AI Paper Atlas

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

Anomaly Detection and Generation with Diffusion Models: A Survey

Yang Liu, Jing Liu, Chengfang Li, Rui Xi, Wenchao Li, Liang Cao, Jin Wang, Laurence Tianruo Yang, Junsong Yuan, Wei Zhou

TPAMI 2026 · Label-efficient and robust learning › OOD and anomaly detection · approach: Anomaly detection with contaminated data · 20 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 diffusion-based DMAD instead of RTFM for video anomaly detectionvideo AD · UCSD Ped2AUC: 96.3 -> 99.7Improved
use DMAD instead of RTFM on CUHK Avenuevideo AD · CUHK AvenueAUC: 85.1 -> 92.8Improved
use diffusion-based VADiffusion instead of RTFMvideo AD · ShanghaiTechAUC: 73 -> 71.7Hurt
use SensitiveHUE instead of TranAD for time series ADtime series AD · SWaTF1: 74.16 -> 91.08Improved
use SensitiveHUE instead of TranADtime series AD · MSLpoint-adjusted F1: 94.94 -> 98.42Improved
use SensitiveHUE instead of NPSRtime series AD · WADIF1: 61.31 -> 69.96Improved
report point-adjusted F1 instead of plain F1 for AnomalyTran on SWaTAnomalyTran · SWaTF1 vs F1_PA: 21.43 -> 94.07Mixed
use AutoUAD instead of NeuTraLAD for tabular ADtabular AD · 47 tabular datasets (avg)AUC (%): 71.45 -> 92.68Improved
use DPAD instead of SCADtabular AD · 47 tabular datasets (avg)AUC (%): 69.15 -> 88.05Improved
use specialized DNN instead of AutoUADtabular AD · 47 tabular datasets (avg)AUC (%): 92.68 -> 92.27No clear effect
use GLAD instead of Tebbe et al.image AD · VisAI-AUROC: 96.0 -> 99.5Improved
use Tebbe et al. instead of PatchCoreimage AD · BTADPRO: 92.7 -> 95.2Improved

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