Again, this isn't essential for Finals (unless you're doing the AI and vision option). But it's part of the AI culture, and will help you see the place of Evans work in history.
It's reasonable to assume that the output from a vision program will be a description of the objects it thinks it can see. The program might go as far as identifying individuals; or it might stop at classifying each object as being of a given type, leaving some other module to go further and distinguish Fido the dog from Rover the dog. In any case, we need to be able to represent objects. This topic comes up in Marr's work, via the Marr and Nishihara model of object recognition.
Many such representations are derived from Evans' work. At the time he wrote, the idea of representing geometric objects this way was new and exciting. Remember the quote I asked you to read on pages 278---279 of Semantic Information Processing
Regardless of what devices are used to put pictorial information into the computer and get it out again, we must meet the questions of what form the internal representation is to take and how it is to be processed. This area is far less developed ... Among pattern-recognition programs perhaps only that of Grimsdale could be viewed as generating from such input something that could be reasonably regarded as a higher-level description of the input figure ... At least one great merit of description processing is that it permits convenient treatment of complex pictorial material that can best be treated as comprising one or more levels of interrelated parts (as are the figures occurring in the geometric-analogy questions). ...
The idea was taken up by Winston, a student of Evans, in his concept-learning program. Read the description on pages 174--179 of Vision by Bruce and Green, and then Chapter 11 of Winston or Chapter 10 of Boden. That will give you a good idea of the kind of descriptions this method could handle. (Winston's original paper is in []).
Finally, for more recent work, look again at the Image Understanding article in []. I have also put two papers into the Psychology photocopies: [] and [].