September 2024
Volume 24, Issue 10
Open Access
Vision Sciences Society Annual Meeting Abstract  |   September 2024
Deep learning models for lightness constancy can exploit both natural lighting cues and rendering artifacts.
Author Affiliations & Notes
  • Alban Flachot
    York University
  • Jaykishan Patel
    York University
  • Marcus A. Brubaker
    York University
  • Thomas S.A. Wallis
    TU Darmstadt
  • David H. Brainard
    University of Pennsylvania
  • Richard F. Murray
    York University
  • Footnotes
    Acknowledgements  Funded by a VISTA postdoctoral fellowship to Alban Flachot
Journal of Vision September 2024, Vol.24, 1463. doi:https://doi.org/10.1167/jov.24.10.1463
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Alban Flachot, Jaykishan Patel, Marcus A. Brubaker, Thomas S.A. Wallis, David H. Brainard, Richard F. Murray; Deep learning models for lightness constancy can exploit both natural lighting cues and rendering artifacts.. Journal of Vision 2024;24(10):1463. https://doi.org/10.1167/jov.24.10.1463.

      Download citation file:


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

      ×
  • Supplements
Abstract

We previously showed that deep learning models that estimate intrinsic image components (albedo and illuminance) outperform classic models on lightness constancy tasks. Here, we examine what cues these models rely on. We considered two cue types: natural features such as shadows and shading, and artifacts of ray tracing softwares, which typically produce a residual rendering noise that varies with local illumination. We rendered training, validation and test sets via (1) ray tracing (Blender/Cycles) with 128 photons sampled per pixel (high residual noise); (2) same as (1) but 1024 photons sampled (low noise); (3) Blender’s EEVEE renderer (rasterization engine, no noise). (Noise artifacts are also found in other ray tracing renderers, including Mitsuba.) Networks trained on EEVEE images showed similar performance on all three test sets (and performed much better than classic models), whereas networks trained on Cycles showed best performance with Cycles test images, and worst performance with EEVEE images. To assess dependence on naturalistic cues, we tested the networks on test images with various scene elements removed: (1) cast shadows on the floor; (2) shading; (3) all shadows and shading. In (3), no naturalistic lighting cues were available, and yet models trained on Cycles keep a partial, if low, constancy. These models were also almost unaffected by the removal of shadows and shading (less than 10% decrease in constancy). However, the model trained on EEVEE showed a 50% decrease in constancy when floor shadows were removed, and had lowest constancy in condition (3). These results show that widely used ray tracing methods typically produce artifacts that networks can exploit to achieve lightness constancy. When these artifacts are avoided, networks rely on more naturalistic lighting cues, and still exhibit human levels of constancy. Thus deep networks provide a promising starting point for image-computable models of human lightness and color perception.

×
×

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.

×