The structure of realistic scenes as a determinant guiding the visual search

Abstract

Visual search in scenes is guided by a variety of factors, each based on different types of scene-related information. One of these is visual salience, which relies on visual features within the scene and directs attention from the most salient areas to the least. Another factor is the scene's meaning and structure, which involves the congruence of objects with the scene they are located in. Congruence can refer to an object's high likelihood of belonging to the scene (meaning) and its spatially expected placement (structure). This paper is aimed to review these factors, identifying the limitations of early models of visual search in scenes, comparing J. Wolfe's model with recent findings on structural-meaning factors, and addressing a series of emerging theoretical questions.