MétaCan
Menu
Back to cohort
Record W2569333680 · doi:10.1167/16.12.403

The role of category-specific global orientation statistics for scene categorization

2016· article· en· W2569333680 on OpenAlex

Why this work is in the frame

A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.

affAt least one author lists a Canadian institution in the pinned OpenAlex snapshot.

Bibliographic record

VenueJournal of Vision · 2016
Typearticle
Languageen
FieldEngineering
TopicRemote-Sensing Image Classification
Canadian institutionsUniversity of Toronto
Fundersnot available
KeywordsCategorizationArtificial intelligenceOrientation (vector space)NaturalnessNormalization (sociology)Coherence (philosophical gambling strategy)Scene statisticsPattern recognition (psychology)MathematicsComputer visionComputer scienceRotation (mathematics)StatisticsPerceptionPsychologyGeometry

Abstract

fetched live from OpenAlex

Real-world scenes contain category-specific regularities in global orientations that correlate well with broad descriptions of scene content, such as naturalness and openness. Here we test the role of global orientation statistics for scene categorization behavior, using line drawings and photographs of real-world scenes. To this end, we selectively disrupted global orientation distributions or local edge coherence and briefly presented the modified scenes to observers, who were asked to categorize each image as beach, city street, forest, highway, mountain, or office. In Experiment 1, we disrupted global orientations of line drawings by random image rotation, local edge coherence by random contour-shifting, or both. We found that contour-shifting impaired categorization accuracy significantly more than image rotation. When line drawings were under both manipulations, scene categorization was the least accurate. These findings suggest that contour orientation contributes to accurate scene categorization, although less so than local edge coherence. In Experiment 2, we normalized the spectral amplitude of grayscale photographs either within a category to preserve category-specific global orientation distributions or across all six scene categories to remove them. We found that category-specific mean amplitude significantly improved participants' categorization accuracy. How does the distribution of category-specific global orientation affect representational structure of scene categories? Across the two experiments, we compared error patterns for manipulated images with those for the intact conditions. The results showed that the error patterns for the images with global orientations disrupted (image rotation, amplitude normalization within or across categories) showed significant correlation with error patterns for intact images. On the other hand, when localized edge coherence was disrupted (contour shifting with or without image rotation), error patterns did not match those for intact images. We conclude that category-specific global orientation distribution aids in accurate scene categorization, but that it has no impact on the categorical representations underlying human scene perception. Meeting abstract presented at VSS 2016

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.

Full frame distilled prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.

metaresearch head score (Codex)0.000
metaresearch head score (Gemma)0.000
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Bench or experimental · Consensus signal: none
GenreCandidate signal: Empirical · Consensus signal: none
Teacher disagreement score0.936
Threshold uncertainty score0.164

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0000.000
Meta-epidemiology (narrow)0.0000.000
Meta-epidemiology (broad)0.0000.000
Bibliometrics0.0000.000
Science and technology studies0.0000.000
Scholarly communication0.0000.000
Open science0.0000.000
Research integrity0.0000.000
Insufficient payload (model declined to judge)0.0000.000

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.

Opus teacher head0.010
GPT teacher head0.260
Teacher spread0.250 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it