October 2020
Volume 20, Issue 11
Open Access
Vision Sciences Society Annual Meeting Abstract  |   October 2020
Modeling biases of perceived slant in curved surfaces
Author Affiliations & Notes
  • Jonathan Tong
    York University
  • Robert Allison
    York University
  • Laurie Wilcox
    York University
  • Footnotes
    Acknowledgements  This work was funded by an NSERC Collaborative Research & Development (CRD) grant, in partnership with Qualcomm Canada Inc, and the Canada First Research Excellence Fund (CFREF) for the Vision: Science to Application (VISTA) program.
Journal of Vision October 2020, Vol.20, 561. doi:https://doi.org/10.1167/jov.20.11.561
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Jonathan Tong, Robert Allison, Laurie Wilcox; Modeling biases of perceived slant in curved surfaces. Journal of Vision 2020;20(11):561. https://doi.org/10.1167/jov.20.11.561.

      Download citation file:

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

  • Supplements

Veridical perception of surface slant is important to everyday tasks such as traversing terrain and interacting with or placing objects on surfaces. However, natural surfaces contain higher-order depth variation, or curvature, which may impact how slant is perceived. We propose a computational model which predicts that curvature, real or distortion-induced, biases the perception of surface slant. The model is based on the perspective projection of surfaces to form “retinal images” containing monocular and binocular texture cues (gradients) for slant estimation. Curvature was either intrinsic to the modelled surface or induced by non-uniform magnification i.e. radial distortion (typical in wide-angle lenses and head-mounted display optics). The resulting binocular and monocular texture gradients derived from these conditions make specific predictions regarding perceived surface slant. In a series of psychophysical experiments we tested these predictions using slant discrimination and magnitude estimation tasks. Our results confirm that local slant estimation is biased in a manner consistent with apparent surface curvature. Further we show that for concave surfaces, irrespective of whether curvature is intrinsic or distortion-induced, there is a net underestimation of global surface slant. Somewhat surprisingly, we also find that the observed biases in global slant are driven largely by the texture gradients and not by the concurrent changes in binocular disparity. This is due to vertical asymmetry in texture gradients of curved surfaces with overall slant. Our results show that while there is a potentially complex interaction between surface curvature and slant perception, much of the perceptual data can be predicted by a relatively simple model based on perspective projection. The work highlights the importance of evaluating the impact of higher-order variations on perceived surface attitude, particularly in virtual environments in which curvature may be intrinsic or caused by optical distortion.


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.