August 2023
Volume 23, Issue 9
Open Access
Vision Sciences Society Annual Meeting Abstract  |   August 2023
Free and fast implementations of a novel lightness computational model
Author Affiliations & Notes
  • Yuki Kobayashi
    Ritsumeikan University
    Japan Society for the Promotion of Science
    American University
  • Masanori Kanamaru
    Japan Society for the Promotion of Science
    The University of Tokyo
  • Akiyoshi Kitaoka
    Ritsumeikan University
  • Footnotes
    Acknowledgements  The first author received financial support from the Japan Society for the Promotion of Science (Grant Nos. 20J00606 and 22K13878).
Journal of Vision August 2023, Vol.23, 4951. doi:https://doi.org/10.1167/jov.23.9.4951
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Yuki Kobayashi, Masanori Kanamaru, Akiyoshi Kitaoka; Free and fast implementations of a novel lightness computational model. Journal of Vision 2023;23(9):4951. https://doi.org/10.1167/jov.23.9.4951.

      Download citation file:


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

      ×
  • Supplements
Abstract

Several researchers in the field of lightness/brightness perception propose computational models that aim to accurately predict human perception. Although models that employ spatial filters is the major approach in this research topic, a novel computational model called Markov illuminance and reflectance (MIR) employs a probabilistic method and demonstrates a high level of performance in predicting lightness illusions. In addition, MIR is extensible; thus, it can be a novel pathway for the pursuit of better models for lightness/brightness. However, its availability is restricted, because the original code of MIR is provided in MATLAB, which is a relatively expensive language. Moreover, processing even extremely small images takes a relatively long time due to the complexity of the computation of MIR. To address these concerns, the current study translates the original code into two free languages, namely, Python and Julia. Python is one of the most popular languages; thus, Python implementation should render MIR open to more people. Alternatively, Julia is currently less popular than Python or MATLAB; however, it is known to run fast with simple coding. The study notes that the Julia implementation ran nearly three times faster than the original MATLAB code in our environment. To improve existing models and approach an ideal one, repeatedly testing models using a large number of images is indispensable, and the proposed models must be testable by anyone. Therefore, the free and fast implementations of MIR presented in the study should be a beneficial tool for constructing a better computational representation of lightness/brightness perception in humans.

×
×

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.

×