Dinomaly: The Less Is More Philosophy in Multi-Class Unsupervised Anomaly Detection

Jia Guo, Shuai Lu, Weihang Zhang, Huiqi Li

Keywords: Unsupervised Anomaly Detection, Unsupervised Learning, Multi-Class UAD, Unified Model

TL;DR: We introduce Dinomaly, a minimalistic UAD method, that can bridge the substantial performance gap between multi-class UAD and class-separated UAD.

Abstract: Recent studies highlighted a practical setting of unsupervised anomaly detection (UAD) that builds a unified model for multi-class images, serving as an alternative to the conventional one-class-one-model setup. Despite various advancements addressing this challenging task, the detection performance under the multi-class setting still lags far behind state-of-the-art class-separated models. Our research aims to bridge this substantial performance gap. In this paper, we introduce a minimalistic reconstruction-based anomaly detection framework, namely Dinomaly, which leverages pure Transformer architectures without relying on complex designs, additional modules, or specialized tricks. Given this powerful framework consisted of only Attentions and MLPs, we found four simple components that are essential to multi-class anomaly detection: (1) _Foundation_ _Transformers_ that extracts universal and discriminative features, (2) _Noisy_ _Bottleneck_ where pre-existing Dropouts do all the noise injection tricks, (3) _Linear_ _Attention_ that naturally cannot focus, and (4) _Loose_ _Reconstruction_ that does not force layer-to-layer and point-by-point reconstruction. Extensive experiments are conducted across three popular anomaly detection benchmarks including MVTec-AD, VisA, and the recently released Real-IAD. Our proposed Dinomaly achieves impressive image AUROC of 99.6\%, 98.7\%, and 89.3\% on the three datasets respectively, which is not only superior to state-of-the-art multi-class UAD methods, but also surpasses the most advanced class-separated UAD records.

Primary Area: Machine vision

Submission Number: 2463