This research addresses the representation of point patterns using nonmetric information. Point patterns are especially useful as feature point models of more complex images. Image representation is essential to both human and machine vision. Nonmetric information is of particular importance since there is considerable evidence that human vision utilizes it extensively, as opposed to precise, metric measurement. For many visual cognition tasks, it seems important for humans to be able to look for visual patterns, which must be very general and flexible thus not metric. This research will contribute to both the understanding of human visionand the development of machine vision based on relative imprecise, inexpensive sensors
