ABSTRACT
A basic fact of vision is that all opaque, solid objects, seen from any vantage point, have visible and invisible parts. Any model of solid shape that is tailored to vision must at least be able to predict which surface patches go in or out of sight when the observer moves with respect to the object. Modern mathematics is able to provide a complete inventory of qualitatively different solid shapes and provides the means to describe all visual events when the observer makes exploratory movements. The structures revealed by such analysis must form the backbone of any useful (for sensorimotor coordination) model of shape. They also form a basis for any psychophysical research on solid shape perception.
