Computational models of attention

TL;DR

This chapter reviews recent computational models of visual attention, covering bottom-up stimulus-driven models and top-down goal-oriented models, highlighting their categories and differences.

Contribution

It provides a comprehensive overview of recent models of visual attention, categorizing and comparing bottom-up and top-down approaches.

Findings

Seven categories of bottom-up attention models analyzed

Discussion of models guiding attention based on relevance to tasks

Comparison of stimulus-driven and goal-oriented attention mechanisms

Abstract

This chapter reviews recent computational models of visual attention. We begin with models for the bottom-up or stimulus-driven guidance of attention to salient visual items, which we examine in seven different broad categories. We then examine more complex models which address the top-down or goal-oriented guidance of attention towards items that are more relevant to the task at hand.