Automatic synthesis of graphical object descriptions

Abstract

A technique is presented for automatically synthesizing graphical object descriptions from high-level specifications. The technique includes mechanisms for describing, selecting, and combining primitive elements of object descriptions. Underlying these mechanisms are a referential framework for describing information used in the construction of object descriptions and a computational model of the object-synthesis process. This technique has been implemented in two prototype systems to synthesize object descriptions in near-real time. One system creates graphical displays of information that resides in a conventional database. The other system is a computer graphicist's tool for creating backgrounds of complex, three-dimensional scenes.