Role of Knowledge in Visual Shape Representation

摘要

This thesis shows how knowledge about the visual world can be built into a shape representation in the form of a descriptive vocabulary making explicit the important spatial events and geometrical relationships comprising an object's shape. We offer two specific computational tools establishing a framework by which a shape representation may support a variety of later visual processing tasks: 1) By maintaining shape tokens on a Scale-Space Blackboard, information about configurations of shape events such as contours and regions can be manipulated symbolically, while the pictorial organization inherent to a shape's spatial geometry is preserved. 2) Through the device of dimensionality-reduction, configurations of shape tokens can be interpreted in terms of their membership within deformation classes; this provides leverage in distinguishing shapes on the basis of subtle variations reflecting deformations in their forms. The power in these tools derives from their contributions to capturing knowledge about the visual world. In contrast to building block approaches to shape representation we employ a large and extensible vocabulary of shape descriptions tailored to the constraints and regularities of particular shape worlds. The approach is illustrated through a computer implementation of a hierarchical shape vocabulary designed to offer flexibility in supporting important aspects of shape recognition and shape comparison in the two-dimensional shape domain of the dorsal fins of fishes. Keywords: Shape representation; Dimensionality; Scale space; Later vision. Theses. (EDC)