problem. At MIT, Richard Greenblatt was hired to lead a group of undergraduates for a summer project to solve vision. This was while Greenblatt was still an undergraduate himself and before he did his work on chess. Needless to say, they made no real progress, and Greenblatt never worked on vision again.