A Normalized Redundancy representation for 2D digital images

This paper presents a representational model in which spatial information i s analyzed by multiple filters, each of which is sensitive to patterns that contain spatial frequencies in a restricted range. The model bases its res ponses on only those filters sensitive to shapes that produce an emergent d istribution of activity (clumps of energy) upon the image spectrum. Spatial structures are described at their best scales as given by a novel Normaliz ed Redundancy approach based on the intrinsic redundancy of natural scenes. (C) 1998 Elsevier Science B.V. All rights reserved.