RAID: Retrieval-Augmented Anomaly Detection

TL;DR

RAID introduces a retrieval-augmented framework for unsupervised anomaly detection that leverages retrieved normal samples to improve noise resilience and localization accuracy across multiple benchmarks.

Contribution

This work reinterprets UAD through the RAG paradigm, proposing a hierarchical retrieval and guided MoE network to enhance anomaly detection and localization.

Findings

Achieves state-of-the-art results on multiple benchmarks.

Effective noise suppression using retrieved normal samples.

Robust performance in full-shot, few-shot, and multi-dataset settings.

Abstract

Unsupervised Anomaly Detection (UAD) aims to identify abnormal regions by establishing correspondences between test images and normal templates. Existing methods primarily rely on image reconstruction or template retrieval but face a fundamental challenge: matching between test images and normal templates inevitably introduces noise due to intra-class variations, imperfect correspondences, and limited templates. Observing that Retrieval-Augmented Generation (RAG) leverages retrieved samples directly in the generation process, we reinterpret UAD through this lens and introduce \textbf{RAID}, a retrieval-augmented UAD framework designed for noise-resilient anomaly detection and localization. Unlike standard RAG that enriches context or knowledge, we focus on using retrieved normal samples to guide noise suppression in anomaly map generation. RAID retrieves class-, semantic-, and instance-level representations from a hierarchical vector database, forming a coarse-to-fine pipeline. A matching cost volume correlates the input with retrieved exemplars, followed by a guided Mixture-of-Experts (MoE) network that leverages the retrieved samples to adaptively suppress matching noise and produce fine-grained anomaly maps. RAID achieves state-of-the-art performance across full-shot, few-shot, and multi-dataset settings on MVTec, VisA, MPDD, and BTAD benchmarks. \href{https://github.com/Mingxiu-Cai/RAID}{https://github.com/Mingxiu-Cai/RAID}.