Abstract
Past research strongly suggests that object representations influence attentional selection. The classic two-rectangle paradigm (Egly, Rafal & Driver, 1994) demonstrated the object based attention (OBA) effect, wherein attention is prioritized to regions within an attended object compared to regions inside a separate, equidistant object. This effect is accentuated when low-level features differ between the two objects (Shomstein & Behrmann, 2008). Although OBA research has concentrated on attentional allocation within simple shapes, our natural environment is rich in objects inherent with both low-level (e.g., color) as well as high-level (e.g., a knife) properties, and an efficient visual system would benefit from utilizing either source of information. However, the effect of high-level semantic information on OBA has yet to be elucidated. In a series of experiments, we adapted the two-rectangle paradigm to contain real-world objects and manipulated low-level feature and high-level semantic properties, independently. Low-level features were manipulated by rendering the color between objects same or different. Conversely, high-level properties were manipulated by varying the strength of the semantic relationship between the objects: high (e.g., fork-knife), low (e.g., fork-marker), or same (e.g., fork-fork). Attentional prioritization between object pairs varied with the strength of the semantic relationship, independently of low-level features. A strong semantic relationship increased prioritization within the same, attended object, while both weak and same semantic object pairs showed increased prioritization within the unattended, different object. We thus conclude that low-level feature and high-level semantic information affect object-based attentional allocation independently, and that semantic information can override low-level features when guiding attention. These findings strongly advocate the need to investigate object-based attentional guidance in real-world images, and provide further constraints on mechanisms of attentional selection.
Meeting abstract presented at VSS 2013