SafeRestore introduces a framework for certifying when an industrial image restoration should be automatically returned to a detector or require human review. It ranks five restoration candidates using action‑specific fitted scores, selects a threshold gate on tuning data, and evaluates the gate on a separate certification sample with two one‑sided exact binomial bounds—one for evidence‑loss incidents and one for excess‑activation incidents. In a retrospective study of 4,591 Carinthia‑S images, the protocol demonstrates auditable risk‑coverage behavior, with varying pass rates across different policies and morphologies.
By Shaoliang Yang, Jun Wang
The paper investigates how to improve patch‑memory anomaly detectors for steel defect detection when additional industrial images may contain unseen defects. By filtering out the most suspicious 20% of patches from a contaminated reference bank and merging the remaining patches with a clean seed bank, the authors reduce contamination from 9.46% to 2.59% and achieve higher AUPRC scores compared to naive expansion or random removal. The method demonstrates that reference purity is a critical design factor and that unverified images can be beneficial only after explicit filtering.
By Hannaneh Kalantari, Javad Khoramdel
The paper introduces a training‑free, human‑in‑the‑loop anomaly detection framework that allows a domain expert to correct a PatchCore detector by editing its memory bank, without retraining or using gradients. Using only ten golden samples, operator corrections close a median 66% of the performance gap to a fully trained bank, improving 12 of 15 MVTec AD categories while harming none. The approach is evaluated with a rigorous held‑out protocol and shows that passive and active querying yield statistically indistinguishable gains, with a defect‑memory extension failing decisively.
By Ayusha Abbas, Saram Abbas, Kabita Adhikari
The paper investigates how many normal samples are required to reliably set an alarm threshold for few‑shot anomaly detectors, focusing on distribution‑free certification limits. Using a frozen DINOv2 PCA residual ranker on 15 MVTec and 12 VisA categories, the authors show that simple leave‑one‑image‑out calibration is limited by resolution and shift, leading to empirical false‑alarm rates far above the nominal level. They derive a category‑count feasibility calculus, demonstrating that at least 14, 29, and 59 independent category draws are needed for 95% upper confidence bounds at α=0.20, 0.10, and 0.05, and propose the CRESS protocol to split source categories into reference, proposal, and certification roles.
whyItMatters":"The study provides concrete numerical thresholds for the amount of source evidence needed to guarantee reliable anomaly detection in new categories, informing practical deployment of few‑shot detectors."
By Gia Huy Thai, Nguyen Thai Anh
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:2606. 13780v1 Announce Type: cross Abstract: Machine-learned anomaly detection is reshaping searches for new physics, but it has outrun the statistics used to interpret it.
By Jack Y. Araz, Michael Spannowsky
Vision-language models (VLMs) are often reported to outperform task-specific vision backbones for unmanned aerial vehicle (UAV) power-line defect assessment. We test that claim on ElecVQA-Bench, a 56,...
arXiv:2606. 22054v2 Announce Type: replace-cross Abstract: Detectors for GNSS radio-frequency impairments (jamming, spoofing, multipath) are usually reported with a single AUC measured on the distribution they were tuned on.
By Chakshu Baweja
arXiv:2608.23299v1 Announce Type: new
Abstract: Memory-based anomaly detectors store nominal training patches and score test patches against this memory. A patch selected for coverage therefore becom...
By Joongwon Chae, Runming Wang, Peiwu Qin
The paper evaluates the claim that vision‑language models (VLMs) outperform task‑specific vision backbones for UAV power‑line defect assessment using the ElecVQA‑Bench benchmark. Across various evaluation settings—partitioning, item sets, label spaces, replication, resolution, and side information—the performance gap between VLMs and traditional backbones is minimal or even reversed when controlling for resolution and token budget. The study concludes that VLM superiority is not universally supported and emphasizes the importance of rigorous benchmark audits.
By Linghao Zhang, Siyu Xiang, Junwei Kuang, Peiyu Yi
arXiv:2608. 12652v1 Announce Type: cross Abstract: Benchmark contamination is diagnosed today with n-gram overlap, with likelihood-based membership inference, or with canary strings, and each needs something usually unavailable: the training corpus, a well-chosen test statistic, or foresight at dataset release.
By Florian Braun
arXiv:2608. 07914v1 Announce Type: new Abstract: Behavioral contamination detectors can return "no evidence" either because a benchmark is clean or because the audit has little power.
By Ibne Farabi Shihab, Sanjeda Akter, Anuj Sharma