Structured Descriptions of Complex Objects

Experimental techniques are demonstrated which generate segmented symbolic descriptions for complex objects with joints, such as a hammer or a glove. Complete "descriptions with relationship of parts at joints and descriptions of joints are presented. These techniques are elements of a larger scheme for description mechanisms for hypotheses, and for visual memory and recognition. 1. Introduction We describe experiments with symbolic description of curved objects. Tn previous work1,2 descriptions were generated for the pieces of objects, according to a volume representation of shape3. The pieces are an adequate description only of objecls with a single part, like a torus or a snake. We have generalized and improved techniques for descriptions of pieces. We have made complete descriptions which join the scattered pieces into a whole. These descriptions are the basis for recognition, and a few examples are discussed. The motivations for studying representation of shape are: for visual systems of robots which use depth or image data; for the interactive programming of assembly robots for industrial tasks; for the computerization of industry-machining, parts description, systemstization of assembly operations; for display; and for its relevance to other areas of A.J. and biology. Our concern is implementation of symbolic descriptions oi shape and space which are adequate for integrated robot systems. We do not think a single representation is adequate throughout a system, but we have attempted to analyze tasks in terms of computational primitives originating from topological and geometrical primitives, and design representation abilities accordingly. Representative-tasks are manipulation , display, and recognition. Manipulation requires mass and momenta of articulated parts, and ability to calculate overall moments in various positions and orientations. Display requires description of Surface properties and two dimensional proximity, to find overlapping surfaces tn the image. Recognition requires visual memory for generation of hypotheses, prediction and verification facilities, geometric relations among parts, and a good choice of parts. 2. Representation and Models The representations-' depend upon segmentation of complex objects into parts. Parts can themselves be composed of subparts. The value of the part/whole segmentation depends on a useful representation for primitive parts. The basic topological operations of cutting and pasting are used in joining parts. The representation oi parts and joints amounts to the basis tor an intelligent guess about structure. Special knowledge about joints, e.g. the anatomy of humans, is much more powerful. Primitive parts arc described by volume representations. The primitive parts are arm-like pieces which …