Viewing the molecular display: Hints from the artificial intelligence of vision

Abstract

In this paper I describe a number of “depth cues” which permit the viewer of computer graphics displays to deduce three-dimensional information from the two-dimensional display. These may be divided into “bottom-up” and “top-down” aids to interpretation. These depth cues can be emphasized in the computer graphics displays to assist the novice viewer, with an expert system adapted to the purpose. I provide a brief and general description of expert systems, including a caricature which deals with some of the problems encountered in molecular modeling with computer graphics. The central construct of the expert system DOCENT is its capacity to represent a macromolecule by “generalized cylinders”, and to permit manipulation of the cylinders directly, instead of the usual internal coordinates or space-fixed axes.