What image features guide lightness perception?
Bibliographic record
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.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.005 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.002 | 0.003 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".