Graphics give an alternative representation model for images and graphic rewriting gives another viewpoint for image matching, image retrieval and identity of shape constancy - all subjects that have raised passionate discussion of mathematical models for both machine and human vision[1,2,3,4,5,6,7,8]. A graphic is a graph where all vertices get both a label and other attribute values from a \sum-algebra, so our graphic grammars are a particular kind of attributed graph grammar: Hess et al [9,10]. Attributed graph grammars are a powerful way of describing transformations of complex structures. Many kinds of attributed graph grammars have been studied but they all seem to describe graph transformations as pushouts in an adhesive HLR category [11], so they enjoy such desirable theoretic properties as locally Church-Rosser and Ehrig Parallelism and Concurrency [12]. Attributed graph grammars have been applied to describe visual languages [13], distributed systems [14], chemical reactions [15], music analysis [16] and much else. In this paper we describe a new application: analysis of how we perceive shape and motion.