arXiv Computer Vision

Training-Free Logical and Structural Anomaly Detection via Calibrated Fusion

The paper introduces a training‑free anomaly detector that simultaneously handles structural and logical defects by calibrating heterogeneous anomaly cues with statistics from normal images. This calibration aligns frozen representations, allowing their fusion without extra training or part‑level supervision. The resulting method achieves state‑of‑the‑art AUROC scores on MVTec‑LOCO and remains competitive on MVTec‑AD.

arXiv Computer Vision
Sep 3

DPA: Decoupling Product-Agnostic Anomaly Representations for Zero-shot Anomaly Generation

The paper introduces DPA, a diffusion-based framework that decouples product-agnostic anomaly representations to enable zero-shot anomaly generation. By reusing real anomalies from existing source products and filtering them for plausibility, DPA learns product-irrelevant anomaly embeddings that can be transferred across products. An adaptive mask-guided pipeline and a training-free labeling module further refine the realism and localization of generated anomalies, leading to improved performance on MVTec-AD, VisA, and a new anomaly-transfer benchmark.

By Hang Yao, Yansheng Fu, Ming Liu, Zifei Yan, Yanli Ji, Hongzhi Zhang, Wangmeng Zuo
arXiv Computer Vision
Aug 27

See More, Detect Less? Taming Information Leakage in Multi-View Anomaly Detection

The paper identifies a problem in multi‑view anomaly detection called cross‑view information leakage, where fusing multiple inspection views can cause normal features to mask anomalies during reconstruction. To address this, the authors propose GLAD, a framework that uses a Global‑Local Attention Driven approach, combining vision foundation model features with two fusion modules: Multi‑view Merging Attention for local, weighted fusion and Object‑Guided Attention for global context aggregation. Experiments on Real‑IAD and MANTA‑Tiny demonstrate that GLAD outperforms existing methods across various metrics, underscoring the importance of restricting information flow to preserve the reconstruction gap.

By Shang-Fu Chen, Kuan-Chuan Peng, Jhih-Ciang Wu, Wen-Huang Cheng, Kai-Lung Hua
arXiv Machine Learning
Jul 30

When Rule Violations Are Rare: Chimera Training for Logical Anomaly Detection

arXiv:2605. 26171v2 Announce Type: replace Abstract: Many practical anomalies are not merely rare inputs, but violations of semantic constraints: objects co-occur in structured ways, actions imply preconditions, and events satisfy temporal or relational regularities.

By Alejandro Ascarate, Leo Lebrat, Rodrigo Santa Cruz, Clinton Fookes, Olivier Salvado
arXiv Computer Vision
Sep 4

Neural-Collapse-guided Task-Free Continual Anomaly Detection

The paper introduces NC‑TFAD, a task‑free continual anomaly detection framework that leverages neural‑collapse geometry to learn from non‑stationary data streams without task boundaries. It freezes a pretrained backbone, aligns streaming features to a simplex Equiangular Tight Frame prototype space, and uses synthetic anomaly anchors, inter‑ and intra‑class regularization, and a Focal Neural Collapse Contrastive loss to stabilize representations and enhance normal‑anomaly separability. A normal‑patch‑prototype‑guided localization branch generates calibrated anomaly heatmaps, and extensive experiments on MVTec AD and VisA demonstrate that NC‑TFAD outperforms existing task‑free continual learning and unified anomaly detection baselines in both image‑level detection and pixel‑level localization.

By Xiaotong Kong, Chaoyang Song, Ziai Zhou, Jinxia Zhang, Kanjian Zhang, Haikun Wei
arXiv Computer Vision
Sep 4

ISP-AD: A Large-Scale Real-World Dataset for Advancing Industrial Anomaly Detection with Synthetic and Real Defects

The paper introduces ISP-AD, the largest publicly available industrial anomaly detection dataset, featuring both synthetic and real defects from a factory floor. It focuses on challenging, small, weakly contrasted surface defects within highly variable structured patterns, addressing the bias of existing datasets toward ideal imaging conditions. Experiments demonstrate that even a small amount of weakly labeled real defects improves model generalization and that synthetic defects can serve as a useful cold‑start baseline for scalable training.

By Paul J. Krassnig, Dieter P. Gruber