Article

A deep convolutional neural network trained for lightness constancy is susceptible to lightness illusions

2025-11-12

Abstract excerpt

Human viewers are able to perform tasks that depend on accurate estimates of surface reflectance, even across large changes in illumination and context. This is a remarkable ability, and successful image-computable models of how the visual system achieves this have remained elusive. Recently, deep convolutional neural networks (CNNs) have been developed that are adept at estimating surface reflectance. Here we eva...

Topics

Open a Topic to create a Post that cites this publication.

Identifiers and source

Literature Corpus work: 0e1c0568-4293-58b8-8c11-91440a9507f0
DOI: 10.1101/2025.11.10.687742

Open publication

Related research

Semantic proximity does not establish scientific evidence.

Click a neighbor to travelStep 1 · 12 closest

A deep convolutional neural network trained for lightness constancy is susceptible to lightness illusionsDOI 10.1101/2025.11.10.687742Read

Empirical Evaluation of Computational Models of Lightness PerceptionSemantic proximity Read

Human gloss perception reproduced by tiny neural networksSemantic proximity Read

Computational Luminance Constancy from Naturalistic ImagesSemantic proximity Read

Unsupervised learning predicts human perception and misperception of glossSemantic proximity Read

Color constancy based on the geometry of color distributionSemantic proximity Read

Characterization of human lightness discrimination thresholds for independent spectral variationsSemantic proximity Read

Modeling Metallicity: Low Level Visual Features Support Robust Material PerceptionSemantic proximity Read

Human color constancy in cast shadowsSemantic proximity Read

Optimal colors can predict luminosity thresholds in natural scenesSemantic proximity Read

Simple assumptions to improve Markov illuminance and reflectanceSemantic proximity Read

Simple assumptions to improve Markov illuminance and reflectanceSemantic proximity Read

A parametric texture model based on deep convolutional features closely matches texture appearance for humansSemantic proximity Read