The Stochastic Dynamics Of Visual Attention

Abstract

This project develops and tests a class of mathematical models of how information is represented statistically in the human visual system and how that representation is modified by selective attention. These models synthesise current theories about how the visual system encodes information and how people make judgements on the basis of that information. The domain of application of the existing class of models will be extended to allow them to predict performance in more complex settings than has hitherto been the case. A program of experimental studies will test the models by measuring human performance in response to computer-generated stimuli.