AI Paper Atlas

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

Natural Synthetic Anomalies for Self-Supervised Anomaly Detection and Localization

Hannah M. Schlüter, Jeremy Tan, Benjamin Hou, Bernhard Kainz

ECCV 2022 · Label-efficient and robust learning › OOD and anomaly detection · approach: Unified multi-class anomaly detection · 261 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 NSA (Poisson blended, resized, shifted patches) instead of CutPaste patchesend-to-end anomaly segmentation model · MVTec AD image-levelAUROC: 84.5 -> 97.2Improved
use logistic labels instead of binary labels (ablation)NSA · MVTec AD image-levelAUROC: 95.9 -> 97.2Improved
use bounded logistic labels instead of unbounded continuous labels (ablation)NSA · MVTec AD image-levelAUROC: 95.2 -> 97.2Improved
use binary labels instead of unbounded continuous labels (ablation)NSA · MVTec AD pixel-levelAUROC: 93.1 -> 96.3Improved
use logistic or binary labels instead of continuous (ablation)NSA · rCXR (female patients)AUROC: 93.0 -> 94.3No clear effect
use NSA instead of PIIanomaly self-supervision · rCXR (male patients)AUROC: 91.7 -> 94.0Improved
shift and resize patches, as in NSA, instead of PII same-location blendinganomaly self-supervision · MVTec AD object classes image-levelAUROC: 87.7 -> 96.5Improved
add random number of patches instead of single patch (ablation)NSA (logistic) · MVTec AD image-levelAUROC: 95.3 -> 97.2Improved
use NSA patch selection instead of CutPaste-style patch selection (ablation)NSA (logistic) · MVTec AD image-levelAUROC: 91.4 -> 97.2Improved
mask patches with random ellipses (ablation)NSA (logistic) · MVTec AD image-levelAUROC: 97.2 -> 97.0No clear effect
add foreground constraints (ablation)NSA (logistic) · MVTec AD capsule image-levelAUROC: 91.7 -> 95.2Improved
use NSA instead of DRAEM without its reconstructive sub-networkanomaly localization model · MVTec AD image-levelAUROC: 93.9 -> 97.2Improved

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