John Tsotsos (York University, Canada)
Thursday, 26.5.2016, 11:30
A functional, computational and mechanistic, explanation of the relationship among visual attention, interpretation of visual stimuli, and eye movements, and how these produce visual behavior, seems elusive. Here, we focus on one component, how selection is accomplished for the next fixation. The popularity of saliency map models drives the inference that this is solved; but we argue otherwise. We provide arguments that a cluster of complementary, conspicuity representations drive selection, modulated by task goals and history, leading to a blended process that encompasses early, mid-level and late attentional selection and reflects the differences between central and peripheral processes. This design is also constrained by the architectural characteristics of the visual processing pathways, specifically, the boundary problem, as well as retinal photoreceptor distribution. These elements combine into a new strategy, caricatured below, for computing fixation targets and a first simulation of its performance is presented.