August 2016
Volume 16, Issue 12
Open Access
Vision Sciences Society Annual Meeting Abstract  |   September 2016
How Natural Distributions of Blur Affect 3D Percepts
Author Affiliations
  • Martin Banks
    Vision Science Program, School of Optometry, UC Berkeley
  • William Sprague
    Vision Science Program, School of Optometry, UC Berkeley
  • Emily Cooper
    Department of Psychology, Dartmouth College
  • Sylvain Reissier
    Vision Science Program, School of Optometry, UC Berkeley
Journal of Vision September 2016, Vol.16, 195. doi:10.1167/16.12.195
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Martin Banks, William Sprague, Emily Cooper, Sylvain Reissier; How Natural Distributions of Blur Affect 3D Percepts. Journal of Vision 2016;16(12):195. doi: 10.1167/16.12.195.

      Download citation file:


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

      ×
  • Supplements
Abstract

We asked whether natural distributions of blur in different parts of the visual field influence 3D percepts. With a mobile eye-and-scene tracker, we measured distances in the scene, fixations, and pupil diameters as people engaged in everyday activities. From those measurements, we calculated point-spread functions for all positions in the central 20 degrees of the visual field across all frames. The distributions of blur varied from task to task, but a weighted combination across tasks revealed consistent patterns. As shown in Figure 1A, there is a consistent vertical gradient of signed blur: Positions below the fovea are generally blurred because most scene points are nearer than fixation (orange); points above the fovea are also blurred because most scene points are farther than fixation (purple). There is no consistent horizontal gradient. These data manifest the prior distribution of blur across the visual field. We next asked whether this distribution affects 3D percepts. We presented random textures with different blur gradients (Figure 1B). When no gradient was present (i.e., the texture was equally sharp everywhere), subjects perceived a frontoparallel plane. When a vertical blur gradient was present, subjects generally perceived a surface slanted top-back even though that stimulus is also geometrically consistent with top-forward slant and with convex and concave wedges. When a horizontal gradient was present, subjects perceived surfaces slanted left-side forward and right-side forward with equally probability. These data are quite consistent with the natural distributions of blur across the visual field. They show that humans use this prior blur distribution to infer 3D shape.

Meeting abstract presented at VSS 2016

×
×

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.

×