Inexact visualization: qualitative object representation for recognizable reconstruction

The use of qualitative models as opposed to quantitative models for object recognition and reconstruction is discussed. Central to the author´s model is an attempt to identify a minimal subset of attributes relevant to visualization. Each object is decomposed into simpler components, with each component described as a qualitative generalized cylinder and the joining operations specified by a three-dimensional qualitative spatial reasoning scheme. The spatial reasoning used is systematically derived and provides a complete description under considerations of tangency, nocontact, and overlap. The representation must be compact, yet contain sufficient information to be ´recognizable´-a term that is defined in terms of human perception and constitutes a measure of the goodness of the model