AI Paper Atlas

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

MoECLIP: Patch-Specialized Experts for Zero-shot Anomaly Detection

Jun Yeong Park, JunYoung Seo, Minji Kang, Yu Rang Park

CVPR 2026 · Label-efficient and robust learning › OOD and anomaly detection · approach: Unified multi-class anomaly detection · 0 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 MoE patch-specialized LoRA experts with routing to frozen CLIPCLIP · ZSAD avg image-levelAUROC: 85.5 -> 89.6 vs AA-CLIPImproved
add MoE patch-specialized LoRA expertsCLIP · ZSAD avg pixel-levelAUROC: 93.2 -> 94.3Improved
add FOFS (frozen orthogonal feature separation) (ablation)MoECLIP · 4-dataset averageimage AUROC: 93.9 -> 95.3Improved
add ETF loss (ablation)MoECLIP · 4-dataset averageimage AUROC: 94.1 -> 95.3Improved
add both FOFS and ETF loss over naive MoE (ablation)MoECLIP · 4-dataset averagepixel AUROC: 91.2 -> 92.2Improved
add depth-wise adapter (ablation)MoECLIP · 4-dataset averageAUROC: 93.6 -> 95.3 imageImproved
add PAA (ablation)MoECLIP · ColonDB pixel-levelAUROC: 81.9 -> 85.4Improved
add PAA (ablation)MoECLIP · MVTec-AD pixel-levelAUROC: 92.1 -> 92.5Improved
add FOFS and ETF loss to reduce inter-expert similarity (ablation)MoE experts · MVTec-AD expert cosine similaritycosine similarity: 0.45 -> 0.02 (experts 1-2)Improved
increase number of experts beyond 4 (ablation)MoECLIP · MVTec-ADAUROCNo clear effect
add CLIP adaptation (naive MoE LoRA) over vanilla CLIP (ablation)CLIP · 4-dataset averageAUROC: 40.6 -> 91.2 pixelImproved
use MoECLIP on Liver CT vs Bayes-PFL/AnomalyCLIPMoECLIP · Liver CT image-levelAUROC: 68.2 -> 74.0Improved

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