AI Paper Atlas

Research map / Learning theory and algorithms / Prediction and decision losses

Saliency Benchmarking Made Easy: Separating Models, Maps and Metrics

Matthias Kummerer, Thomas S. A. Wallis, Matthias Bethge

ECCV 2018 · Learning theory and algorithms › Prediction and decision losses · approach: Theory of neural network sensitivity · 122 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use the metric-specific optimal saliency map instead of one saliency map for all metricssaliency models (AIM, BMS, eDN, OpenSALICON, SalGAN, DeepGaze II) · MIT-style fixation data, seven metricsAUC, sAUC, NSS, IG, CC, KL-Div, SIMImproved
evaluate with saliency maps derived from the model density rather than original mapsall six saliency models · seven saliency metricsconsistency of model rankings across metricsImproved
use a map derived for another metric (instead of the matching one) (ablation)saliency models · seven saliency metricsmetric performanceHurt
evaluate log-density maps (prior unification) instead of metric-specific maps (ablation)saliency models · NSS, CC, KL-Div, SIMmetric performance and ranking agreementHurt
use CC saliency map instead of NSS saliency map for AIM (ablation)AIM · NSS metricNSSMixed
use the SIM-specific map instead of the CC (blurred density) map (ablation)saliency models · SIM metricSIM: average difference 0.006No clear effect
add 8-bit binning to the raw density saliency map (ablation)DeepGaze II · AUCAUC: 0.88355 -> 0.87931Hurt
add 8-bit binning to the raw density saliency map (ablation)AIM · AUCAUC: 0.82883 -> 0.82855Hurt
equalize the histogram to be uniform before 8-bit binningDeepGaze II · AUCAUC under 8-bit binning: 0.87931 -> 0.88351Improved
equalize the histogram to be uniform before 8-bit binningOpenSALICON · AUCAUC under 8-bit binning: 0.86200 -> 0.86347Improved
use mean empirical saliency map instead of mean normalized empirical map for CC (ablation)CC saliency map · simulated fixationsCCNo clear effect
compute SIM map with a different number of fixations per sample than the test data (ablation)SIM saliency map · simulated fixationsSIMHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy