What image features guide lightness perception?

Minjung Kim,Richard F Murray,Jason M Gold

doi:10.1167/18.13.1

Minjung Kim, Richard F Murray + Show 1 more

Open Access

https://doi.org/10.1167/18.13.1

Copy DOI

Abstract

Lightness constancy is the ability to perceive black and white surface colors under a wide range of lighting conditions. This fundamental visual ability is not well understood, and current theories differ greatly on what image features are important for lightness perception. Here we measured classification images for human observers and four models of lightness perception to determine which image regions influenced lightness judgments. The models were a high-pass-filter model, an oriented difference-of-Gaussians model, an anchoring model, and an atmospheric-link-function model. Human and model observers viewed three variants of the argyle illusion (Adelson, 1993) and judged which of two test patches appeared lighter. Classification images showed that human lightness judgments were based on local, anisotropic stimulus regions that were bounded by regions of uniform lighting. The atmospheric-link-function and anchoring models predicted the lightness illusion perceived by human observers, but the high-pass-filter and oriented-difference-of-Gaussians models did not. Furthermore, all four models produced classification images that were qualitatively different from those of human observers, meaning that the model lightness judgments were guided by different image regions than human lightness judgments. These experiments provide a new test of models of lightness perception, and show that human observers' lightness computations can be highly local, as in low-level models, and nevertheless depend strongly on lighting boundaries, as suggested by midlevel models.

Highlights

Lightness constancy is the remarkable ability of the human visual system to maintain a stable percept of surface reflectance across a wide range of lighting conditions
Our results show that human observers’ lightness judgments depend strongly on the luminances in the immediate neighborhood of the test patches being judged, in a way that tracks the boundaries of local lighting frameworks
In the Modeling section, we examine classification images from four current models of lightness perception to see how well they account for the local, anisotropic, contrast-like effects we found in human classification images

Summary

Introduction

Lightness constancy is the remarkable ability of the human visual system to maintain a stable percept of surface reflectance across a wide range of lighting conditions. We chose the argyle illusion as our ‘‘fruit fly’’ because it is one of the strongest known lightness illusions, and one which has consistently resisted explanations by low-level models (e.g., Blakeslee & McCourt, 2012) It poses a difficult and interesting problem for modeling visual perception, and understanding it may reveal general principles of lightness constancy. None of the models that we tested are able to both replicate the argyle illusion and produce classification images that are even qualitatively similar to those from human observers These findings show that making progress with computational models of lightness perception will require a better understanding of how lighting frameworks are established and of how the luminance of elements within lighting frameworks contributes to lightness percepts. We measured classification images to determine what parts of the image contributed most strongly to the illusion

Methods

Procedure

Results and discussion