Academic journal article Perception and Psychophysics

Attentional Set for Axis of Symmetry in Symmetry-Defined Visual Search

Academic journal article Perception and Psychophysics

Attentional Set for Axis of Symmetry in Symmetry-Defined Visual Search

Article excerpt

Olivers and van der Helm (1998) showed that symmetry-defined visual search (for both symmetry and asymmetry) requires selective spatial attention. We hypothesize that an attentional set for the orientation of a symmetry axis also is involved in symmetry-defined visual search. We conducted three symmetry-defined visual search experiments with manipulations of the axis of symmetry orientations, and performance was better when the axis orientations within the search array were uniform, rather than a mixture of two orientations, and the attentional set for the axis orientation could be kept. In addition, search performance when the target was defined by the presence of symmetry was equivalent to that when the target was defined by a difference of symmetry axis orientation. These results suggest that attentional set for axis orientation plays a fundamental role in symmetry-defined visual search.

The human visual system can easily detect visual bilateral (mirror) symmetry. An efficient and robust perception of symmetry has been demonstrated in many experimental studies (e.g., Barlow & Reeves, 1979; Julesz, 1971), and it has been hypothesized that the human visual system is equipped with a mechanism that processes symmetric patterns holistically. Symmetry is present everywhere in everyday scenes, especially in objects (both natural and artificial). The efficient perception of symmetry seems helpful in segmenting and recognizing visual objects (Parovel & Vezzani, 2002; Rock, 1983). But what is the relationship between the perception of symmetry and the visual attention that selects the object? If symmetry is automatically encoded in the early stage of visual processing, it is plausible that the mechanism precedes attentional selection. However, if the mechanism requires complex and mass processing of a symmetric pattern that comprises many distributed visual elements (e.g., dots in a dot pattern), it can be presumed to work at a late stage, after attentional selection.

The visual search experiments of Olivers and van der Helm (1998) provide one clear answer to the question: Symmetry search requires visual spatial attention. Olivers and van der Helm presented several patterns simultaneously and required their observers to search for one symmetric (asymmetric) pattern as a target among asymmetric (symmetric) distractor patterns. According to the feature integration theory of attention (Treisman & Gelade, 1980), if symmetry is a basic visual feature that is automatically encoded, it pops out. However, Olivers and van der Helm found inefficient (i.e., serial) search functions (Neisser, 1963; Treisman & Gelade, 1980): Longer response times were found for larger display sizes. Thus, the detection of symmetry within multiple patterns required attention's focusing on each pattern.

However, one question remained: What attribute was detected in such symmetry-defined visual search? In Olivers and van der Helm's (1998) experiments, the criterion that differentiated targets from distractors was the presence (or absence) of vertical symmetry. Hence, the targets could be found by detection of (1) vertical symmetry, (2) symmetry in general (i.e., coded as the presence of symmetry, irrespective of its axis orientation), or (3) asymmetry per se. A vertically symmetric target among asymmetric distractors can be determined by the presence of any vertical symmetry or symmetry in general or by the absence of asymmetry per se. An asymmetric target among vertically symmetric distractors can be determined by the absence of any vertical symmetry or symmetry in general or by the presence of asymmetry per se. Depending on the observer's attentional set for which kind of attributes should be detected, the performance of symmetry-defined visual search would be altered.

Which of these three attributes influenced the results of Olivers and van der Helm's (1998) symmetry-defined visual search? Even in their conditions of search for asymmetry, it is hard to imagine that the observers detected asymmetry per se, rather than symmetry. …

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.