August 2014
Volume 14, Issue 10
Free
Vision Sciences Society Annual Meeting Abstract  |   August 2014
Predicting lightness rankings from image statistics of matte and glossy surfaces
Author Affiliations
  • Matteo Toscani
    Department of Psychology, Justus Liebig University Giessen
  • Matteo Valsecchi
    Department of Psychology, Justus Liebig University Giessen
  • Karl Gegenfurtner
    Department of Psychology, Justus Liebig University Giessen
Journal of Vision August 2014, Vol.14, 77. doi:https://doi.org/10.1167/14.10.77
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Matteo Toscani, Matteo Valsecchi, Karl Gegenfurtner; Predicting lightness rankings from image statistics of matte and glossy surfaces. Journal of Vision 2014;14(10):77. https://doi.org/10.1167/14.10.77.

      Download citation file:


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

      ×
  • Supplements
Abstract

Humans are able to estimate the reflective properties of the surface (albedo) of an object despite the large variability in the reflected light due to shading. We investigated which statistics of the luminance distribution of matte and glossy three-dimensional virtual objects are used to estimate albedo. Eight naive observers were asked to sort twelve objects in an achromatic virtual scene in terms of their albedo. The objects were positioned uniformly spaced on a horizontal plane, the scene was illuminated by a light probe captured in a natural scene. We chose twelve different reflectances which allowed observers to rank the objects better than chance but not perfectly. The scenes were rendered using radiance, a physically based rendering software. The twelve reflectance values were assigned randomly to the objects in the scenes. The twelve object placed in each scene were randomly chosen from a pool of twenty four tridimensional models, ranging from simple geometrical shapes to complex real object models. Observers were significantly better in ranking matte objects (82% correct) than glossy ones (72% correct). The physical ranking of matte objects was best predicted by the maximum of the luminance distribution whereas the best predictor for the glossy objects was the mean of the distribution. Similarly, the observers judgments for matte objects were best predicted based on the mean, maximum and quartiles of the distribution whereas for glossy objects the maximum was a poor predictor of the observers' judgments. In summary our data suggest that histogram statistics of the luminance distributions of complex objects can support the recovery of their surfaces albedo, despite the fact that this distributions results from the complex interplay of geometry and the structure of the illuminant.

Meeting abstract presented at VSS 2014

×
×

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.

×