ToCoAD: Two-Stage Contrastive Learning for Industrial Anomaly Detection
This work addresses the domain gap issue in industrial anomaly detection, offering a method that enhances performance on specific datasets, but it appears incremental as it builds on existing contrastive learning and self-supervised techniques.
The paper tackles the problem of unsupervised anomaly detection in industrial settings struggling with domain gaps by proposing ToCoAD, a two-stage contrastive learning method that uses synthetic anomalies and bootstrap contrastive learning to improve generalizability, achieving pixel-level AUROC scores of 98.21%, 98.43%, and 97.70% on MVTec AD, VisA, and BTAD datasets.
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.