December 2022
Volume 22, Issue 14
Open Access
Vision Sciences Society Annual Meeting Abstract  |   December 2022
Complementary methodologies to investigate spatial frequencies in facial expression recognition
Author Affiliations & Notes
  • Isabelle Charbonneau
    Universite du Quebec en Outaouais
  • Joël Guérette
    Universite du Quebec en Outaouais
  • Caroline Blais
    Universite du Quebec en Outaouais
  • Fraser Smith
    School of Psychology, University of East Anglia
  • Daniel Fiset
    Universite du Quebec en Outaouais
  • Footnotes
    Acknowledgements  Natural Sciences and Engineering Research Council of Canada (NSERC)
Journal of Vision December 2022, Vol.22, 4070. doi:
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Isabelle Charbonneau, Joël Guérette, Caroline Blais, Fraser Smith, Daniel Fiset; Complementary methodologies to investigate spatial frequencies in facial expression recognition. Journal of Vision 2022;22(14):4070.

      Download citation file:

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

  • Supplements

Recently, using the SFs Bubbles method, we revealed a broad range of important SFs for pain recognition, starting from low to relatively high SFs (Charbonneau et al., 2021). We also included a more ecological method that simulates the distance of stimuli presentation which revealed that pain recognition is optimal in a short to medium distance (1.2–4.8m) but declines significantly when mid SFs are no longer available. Here, we were interested in the generalization of these results for other basic expressions. Twenty participants took part in a basic emotion categorization task where reduced size images simulating increasing viewing distance were presented (3.26, 1.63, 0.815, 0.41, 0.20, 0.10 degree of visual angle using the Laplacian Pyramid toolbox; Burt & Adelson, 1983). Unbiased hit rates (Wagner, 1993) were then calculated to quantify performance at each distance. Another twenty participants completed an emotion categorization task using the SFs Bubbles method, a data-driven method which randomly samples SFs on each trial. Multiple regression analysis on the SF filters and accuracies across trials were completed and SF peaks were computed by submitting the classification vector to a 50% area SF measure (ASFM). The results show ASFM peaks in the medium SF range for all emotions (anger = 21.33cpf, disgust = 18.67cpf, fear = 20.67cpf, happiness = 10.33cpf, neutral = 15.67cpf, sadness = 18.33cpf) except for surprise (7.67cpf) which falls in the low SF range. As for pain, the results suggest that for basic emotion a large decrease in performance occurs when the MSFs are no longer available (i.e. between 32 and 16cpf and between 16 and 8cpf). Interestingly, surprise is the most recognized emotion at the furthest distance, which is consistent with the use of lower frequencies for this emotion.


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.