AI Paper Atlas

Research map / Label-efficient and robust learning / OOD and anomaly detection

Triggering Failures: Out-of-Distribution Detection by Learning From Local Adversarial Attacks in Semantic Segmentation

Victor Besnier, Andrei Bursuc, David Picard, Alexandre Briot

ICCV 2021 · Label-efficient and robust learning › OOD and anomaly detection · approach: Spurious-aware OOD detection · 49 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
add local adversarial attack training (ablation)ObsNet · CamVid OODfpr95tpr: 54.2 -> 44.6Improved
add local adversarial attack training (ablation)ObsNet · StreetHazardsfpr95tpr: 50.1 -> 44.7Improved
add local adversarial attack training (ablation)ObsNet · BDD Anomalyfpr95tpr: 62.4 -> 60.3Improved
add skip connections (ablation)ObsNet · CamVid OODfpr95tpr: 81.3 -> 54.2Improved
feed input image to observer (ablation)ObsNet · CamVid OODfpr95tpr: 57.0 -> 54.2Improved
use SegNet-mimicking architecture instead of small CNN (ablation)ObsNet · CamVid OODfpr95tpr: 60.3 -> 54.2Improved
add pretraining (ablation)ObsNet · CamVid OODfpr95tpr: 55.7 -> 54.2Improved
attack random-shape regions instead of all pixels (ablation)LAA · CamVid OODfpr95tpr: 51.9 -> 44.6Improved
attack random-shape regions instead of sparse pixels (ablation)LAA · CamVid OODfpr95tpr: 54.2 -> 44.6Improved
adversarial augmentation applied to main segmentation network (ablation)Bayesian SegNet · CamVid OODMean IoU: 49.6 -> 41.6Hurt
adversarial augmentation applied to main segmentation network (ablation)Bayesian SegNet · StreetHazardsMean IoU: 44.3 -> 37.8Hurt
replace MC Dropout with ObsNet + LAA (ablation)OOD detector · CamVid OODfpr95tpr: 49.3 -> 44.6Improved

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