Specifying the Components of Attention in a Visual Search Task

Although commonly treated as a unitary process, attention is more likely a collection of task-related but separable operations. Three components of attention (set, selection, and movement) are identified and defined within the context of a computationally explicit model of eye movements during visual search. The model compares filter-based representations of the target and search displays to derive a salience map indicating likely target candidates in a scene. Eye position is defined as the centroid of activity on this saliency map. As this map is put on the threshold over time, the changing centroid produces a sequence of movements that eventually cause simulated gaze to become aligned with the target. By adopting a more computational language and making explicit the underlying operations of the task, visual search, a behavior that has long been hobbled to the concept of attention, can be well described without appeal to an abstracted attention theory.