Learning Unified Reference Representation for Unsupervised Multi-class Anomaly Detection

TL;DR

This paper introduces RLR, a novel anomaly detection framework using learnable reference representations to explicitly learn normal patterns and prevent shortcut learning, achieving superior results on benchmark datasets.

Contribution

The paper proposes a unified feature reconstruction framework with learnable references and locality constraints to improve multi-class anomaly detection.

Findings

RLR outperforms state-of-the-art methods on MVTec-AD and VisA datasets.

Learnable references effectively prevent shortcut learning in anomaly detection.

Locality constraints enhance the model's ability to capture normal patterns.

Abstract

In the field of multi-class anomaly detection, reconstruction-based methods derived from single-class anomaly detection face the well-known challenge of "learning shortcuts", wherein the model fails to learn the patterns of normal samples as it should, opting instead for shortcuts such as identity mapping or artificial noise elimination. Consequently, the model becomes unable to reconstruct genuine anomalies as normal instances, resulting in a failure of anomaly detection. To counter this issue, we present a novel unified feature reconstruction-based anomaly detection framework termed RLR (Reconstruct features from a Learnable Reference representation). Unlike previous methods, RLR utilizes learnable reference representations to compel the model to learn normal feature patterns explicitly, thereby prevents the model from succumbing to the "learning shortcuts" issue. Additionally, RLR incorporates locality constraints into the learnable reference to facilitate more effective normal pattern capture and utilizes a masked learnable key attention mechanism to enhance robustness. Evaluation of RLR on the 15-category MVTec-AD dataset and the 12-category VisA dataset shows superior performance compared to state-of-the-art methods under the unified setting. The code of RLR will be publicly available.