Academic journal article Psychonomic Bulletin & Review

Attentional Selection of Complex Objects: Joint Effects of Surface Uniformity and Part Structure

Academic journal article Psychonomic Bulletin & Review

Attentional Selection of Complex Objects: Joint Effects of Surface Uniformity and Part Structure

Article excerpt

What object properties warrant selection by object-based attention? Previous research has suggested that surface uniformity is required for object-based attentional selection (Watson & Kramer, 1999), yet nonuniform objects are encountered frequently. In the present experiments, we investigated the interplay between surface uniformity and part boundaries and their effect on object-based attention. Specifically, we asked if attention can select nonuniform objects whose surface changes occur at part boundaries. Although uniformly colored objects did exhibit object-based effects, we only observed an object-based effect for multicolored objects when surface changes occurred at part boundaries. These findings suggest that attention can only select nonuniform objects when the surface change occurs at a part boundary.

Visual scenes are rich with information, but elementary visual information, such as local luminance values, do not appear to be perceived directly. Instead, elementary information is organized into perceptual groups (Palmer, 1999, 2002). Typical visual scenes will contain many perceptual groups, including objects and the regions that fall behind objects. To further manage the wealth of visual information in any scene, perceptual groups can guide attention in an object-based manner, allowing relevant objects to be selectively attended.

Several experimental designs have been used to examine object-based attention. Accuracy in feature reports (e.g., Duncan, 1984; Vecera & Farah, 1994) and speeded spatial cuing tasks (e.g., Egly, Driver, & Rafal, 1994; Vecera, 1994) are among the most common paradigms. For example, Egly et al. used a spatial cuing task in which observers viewed displays containing two simple objects, similar to those shown in Figure 1. Attention was summoned to a cued location at the end of one of two rectangles, and observers detected the onset of a target Targets appeared at validly or invalidly cued locations; when targets were invalidly cued, they could appear in the cued object or in the uncued object. These two invalid locations were equidistant from the cued location, preventing any explanation of the results using location-based attention. Results from this task typically find that observers are fastest to respond to targets at the cued location and, more importantly, are faster to uncued targets in the cued object than those in the uncued object This latter result defines an object-based attention effect.

Although there have been many demonstrations of object-based attention, few studies have attempted to determine the perceptual properties that define the objects selected by object-based attention. One relevant study by Watson and Kramer (1999) examined the perceptual grouping cues to which object-based attention was sensitive. Specifically, using a theory of perceptual organization proposed by Palmer and Rock (1994; Palmer, 1999, 2002), Watson and Kramer focused on the "uniform connectedness" cue, which states that uniform visual surfaces (i.e., surfaces with uniform color, luminance, or texture) are the foundation for perceptual organization. Object-based attention is also sensitive to uniform connectedness: An object-based effect was found for uniform objects that were composed of a single color and luminance but not for nonuniform objects that were composed of regions of different color and luminance. Thus, the objects selected by attention must have uniform surfaces, unless the nonuniform surfaces are task relevant; see Experiment 2 in Watson & Kramer. This surface uniformity likely involves postconstancy reflectance of the surface (Rock, Nijhawan, Palmer, & Tudor, 1992).

The finding that object-based attention typically selects only single-region objects may limit the generalizability of object-based attention because real world objects typically contain several regions (see Matsukura & Vecera, 2006, for discussion; also see Martin, Fowlkes, & Malik, 2004 for evidence from computer vision concerning image regions). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.