September 2017
Volume 17, Issue 10
Open Access
Vision Sciences Society Annual Meeting Abstract  |   August 2017
Statistical variations in the power spectrum of daylight over a day predict communicative efficiency of color-language
Author Affiliations
  • Sivalogeswaran Ratnasingam
    Laboratory of Sensorimotor Research, National Eye Institute, NIH
  • Edward Gibson
    Department of Brain and Cognitive Sciences, MIT
  • Richard Futrell
    Department of Brain and Cognitive Sciences, MIT
  • Julian Jara-Ettinger
    Department of Brain and Cognitive Sciences, MIT
  • Kyle Mahowald
    Department of Brain and Cognitive Sciences, MIT
  • Leon Bergen
    Department of Brain and Cognitive Sciences, MIT
  • Steven Piantadosi
    Department of Brain and Cognitive Sciences, University of Rochester
  • Bevil Conway
    Laboratory of Sensorimotor Research, National Eye Institute, NIH
Journal of Vision August 2017, Vol.17, 1171. doi:https://doi.org/10.1167/17.10.1171
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Sivalogeswaran Ratnasingam, Edward Gibson, Richard Futrell, Julian Jara-Ettinger, Kyle Mahowald, Leon Bergen, Steven Piantadosi, Bevil Conway; Statistical variations in the power spectrum of daylight over a day predict communicative efficiency of color-language. Journal of Vision 2017;17(10):1171. https://doi.org/10.1167/17.10.1171.

      Download citation file:


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

      ×
  • Supplements
Abstract

All languages appear to distinguish warm colors from cool colors, but the reason underlying this universal color-naming pattern is unknown. Using original data collected in English, Spanish, and Tsimane', along with data from the 110 languages of the World Color Survey, we discovered that all languages convey more information about warm colors (reds, oranges) than cool colors (greens, blues). In this study, we conducted a new simple information theoretic measure of a particular color relative to a set of colors: average surprisal, which allows us to rank the colors for their relative communicative efficiency within a language. We investigate two potential causes for this asymmetry: (1) color statistics of foreground objects versus backgrounds in natural scenes; and (2) statistical variation of daylight spectra at different phase of day. First, using an analysis of 20000 images independently curated for salient objects [1], we discovered that objects have a higher probability of having a warm color than a cool color compared to backgrounds. Second, we discovered that the communication efficiency of a color term is correlated with equivalent corrected color temperature (CCT) of the color of the Munsell chips. Specifically, the color-language communication efficiency decreases with increasing equivalent CCT of the Munsell chips. Further, the CCT of daylight is low during midday and increases during sunrise and sunset [2]. This variation in spectral content of the daylight is largely discounted by the visual system through color constancy operations, although an object's color is most close to the color observed under equal energy light during midday, under low CCT daylight. [1]. Tie Liu, Jian et.al. Learning to Detect A Salient Object. In Proc. IEEE-CVPR, Minnesota, 2007. [2]. Sivalogeswaran Ratnasingam, et.al. Analysis of colour constancy algorithms using the knowledge of variation of correlated colour temperature of daylight with solar elevation. EURASIP, pp 1-13, 2013.

Meeting abstract presented at VSS 2017

×
×

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.

×