Abstract:Anomaly synthesis strategies can effectively enhance unsupervised anomaly detection. However, existing strategies have limitations in the coverage and controllability of anomaly synthesis, particularly for weak defects that are very similar to normal regions. In this paper, we propose Global and Local Anomaly co-Synthesis Strategy (GLASS), a novel unified framework designed to synthesize a broader coverage of anomalies under the manifold and hypersphere distribution constraints of Global Anomaly Synthesis (GAS) at the feature level and Local Anomaly Synthesis (LAS) at the image level. Our method synthesizes near-in-distribution anomalies in a controllable way using Gaussian noise guided by gradient ascent and truncated projection. GLASS achieves state-of-the-art results on the MVTec AD (detection AUROC of 99.9\%), VisA, and MPDD datasets and excels in weak defect detection. The effectiveness and efficiency have been further validated in industrial applications for woven fabric defect detection. The code and dataset are available at: \url{<a class="link-external link-https" href="https://github.com/cqylunlun/GLASS" rel="external noopener nofollow">this https URL</a>}.

What problem does this paper attempt to address?

### What problems does this paper attempt to solve? This paper aims to solve several key challenges in industrial anomaly detection and localization, especially the limitations of existing anomaly synthesis strategies in terms of coverage and controllability, particularly when dealing with weak defects that are very similar to normal regions. Specifically, this research proposes a new unified framework - the Global and Local Anomaly Synergistic Synthesis Strategy (GLASS) to synthesize anomaly samples in a more extensive and controllable way. #### Main problems: 1. **Limitations of existing methods**: - **Image - level anomaly synthesis**: Although it can provide detailed abnormal textures, it lacks diversity and a sense of reality. - **Feature - level anomaly synthesis**: Although it is more efficient, it has deficiencies in synthesis directionality and controllability, especially when dealing with near - distribution anomalies. 2. **Weak defect detection**: - Existing methods have difficulty in effectively detecting weak defects with small areas or low contrast, which are very similar to normal regions. #### Solutions: - **Global Anomaly Synthesis (GAS)**: Use Gaussian - noise - guided gradient ascent and truncated projection at the feature level to synthesize anomaly samples, ensuring that the synthesized anomaly samples are close to the normal sample distribution, thereby improving the ability to detect weak defects. - **Local Anomaly Synthesis (LAS)**: Synthesize anomaly samples through texture overlay at the image level, providing more anomaly diversity. - **Combining GAS and LAS**: By combining the global and local anomaly synthesis strategies, GLASS can cover anomaly samples more extensively at both the feature and image levels and maintain controllability during the synthesis process. #### Experimental results: - GLASS has achieved state - of - the - art performance on multiple datasets such as MVTec AD, VisA, and MPDD, especially performing excellently in weak defect detection. - Its effectiveness and efficiency have been verified through practical industrial applications, such as in fabric defect detection. In summary, this paper solves the limitations of existing anomaly synthesis strategies in terms of coverage and controllability by proposing the GLASS framework and significantly improves the effect of weak defect detection.

A Unified Anomaly Synthesis Strategy with Gradient Ascent for Industrial Anomaly Detection and Localization

AnomalyDiffusion: Few-Shot Anomaly Image Generation with Diffusion Model

AnomalyFactory: Regard Anomaly Generation as Unsupervised Anomaly Localization

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

Unsupervised Anomaly Detection for Surface Defects with Dual-Siamese Network

AnomalyNCD: Towards Novel Anomaly Class Discovery in Industrial Scenarios

Focus Your Distribution: Coarse-to-Fine Non-Contrastive Learning for Anomaly Detection and Localization

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

AnomalyControl: Learning Cross-modal Semantic Features for Controllable Anomaly Synthesis

A Diffusion-Based Framework for Multi-Class Anomaly Detection

DAS3D: Dual-modality Anomaly Synthesis for 3D Anomaly Detection

Mixed Noise-Guided Mutual Constraint Framework for Unsupervised Anomaly Detection in Smart Industries

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

Natural Synthetic Anomalies for Self-Supervised Anomaly Detection and Localization

Anomaly-Aware Semantic Segmentation by Leveraging Synthetic-Unknown Data

Collaborative Discrepancy Optimization for Reliable Image Anomaly Localization

AnomalySD: Few-Shot Multi-Class Anomaly Detection with Stable Diffusion Model

AnomalyXFusion: Multi-modal Anomaly Synthesis with Diffusion

GLAD: Towards Better Reconstruction with Global and Local Adaptive Diffusion Models for Unsupervised Anomaly Detection

SeaS: Few-shot Industrial Anomaly Image Generation with Separation and Sharing Fine-tuning

Autoencoder-Based Visual Anomaly Localization for Manufacturing Quality Control