Anomaly Detection2019en

MVTec Anomaly Detection Dataset

5,354 high-resolution images across 15 object and texture categories. The gold standard for industrial anomaly detection with pixel-level annotations.

Metrics:auroc, pixel-auroc, pro, ap
Paper / WebsiteDownload
Current State of the Art

SimpleNet

Research

99.6

auroc

auroc Progress Over Time

Showing 4 breakthroughs from Nov 2020 to Mar 2023

97.798.298.899.399.8Nov 2020Aug 2021May 2022Mar 2023aurocDate

Key Milestones

Nov 2020
PaDiM

ResNet-18 backbone. Fast inference.

97.9
Source
Jun 2021
PatchCore

Image-level AUROC. State-of-the-art with WideResNet-50 backbone.

99.1
+1.2%
Source
Nov 2021
FastFlow

2D normalizing flow. Good speed-accuracy tradeoff.

99.4
+0.3%
Source
Mar 2023
SimpleNetCurrent SOTA

State-of-the-art on MVTec AD. Simple architecture.

99.6
+0.2%
Source
Total Improvement
1.7%
Time Span
2y 4m
Breakthroughs
4
Current SOTA
99.6

Top Models Performance Comparison

Top 8 models ranked by auroc

auroc1SimpleNet99.6100.0%2FastFlow99.499.8%3PatchCore99.199.5%4EfficientAD99.199.5%5Reverse Distillation98.598.9%6CFLOW-AD98.398.7%7DRAEM98.098.4%8PaDiM97.998.3%0%25%50%75%100%% of best
Best Score
99.6
Top Model
SimpleNet
Models Compared
8
Score Range
1.7

aurocPrimary

#ModelScorePaper / CodeDate
1
SimpleNetOpen Source
Research		99.6
research-paper
Dec 2025
2
FastFlowOpen Source
Research		99.4
research-paper
Dec 2025
3
PatchCoreOpen Source
Amazon		99.1
research-paper
Dec 2025
4
EfficientADOpen Source
Research		99.1
research-paper
Dec 2025
5
Reverse DistillationOpen Source
Research		98.5
research-paper
Dec 2025
6
CFLOW-ADOpen Source
Research		98.3
research-paper
Dec 2025
7
DRAEMOpen Source
Research		98
research-paper
Dec 2025
8
PaDiMOpen Source
Research		97.9
research-paper
Dec 2025

Other Anomaly Detection Datasets

MVTec 3D-ADVisAKolektorSDD2Weld Defect X-RaySeverstal Steel DefectNEU-DET