Awesome-Multi-Setting-UIAD [](https://github.com/sindresorhus/awesome)

November 14, 2025 · View on GitHub

A taxonomy of Unsupervised Industrial Anomaly Detection (UIAD) methods and datasets (updating).

Welcome to follow our papers "A survey on RGB, 3D, and multimodal approaches for unsupervised industrial image anomaly detection".

If you find any errors in our survey and resource repository, or if you have any suggestions, please feel free to contact us via email at: olsunnylo@outlook.com.

News

  • [2025.03.19] 🔥🔥 Accepted by Information Fusion!

Contents

Overview

Roadmap

roadmap.png

Methods

paradigms.png

RGB UIAD

Datasets

DatasetResourceYearTypeTrainTest (good)Test (anomaly)ValTotalClassAnomaly TypeModal Type
MVTec AD
Data2019Real36294671258-53541573RGB
BTAD
Data2021Real1799451290-25403-RGB
MPDD
Data2021Real888176282-13466-RGB
MVTec LOCO-AD
Data2022Real17725759933043644589RGB
VisA
Data2022Real962101200-1082112-RGB
GoodsAD
Data2023Real313613281660-61246-RGB
MSC-AD
-2023Real648021601080-9720125RGB
CID
Data2024Real390033360-429316RGB
Real-IAD
Data2024Real72840078210-151050308RGB
RAD
Data2024Real213731224-15104-RGB
MIAD
Data2023Synthetic700001750017500-105000713RGB
MAD-Sim
Data2023Synthetic42006384951-9789203RGB
DTD-Synthetic
Data2024Synthetic1200357947-250412-RGB

Methods

NameTitlePublicationYearCodeParadigm
USUninformed Students: Student-Teacher Anomaly Detection With Discriminative Latent Embeddings
CVPR2020CodeTeacher-student architecture
MKDMultiresolution knowledge distillation for anomaly detection
CVPR2021CodeTeacher-student architecture
GPGlancing at the patch: Anomaly localization with global and local feature comparison
CVPR2021-Teacher-student architecture
RD4ADAnomaly Detection via Reverse Distillation From One-Class Embedding
CVPR2022CodeTeacher-student architecture
PFMUnsupervised Image Anomaly Detection and Segmentation Based on Pretrained Feature Mapping
TII2023CodeTeacher-student architecture
MemKDRemembering Normality: Memory-guided Knowledge Distillation for Unsupervised Anomaly Detection
ICCV2023CodeTeacher-student architecture
DeSTSegDeSTSeg: Segmentation Guided Denoising Student-Teacher for Anomaly Detection
CVPR2023CodeTeacher-student architecture
EfficientADEfficientAD: Accurate Visual Anomaly Detection at Millisecond-Level Latencies
WACV2024Unofficial CodeTeacher-student architecture
EMMFRKDEnhanced multi-scale features mutual mapping fusion based on reverse knowledge distillation for industrial anomaly detection and localization
TBD2024-Teacher-student architecture
AEKDAEKD: Unsupervised auto-encoder knowledge distillation for industrial anomaly detection
JMS2024-Teacher-student architecture
FCACDLFeature-Constrained and Attention-Conditioned Distillation Learning for Visual Anomaly Detection
ICASSP2024-Teacher-student architecture
DMDDDual-Modeling Decouple Distillation for Unsupervised Anomaly Detection
ACM MM2024-Teacher-student architecture
CutPasteCutPaste: Self-Supervised Learning for Anomaly Detection and Localization
CVPR2021Unofficial CodeOne-class classification
SimpleNetSimpleNet: A Simple Network for Image Anomaly Detection and Localization
CVPR2023CodeOne-class classification
ADShiftAnomaly Detection Under Distribution Shift
ICCV2023CodeOne-class classification
DS2Learning Transferable Representations for Image Anomaly Localization Using Dense Pretraining
WACV2024-One-class classification
GeneralADGeneralAD: Anomaly Detection Across Domains by Attending to Distorted Features
ECCV2024CodeOne-class classification
GLASSA Unified Anomaly Synthesis Strategy with Gradient Ascent for Industrial Anomaly Detection and Localization
ECCV2024CodeOne-class classification
FastFlowFastFlow: Unsupervised Anomaly Detection and Localization via 2D Normalizing Flows
-2021Unofficial CodeDistribution map
DifferNetSame Same but DifferNet: Semi-Supervised Defect Detection With Normalizing Flows
WACV2021CodeDistribution map
CFLOW-ADCFLOW-AD: Real-Time Unsupervised Anomaly Detection With Localization via Conditional Normalizing Flows
WACV2022CodeDistribution map
CS-FlowFully Convolutional Cross-Scale-Flows for Image-Based Defect Detection
WACV2022CodeDistribution map
CDOCollaborative Discrepancy Optimization for Reliable Image Anomaly Localization
TII2023CodeDistribution map
PyramidFlowPyramidFlow: High-Resolution Defect Contrastive Localization Using Pyramid Normalizing Flow
CVPR2023CodeDistribution map
SLADFascinating Supervisory Signals and Where to Find Them: Deep Anomaly Detection with Scale Learning
ICML2023CodeDistribution map
MSFlowMSFlow: Multiscale Flow-Based Framework for Unsupervised Anomaly Detection
TNNLS2024CodeDistribution map
AttentDifferNetAttention Modules Improve Image-Level Anomaly Detection for Industrial Inspection: A DifferNet Case Study
WACV2024CodeDistribution map
PaDiMPaDiM: A Patch Distribution Modeling Framework for Anomaly Detection and Localization
ICPR2021Unofficial CodeMemory bank
PatchCoreTowards Total Recall in Industrial Anomaly Detection
CVPR2022CodeMemory bank
CFACFA: Coupled-Hypersphere-Based Feature Adaptation for Target-Oriented Anomaly Localization
IEEE Access2022CodeMemory bank
DMADDiversity-Measurable Anomaly Detection
CVPR2023CodeMemory bank
PNIPNI : Industrial Anomaly Detection using Position and Neighborhood Information
ICCV2023CodeMemory bank
GraphCorePushing the Limits of Fewshot Anomaly Detection in Industry Vision: Graphcore
ICLR2023-Memory bank
InReaChInter-Realization Channels: Unsupervised Anomaly Detection Beyond One-Class Classification
ICCV2023CodeMemory bank
ReconFAA Reconstruction-Based Feature Adaptation for Anomaly Detection with Self-Supervised Multi-Scale Aggregation
ICASSP2024-Memory bank
ReConPatchReConPatch: Contrastive Patch Representation Learning for Industrial Anomaly Detection
WACV2024Unofficial CodeMemory bank
AE-SSIMImproving Unsupervised Defect Segmentation by Applying Structural Similarity to Autoencoders
-2018Unofficial CodeAutoencoder-based Reconstruction
DFRUnsupervised anomaly segmentation via deep feature reconstruction
Neurocomputing2020CodeAutoencoder-based Reconstruction
DAADDivide-and-Assemble: Learning Block-Wise Memory for Unsupervised Anomaly Detection
ICCV2021-Autoencoder-based Reconstruction
RIADReconstruction by inpainting for visual anomaly detection
PR2021Unofficial CodeAutoencoder-based Reconstruction
DRÆMDRAEM - A Discriminatively Trained Reconstruction Embedding for Surface Anomaly Detection
ICCV2021CodeAutoencoder-based Reconstruction
DSRDSR – A Dual Subspace Re-Projection Network for Surface Anomaly Detection
ECCV2022CodeAutoencoder-based Reconstruction
NSANatural Synthetic Anomalies for Self-Supervised Anomaly Detection and Localization
ECCV2022CodeAutoencoder-based Reconstruction
SSPCABSelf-Supervised Predictive Convolutional Attentive Block for Anomaly Detection
CVPR2022CodeAutoencoder-based Reconstruction
SSMCTBSelf-Supervised Masked Convolutional Transformer Block for Anomaly Detection
TPAMI2024CodeAutoencoder-based Reconstruction
THFRTemplate-guided Hierarchical Feature Restoration for Anomaly Detection
ICCV2023-Autoencoder-based Reconstruction
FastReconFastRecon: Few-shot Industrial Anomaly Detection via Fast Feature Reconstruction
ICCV2023CodeAutoencoder-based Reconstruction
RealNetRealNet: A Feature Selection Network with Realistic Synthetic Anomaly for Anomaly Detection
CVPR2024CodeAutoencoder-based Reconstruction
IFgNetImplicit Foreground-Guided Network for Anomaly Detection and Localization
ICASSP2024CodeAutoencoder-based Reconstruction
LAMPNeural Network Training Strategy To Enhance Anomaly Detection Performance: A Perspective On Reconstruction Loss Amplification
ICASSP2024-Autoencoder-based Reconstruction
PatchAnomalyPatch-Wise Augmentation for Anomaly Detection and Localization
ICASSP2024-Autoencoder-based Reconstruction
MAAEMixed-Attention Auto Encoder for Multi-Class Industrial Anomaly Detection
ICASSP2024-Autoencoder-based Reconstruction
DC-AEDual-Constraint Autoencoder and Adaptive Weighted Similarity Spatial Attention for Unsupervised Anomaly Detection
TII2024-Autoencoder-based Reconstruction
EARVisual defect obfuscation based self-supervised anomaly detection
Scientific Reports2024-Autoencoder-based Reconstruction
FADeRFeature Attenuation of Defective Representation Can Resolve Incomplete Masking on Anomaly Detection
CVPR 2025 workshop2025-Autoencoder-based Reconstruction
SCADNLearning Semantic Context from Normal Samples for Unsupervised Anomaly Detection
AAAI2021CodeGAN-based Reconstruction
OCR-GANOmni-Frequency Channel-Selection Representations for Unsupervised Anomaly Detection
TIP2023CodeGAN-based Reconstruction
MeTALMasked Transformer for Image Anomaly Localization
IJNS2022-Transformer-based Reconstruction
FODFocus the Discrepancy: Intra- and Inter-Correlation Learning for Image Anomaly Detection
ICCV2023CodeTransformer-based Reconstruction
AMI-NetAMI-Net: Adaptive Mask Inpainting Network for Industrial Anomaly Detection and Localization
TASE2024CodeTransformer-based Reconstruction
PNPTPrior Normality Prompt Transformer for Multiclass Industrial Image Anomaly Detection
TII2024-Transformer-based Reconstruction
DDADAnomaly Detection with Conditioned Denoising Diffusion Models
-2023CodeDiffusion-based Reconstruction
DiffADUnsupervised Surface Anomaly Detection with Diffusion Probabilistic Model
ICCV2023-Diffusion-based Reconstruction
RANRemoving Anomalies as Noises for Industrial Defect Localization
ICCV2023-Diffusion-based Reconstruction
TransFusionTransFusion – A Transparency-Based Diffusion Model for Anomaly Detection
ECCV2024CodeDiffusion-based Reconstruction
DiADA Diffusion-Based Framework for Multi-Class Anomaly Detection
AAAI2024CodeDiffusion-based Reconstruction
GLADGLAD: Towards Better Reconstruction with Global and Local Adaptive Diffusion Models for Unsupervised Anomaly Detection
ECCV2024CodeDiffusion-based Reconstruction
AnomalySDAnomalySD: Few-Shot Multi-Class Anomaly Detection with Stable Diffusion Model
-2024-Diffusion-based Reconstruction

3D UIAD

Datasets

DatasetResourceYearTypeTrainTest (good)Test (anomaly)ValTotalClassAnomaly TypeModal Type
Real3D-AD
data2023Real48604602-1254123Point cloud
Anomaly-ShapeNet
data2023Synthetic208780943-1931507Point cloud

Methods

NameTitlePublicationYearCodeParadigm
3D-STAnomaly Detection in 3D Point Clouds Using Deep Geometric Descriptors
WACV2023-Teacher-student architecture
Reg3D-ADReal3D-AD: A Dataset of Point Cloud Anomaly Detection
NeurIPS2024CodeMemory bank
Group3ADTowards High-resolution 3D Anomaly Detection via Group-Level Feature Contrastive Learning
ACM MM2024CodeMemory bank
PointCorePointCore: Efficient Unsupervised Point Cloud Anomaly Detector Using Local-Global Features
-2024-Memory bank
R3D-ADR3D-AD: Reconstruction via Diffusion for 3D Anomaly Detection
ECCV2024-Reconstruction

Multimodal UIAD

Datasets

DatasetResourceYearTypeTrainTest (good)Test (anomaly)ValTotalClassAnomaly TypeModal Type
MVTec 3D-AD
data2021Real265629494829441471041RGB & Point cloud
PD-REAL
data2023Real23993005303003529156RGB & Point cloud
MulSen-AD
data2024Real1391150494-20351514RGB & Infrared & Point cloud
Eyecandies
data2022Synthetic100002250225010001550010-RGB & Depth

Methods

NameTitlePublicationYearCodeParadigm
BTFBack to the Feature: Classical 3D Features Are (Almost) All You Need for 3D Anomaly Detection
CVPR2023Code-
ASTAsymmetric Student-Teacher Networks for Industrial Anomaly Detection
WACV2023CodeTeacher-student architecture
MMRDRethinking Reverse Distillation for Multi-Modal Anomaly Detection
AAAI2024-Teacher-student architecture
M3DMMultimodal Industrial Anomaly Detection via Hybrid Fusion
CVPR2023CodeMemory bank
Shape-GuidedShape-Guided Dual-Memory Learning for 3D Anomaly Detection
ICML2023CodeMemory bank
CPMFComplementary Pseudo Multimodal Feature for Point Cloud Anomaly Detection
Pattern Recognition2024CodeMemory bank
LSFASelf-supervised Feature Adaptation for 3D Industrial Anomaly Detection
ECCV2024CodeMemory bank
ITNMIncremental Template Neighborhood Matching for 3D anomaly detection
Neurocomputing2024-Memory bank
CMDIADIncomplete Multimodal Industrial Anomaly Detection via Cross-Modal Distillation
-2024CodeMemory bank
M3DM-NRM3DM-NR: RGB-3D Noisy-Resistant Industrial Anomaly Detection via Multimodal Denoising
-2024-Memory bank
EasyNetEasyNet: An Easy Network for 3D Industrial Anomaly Detection
ACM MM2023CodeReconstruction
DBRNDual-Branch Reconstruction Network for Industrial Anomaly Detection with RGB-D Data
ISPP2024-Reconstruction
3DSRCheating Depth: Enhancing 3D Surface Anomaly Detection via Depth Simulation
WACV2024CodeReconstruction
CFMMultimodal Industrial Anomaly Detection by Crossmodal Feature Mapping
CVPR2024CodeReconstruction
3DRÆMKeep DRÆMing: Discriminative 3D anomaly detection through anomaly simulation
PRL2024-Reconstruction

Stargazers over time

Stargazers over time

BibTex Citation

If you find this paper and repository useful, please cite our paper:

@article{lin2025survey,
  title={A survey on RGB, 3D, and multimodal approaches for unsupervised industrial image anomaly detection},
  author={Lin, Yuxuan and Chang, Yang and Tong, Xuan and Yu, Jiawen and Liotta, Antonio and Huang, Guofan and Song, Wei and Zeng, Deyu and Wu, Zongze and Wang, Yan and others},
  journal={Information Fusion},
  pages={103139},
  year={2025},
  publisher={Elsevier}
}