Computational Models
This chapter reviews recent progress in computational modelling of visual attention. The authors start with early concepts and models, which have emphasized stimulus-driven guidance of attention towards salient objects in the visual world. They then present a taxonomy of the many different approaches which have emerged in recent research efforts. They then turn to the more complex problem of modelling top-down, task- and goal-driven influences on attention. While early top-down models have been more qualitative in nature, the authors describe several recent fully computational approaches that address top-down biasing in space, over features, and towards objects. This chapter finally provides an outlook and describes promising future research directions.