Grouping and Gestalts.

Our visual worlds are highly complex. Elemental visual input (e.g., color, form, orientation, movement) begets features (as coded by the perceiver), which assume any variety of combinations to form any number of “parts” (bounded spatial segments) which themselves comprise objects or “things”, and we are exquisitely capable of compiling, translating/processing, and utilizing all of this information to our benefit. While we have had much success in identifying the processes by which we perceive basic visual elements, the processes concerning the grouping of features into objects, and the role that grouping plays in visual search, are still subject to speculation.

To the Gestalt psychologists, the resolution of a certain combination of features – a “thing” – into a fundamentally different category than its individual features is a core function of object identification. That is, features combine non-additively, and the resulting conglomerate has a set of visual properties that is different from, not necessarily greater than, the sum of the properties of the individual features (Pomerantz & Kubovy, 1981; Wertheimer, 1912). These qualitatively different properties are novel and cannot be predicted by the original features from which they are comprised.

Treisman and Gelade’s (1980) Feature Integration Theory (FIT), a leading influence in theories of visual attention, proposes that the visual field is composed of basic features that are processed quickly and parallel, and that it is not until we attend to the combination of those features that we are able to spatially bind them into a single, recognizable object. Treisman and Paterson (1984) added to FIT by recognizing that objects retained different features than their individual features, and called these newly-born features emergent features (EFs) a term previously coined by Pomerantz et al (1977). These novel features emerging from the object seem to be processed in the same, parallel nature as basic features. In fact, they are more prominent and salient than the features of the object’s parts (Pomerantz et al., 1977).

In further studies, Pomerantz (2001) considered that grouping is simply the appearance of EFs, in which certain configurations of dots or lines leads to a configural superiority effect (CSE), making those configurations more salient than others in the visual field. In the case of dots, configurations yielding CSEs and therefore EFs include proximity, orientation, symmetry, linearity, and surroundedness (Portillo & Pomerantz, 2005).

Does grouping always help?

Until recently, grouping and CSEs have been observed for their facilitative nature. However, Pomerantz et al (1977) identified configural inferiority effects when certain contexts were added to initial stimuli such that performance was disrupted by the addition of homogenous context to the initial display.Along with configural inferiority effects, other patterns of error in visual search tasks have warranted further observation into the possible interference of pop-out (resulting from grouping) in detecting targets.Pomerantz and Portillo (2004) first deemed one of these phenomenons False Pop-Out (FPO), and described it simply as the “tendency of a target to pose as a distracter”.

The goal of my current research is to determine definitively why this phenomenon occurs. Particularly, I am asking the questions “Why does FPO occur (i.e., what conditions must be met before we expect FPO to occur), is it possible to quantitatively define FPO, and at what level of stimulus complexity, if any, does FPO stop occurring?”