OpenCodePapers

multi-class-anomaly-detection-on-real-iad-unified-multi-class

Anomaly DetectionMulti-class Anomaly Detection

Results over time

Click legend items to toggle metrics. Hover points for model names.

Leaderboard

Paper	Code	Detection AUROC	Segmentation AP	Detection F1-max	Segmentation AUROC	Segmentation F1-max	Segmentation AUPRO@30% FPR	Detection AP	ModelName	ReleaseDate
Dinomaly2: A Unified Framework for Unsupervised Image Anomaly Detection	✓ Link	92.1	48.4	83.4	99.2	51.8	96.2	90.2	Dinomaly2-L	2025-10-20
MATCH: Flow Matching for Multi-View Anomaly Detection	✓ Link	91.17			99.24		94.76		MATCH	2026-06-23
HLRAD: High-dimensional Latent Representation for Unified Anomaly Detection		90.1	47.7	81.1		50.5			HLRAD	2026-09-08
Dinomaly: The Less Is More Philosophy in Multi-Class Unsupervised Anomaly Detection	✓ Link	90.07	44.29	80.90	99.02	48.36	94.37	87.57	Dinomaly (DINOv2-R ViT-Large/14) (extra data)	2024-05-23
Dinomaly: The Less Is More Philosophy in Multi-Class Unsupervised Anomaly Detection	✓ Link	90.07	44.29	80.90	99.02	48.36	94.37	87.57	Dinomaly ViT-Large	2024-05-23
IMMoE: Incomplete Multi-View Anomaly Detection via Mixture of View Experts Fusion	✓ Link	89.7	44.3						IMMoE	2026-07-21
MambaAD: Exploring State Space Models for Multi-class Unsupervised Anomaly Detection	✓ Link	86.3	33.0	77.0	98.5	38.7	90.5	84.6	MambaAD (ResNet34 + Mamba [3,4,6,3]) (extra data)	2024-04-09
MambaAD: Exploring State Space Models for Multi-class Unsupervised Anomaly Detection	✓ Link	86.3	33.0	77.0	98.5	38.7	90.5	84.6	MambaAD	2024-04-09