Structural Teacher-Student Normality Learning for Multi-Class Anomaly Detection and Localization

TL;DR

This paper introduces a novel structural teacher-student learning framework for multi-class anomaly detection and localization, effectively addressing cross-class interference and outperforming existing methods on benchmark datasets.

Contribution

The paper proposes a new Structural Teacher-Student Normality Learning approach with spatial-channel and affinity distillation, plus a residual aggregation module, to improve multi-class anomaly detection.

Findings

Outperforms state-of-the-art distillation methods by 3.9% on MVTecAD

Achieves 1.2% improvement on VisA in multi-class detection

Surpasses current unified models on both datasets

Abstract

Visual anomaly detection is a challenging open-set task aimed at identifying unknown anomalous patterns while modeling normal data. The knowledge distillation paradigm has shown remarkable performance in one-class anomaly detection by leveraging teacher-student network feature comparisons. However, extending this paradigm to multi-class anomaly detection introduces novel scalability challenges. In this study, we address the significant performance degradation observed in previous teacher-student models when applied to multi-class anomaly detection, which we identify as resulting from cross-class interference. To tackle this issue, we introduce a novel approach known as Structural Teacher-Student Normality Learning (SNL): (1) We propose spatial-channel distillation and intra-&inter-affinity distillation techniques to measure structural distance between the teacher and student networks. (2) We introduce a central residual aggregation module (CRAM) to encapsulate the normal representation space of the student network. We evaluate our proposed approach on two anomaly detection datasets, MVTecAD and VisA. Our method surpasses the state-of-the-art distillation-based algorithms by a significant margin of 3.9% and 1.5% on MVTecAD and 1.2% and 2.5% on VisA in the multi-class anomaly detection and localization tasks, respectively. Furthermore, our algorithm outperforms the current state-of-the-art unified models on both MVTecAD and VisA.