July 2013
Volume 13, Issue 9
Vision Sciences Society Annual Meeting Abstract  |   July 2013
Attention and spatial scale selection in scene categorization
Author Affiliations
  • John Brand
    Department of Psychology, Concordia University
  • Aaron Johnson
    Department of Psychology, Concordia University
Journal of Vision July 2013, Vol.13, 431. doi:https://doi.org/10.1167/13.9.431
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      John Brand, Aaron Johnson; Attention and spatial scale selection in scene categorization. Journal of Vision 2013;13(9):431. https://doi.org/10.1167/13.9.431.

      Download citation file:

      © ARVO (1962-2015); The Authors (2016-present)

  • Supplements

Humans can identify a scene’s category within 150 ms. Such efficient categorization has been attributed to the coarse information carried by a scene’s low spatial frequencies (LSFs) as opposed to its high spatial frequencies (HSFs; Schyns & Oliva, 1994). However, scene categorization has been shown to use the spatial scale that optimizes performance, suggesting a role for attention (Oliva & Schyns, 1997). Here, we address attention’s role in scale selection by showing that attention to global or local levels of NAVON stimuli influenced selection of LSFs or HSFs, respectively, in subsequently presented hybrid images (e.g., an image that contains a low-frequency version of one image [a city] and a high-frequency version of another image [a mountain]). Each trial consisted of two displays. The first was a prime display in which participants were instructed to identify a NAVON letter based on its global or local level. The second consisted of a briefly presented hybrid image (30 or 150 ms), in which participants were asked to categorize the hybrid. On 50% of trials, the LSF content of the hybrid corresponded to a natural scene (e.g., beach), whereas on the other 50% of trials the LSF content corresponded to a man-made scene (e.g., city). There were 12 possible hybrid images constructed from four scene categories (2 man-made; 2 natural). We found that when a hybrid image contained a natural and a man-made image, participants preferred to classify the hybrid based on the natural image at short durations, and the man-made image at long durations, irrespective of the NAVON task. However, when the hybrid consisted of two images from the same superordinate category (e.g., both natural), observers preferred to categorize the hybrid based on its LSFs when primed with the global NAVON task, and its HSFs when primed with the local NAVON task.

Meeting abstract presented at VSS 2013


This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.