Distribution Prototype Diffusion Learning for Open-set Supervised Anomaly Detection

TL;DR

The paper introduces Distribution Prototype Diffusion Learning (DPDL), a novel method for open-set supervised anomaly detection that uses Gaussian prototypes and Schrödinger bridges to improve normal sample representation and out-of-distribution anomaly identification.

Contribution

It proposes a new diffusion-based approach with learnable prototypes and dispersion features to enhance anomaly detection performance in open-set scenarios.

Findings

Achieves state-of-the-art results on 9 public datasets.

Effectively encloses normal samples within a discriminative distribution space.

Improves out-of-distribution anomaly detection through hyperspherical dispersion features.

Abstract

In Open-set Supervised Anomaly Detection (OSAD), the existing methods typically generate pseudo anomalies to compensate for the scarcity of observed anomaly samples, while overlooking critical priors of normal samples, leading to less effective discriminative boundaries. To address this issue, we propose a Distribution Prototype Diffusion Learning (DPDL) method aimed at enclosing normal samples within a compact and discriminative distribution space. Specifically, we construct multiple learnable Gaussian prototypes to create a latent representation space for abundant and diverse normal samples and learn a Schr\"odinger bridge to facilitate a diffusive transition toward these prototypes for normal samples while steering anomaly samples away. Moreover, to enhance inter-sample separation, we design a dispersion feature learning way in hyperspherical space, which benefits the identification of out-of-distribution anomalies. Experimental results demonstrate the effectiveness and superiority of our proposed DPDL, achieving state-of-the-art performance on 9 public datasets.