Abstract:In the industrial scenario, anomaly detection could locate but cannot classify anomalies. To complete their capability, we study to automatically discover and recognize visual classes of industrial anomalies. In terms of multi-class anomaly classification, previous methods cluster anomalies represented by frozen pre-trained models but often fail due to poor discrimination. Novel class discovery (NCD) has the potential to tackle this. However, it struggles with non-prominent and semantically weak anomalies that challenge network learning focus. To address these, we introduce AnomalyNCD, a multi-class anomaly classification framework compatible with existing anomaly detection methods. This framework learns anomaly-specific features and classifies anomalies in a self-supervised manner. Initially, a technique called Main Element Binarization (MEBin) is first designed, which segments primary anomaly regions into masks to alleviate the impact of incorrect detections on learning. Subsequently, we employ mask-guided contrastive representation learning to improve feature discrimination, which focuses network attention on isolated anomalous regions and reduces the confusion of erroneous inputs through re-corrected pseudo labels. Finally, to enable flexible classification at both region and image levels during inference, 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, 12.8% $F_1$ gain, 5.7% NMI gain, and 10.8% ARI gain on MTD. The source code is available at <a class="link-external link-https" href="https://github.com/HUST-SLOW/AnomalyNCD" rel="external noopener nofollow">this https URL</a>.

Robust Anomaly Map Assisted Multiple Defect Detection with Supervised Classification Techniques

From Anomaly Detection to Defect Classification

Anomaly detection for industrial quality assurance: A comparative evaluation of unsupervised deep learning models

Pull & Push: Leveraging Differential Knowledge Distillation for Efficient Unsupervised Anomaly Detection and Localization

Industrial Product Surface Anomaly Detection with Realistic Synthetic Anomalies Based on Defect Map Prediction

Robust Unsupervised Anomaly Detection for Surface Defects Based on Stacked Broad Learning System

Score-Based Anomaly Detection for Smart Manufacturing Systems

DRAEM -- A discriminatively trained reconstruction embedding for surface anomaly detection

Anomaly Detection in the Automotive Stamping Process: An Unsupervised Machine Learning Approach

Uni-3DAD: GAN-Inversion Aided Universal 3D Anomaly Detection on Model-free Products

Autoencoder-Based Visual Anomaly Localization for Manufacturing Quality Control

An Improved Anomaly Detection Model for Automated Inspection of Power Line Insulators

Unsupervised Anomaly Detection for Surface Defects with Dual-Siamese Network

A Survey on Unsupervised Anomaly Detection Algorithms for Industrial Images

AnomalyNCD: Towards Novel Anomaly Class Discovery in Industrial Scenarios

AnomalySeg: Deep Learning-Based Fast Anomaly Segmentation Approach for Surface Defect Detection

Unsupervised Automatic Defect Inspection based on Image Matching and Local One-class Classification

Supervised Anomaly Detection for Complex Industrial Images

Anomaly Detection in Surface Mount Technology Process Using Multi-Modal Data: Poster Abstract

A Survey on Unsupervised Visual Industrial Anomaly Detection Algorithms