The RGFF representational model:: A system for the automatically learned partitioning of "visual patterns" in digital images

This paper describes a system for the automatically learned partitioning of "visual patterns" in 2D images, based on a sophisticated, band-pass, filtering operation with fixed scale and orientation sensitivity. In this scheme, the "visual patterns" are defined as the features which have the highest degree of alignment in the statistical structure across different frequency bands. The analysis reorganizes the image according to a constraint of invariance in statistical structure and consists of three stages: 1) pre-attentive stage, 2) integration stage, and 3) learning stage. The first stage takes the input image and performs filtering with a set of log-Gabor filters. Based on their responses, activated filters which are selectively sensitive to patterns in the image are short listed. In the integration stage, common grounds between several activated sensors are explored. The filtered responses are analyzed through a family of statistics. For any given two activated filters, a distance between them is derived via distances between their statistics. The third stage, the learning stage, performs cluster partitioning as a mechanism for learning the subspace of log-Gabor filters needed to partition the image data. The clustering is based on a dissimilarity measure intended to highlight scale and orientation invariance of the filtered responses. The technique is illustrated on real and simulated data sets. Finally, this paper presents a computational visual distinctness measure computed from the image representational model based on visual patterns. It is applied to quantity the visual distinctness of targets in complex natural scenes. Several experiments are performed to investigate the relation between the computational distinctness measure and the visual target distinctness measured by human observers.