The role of local and global clutter in visual search

A Model of Clutter for Complex, Multivariate Geospatial Displays

“…In contrast, the aeronautical charts in the second experiment comprised a much greater diversity of cartographic feature types, which were more difficult to quantify into a single feature set size. We suggest that in the case of these more complex images, in which features are not easily "countable" and set size is not obvious, clutter models such as C3 and Rosenholtz's model may be better predictors of subjective clutter (Beck et al, 2008).…”

“…The correlation between C3 and ratings was higher in the second data set, in which feature set size was irrelevant. We suggest that for more complex displays, in which set size is difficult to determine, C3 may be a better predictor of subjective perceived clutter (Beck et al, 2008).…”

“…As soon as a response was given, the next graphic appeared on the participant's display. Participants first completed a target search task on each graphic, presented in random order, results of which are published in Beck, Lohrenz, Trafton, and Gendron (2008). Following this task (which took an average of 17 min per participant), participants took a 5-min break (on average) and then were shown the same graphics in a different random order and asked to rate how cluttered they thought each appeared, from 0 (not cluttered) to 9 (extremely cluttered).…”

“…The ability to predict subjective impressions of display clutter is a good first step in determining whether clutter detracts from a display's usefulness, but the ultimate test is comparing the clutter model with target search performance. Another set of experiments completed by the authors demonstrated that C3 is also a good predictor of human response time when locating targets in displays of varying clutter (Beck et al, 2008).…”

Evidence of Clutter Avoidance in Complex Scenes

“…Several measures of display clutter have recently been introduced which correlate highly with subjective ratings of clutter (e.g., Lohrenz et al, 2009;Beck et al, 2008;Rosenholtz et al, 2005 and. Further, recent results from Beck et al (in press) demonstrated that at least two interacting measures of clutter are needed to adequately model limitations on attention during visual search of complex displays: global clutter (the average clutter of the entire display) and local clutter (the clutter immediately surrounding some target or feature of interest).…”