ToCoAD: Two-Stage Contrastive Learning for Industrial Anomaly Detection

TL;DR

This paper introduces ToCoAD, a two-stage contrastive learning approach that improves industrial anomaly detection by effectively learning domain-specific anomaly features, achieving high pixel-level AUROC scores on multiple datasets.

Contribution

The paper proposes a novel two-stage training strategy that enhances anomaly detection by combining synthetic anomaly generation and bootstrap contrastive learning, improving domain adaptation.

Findings

Achieves pixel-level AUROC scores over 97% on multiple datasets.

Outperforms existing unsupervised anomaly detection methods.

Demonstrates strong generalization to various industrial anomalies.

Abstract

Current unsupervised anomaly detection approaches perform well on public datasets but struggle with specific anomaly types due to the domain gap between pre-trained feature extractors and target-specific domains. To tackle this issue, this paper presents a two-stage training strategy, called \textbf{ToCoAD}. In the first stage, a discriminative network is trained by using synthetic anomalies in a self-supervised learning manner. This network is then utilized in the second stage to provide a negative feature guide, aiding in the training of the feature extractor through bootstrap contrastive learning. This approach enables the model to progressively learn the distribution of anomalies specific to industrial datasets, effectively enhancing its generalizability to various types of anomalies. Extensive experiments are conducted to demonstrate the effectiveness of our proposed two-stage training strategy, and our model produces competitive performance, achieving pixel-level AUROC scores of 98.21\%, 98.43\% and 97.70\% on MVTec AD, VisA and BTAD respectively.