AnomalyNCD: Towards Novel Anomaly Class Discovery in Industrial Scenarios

TL;DR

AnomalyNCD introduces a novel multi-class anomaly classification network that effectively detects and classifies anomalies in industrial scenarios by addressing weak semantics and non-prominent anomalies, outperforming state-of-the-art methods.

Contribution

The paper proposes innovative techniques like main element binarization, mask-guided representation learning, and region merging for improved anomaly classification in industrial images.

Findings

Achieves significant performance gains on MVTec AD and MTD datasets.

Outperforms state-of-the-art methods with up to 12.8% F1 score improvement.

Effectively handles weak semantics and non-prominent anomalies.

Abstract

Recently, multi-class anomaly classification has garnered increasing attention. Previous methods directly cluster anomalies but often struggle due to the lack of anomaly-prior knowledge. Acquiring this knowledge faces two issues: the non-prominent and weak-semantics anomalies. In this paper, we propose AnomalyNCD, a multi-class anomaly classification network compatible with different anomaly detection methods. To address the non-prominence of anomalies, we design main element binarization (MEBin) to obtain anomaly-centered images, ensuring anomalies are learned while avoiding the impact of incorrect detections. Next, to learn anomalies with weak semantics, we design mask-guided representation learning, which focuses on isolated anomalies guided by masks and reduces confusion from erroneous inputs through corrected pseudo labels. Finally, to enable flexible classification at both region and image levels, we develop a region merging strategy that determines the overall image category based on the classified anomaly regions. Our method outperforms the state-of-the-art works on the MVTec AD and MTD datasets. Compared with the current methods, AnomalyNCD combined with zero-shot anomaly detection method achieves a 10.8% $F_1$ gain, 8.8% NMI gain, and 9.5% ARI gain on MVTec AD, and 12.8% $F_1$ gain, 5.7% NMI gain, and 10.8% ARI gain on MTD. Code is available at https://github.com/HUST-SLOW/AnomalyNCD.