ShiftSplit-AD is a method that separates domain shift from defects in visual anomaly detection by decomposing the residual matrix of DINOv2 features into low‑rank and row‑sparse components. The sparse component is used for scoring anomalies, optionally fused with the low‑rank part. Experiments on AeBAD‑S show that sparse‑only scoring raises image AUROC from 0.6780 to 0.7294 and AUPRC from 0.8052 to 0.8465, but it also lowers clean AUROC on MVTec categories and hurts Bottle localization, highlighting a trade‑off between filtering shift and preserving defect information.
By Muhamathu Ameer Ali Aacaas Muhamath
Reconstruction-based methods are a cornerstone of unsupervised image anomaly detection, but they remain vulnerable to \emph{outlier leakage}, where standard mean squared error (MSE) loss drives the model to faithfully reconstruct anomalous patterns. We propose a Non-linear Reconstruction Loss that applies a sigmoid-based squashing function to suppress high-magnitude features, preventing outliers from dominating optimization while preserving sensitivity to normal patterns.
This paper introduces IDEAL, a framework for few-shot anomaly detection that uses both normal and anomalous reference examples. IDEAL suppresses irrelevant normal variations and encodes intrinsic deviation vectors that capture discriminative anomaly directions. Experiments on eight real-world datasets show that IDEAL generalizes to unseen anomalies and outperforms existing methods.
By Huan Wang, Jun Shen, Jun Yan, Guansong Pang
SPARC is a few‑shot calibration technique for vision‑based industrial anomaly detectors that corrects deployment‑time nuisances by projecting patch features onto a per‑cell subspace, requiring only up to eight verified‑normal images and no gradient updates. It operates between the encoder and detector, using a closed‑form, spatially indexed estimate based on the encoder’s native patch grid. Across seven detectors on shift‑prone benchmarks, SPARC boosts pooled Image AUROC by 13.8 pp and AU‑PRO₀.₃ by 3.5 pp, while showing modest changes on benchmarks without engineered shift.
arXiv:2608.23723v1 Announce Type: new
Abstract: Few-shot anomaly detection (FSAD) has recently benefited from vision-language models such as CLIP, which enable anomaly de?tection by aligning visual f...
By Wenyang Liu, Tianyi Liu, Dongshuo Zhang, Kejun Wu, Adams Wai-Kin Kong
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:2610.07016v1 Announce Type: cross
Abstract: In few-shot industrial anomaly detection, the few normal target images provide no direct defect supervision, making anomaly prompts difficult to lear...
By Mengyang Zhao, Teng Fu, Haiyang Yu, Ke Niu, Bin Li, Xiangyang Xue
arXiv:2610.32964v1 Announce Type: new
Abstract: Few-shot industrial anomaly detection requires modeling category-specific normality from only a handful of defect-free images. Common approaches tie in...
By Bernardo Perrone Ribeiro, Erkut Akdag, Egor Bondarev
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.
By Changyi Li, Miao Yu, Kai Dong, Yu Xiao
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
The paper introduces TED (Text-Axis Evidence Decomposition), a post‑hoc scoring method that improves anomaly localization in CLIP‑based detectors without altering the backbone or prompts. TED evaluates whether ambiguous responses are better supported by defect patches or normal patches, thereby distinguishing true defects from visually complex normal regions. Experiments show that TED significantly enhances pixel‑level localization across frozen VLM backbones and adapted hosts, especially under hard‑false‑positive competition.
By JinYoung Kim, Geonho Kim, GiJeong Park, Geonu Lee, YoungJoon Yoo
The paper introduces Contrastive Dual Gaussian Processes (CDGP), a weakly supervised approach for anomaly segmentation in industrial visual inspection. CDGP models normal and anomaly predictive distributions over dense tokens, using a posterior-dominance statistic that normalizes predictive-mean differences by joint uncertainty to provide spatial evidence and image-level confidence. The method achieves top performance on MVTec AD~2, KSDD2, and VisA datasets without requiring pixel-level annotations or test-time fitting.