# Robust Logical Anomaly Detection under Capture Shifts via Synthetic Anomalies

arXiv: https://arxiv.org/abs/2610.18295 · HTML: https://mlsanity.com/paper/2610.18295/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 24 of its 59 references that are in the map.
- First posted on arXiv 13 Oct 2026.

## Builds on (most cited first, 24 of 24)

- [A ConvNet for the 2020s](https://mlsanity.com/paper/2201.03545/) (2022)
- [Benchmarking Neural Network Robustness to Common Corruptions and Perturbations](https://mlsanity.com/paper/1903.12261/) (2019)
- [InstructPix2Pix: Learning to Follow Image Editing Instructions](https://mlsanity.com/paper/2211.09800/) (2022)
- [Albumentations: fast and flexible image augmentations](https://mlsanity.com/paper/1809.06839/) (2018)
- [CutPaste: Self-Supervised Learning for Anomaly Detection and Localization](https://mlsanity.com/paper/2104.04015/) (2021)
- [DRAEM -- A discriminatively trained reconstruction embedding for surface anomaly detection](https://mlsanity.com/paper/2108.07610/) (2021)
- [Learning the Difference that Makes a Difference with Counterfactually-Augmented Data](https://mlsanity.com/paper/1909.12434/) (2019)
- [No Subclass Left Behind: Fine-Grained Robustness in Coarse-Grained Classification Problems](https://mlsanity.com/paper/2011.12945/) (2020)
- [Object-centric Auto-encoders and Dummy Anomalies for Abnormal Event Detection in Video](https://mlsanity.com/paper/1812.04960/) (2018)
- [Envisioning Outlier Exposure by Large Language Models for Out-of-Distribution Detection](https://mlsanity.com/paper/2406.00806/) (2024)
- [The MVTec AD 2 Dataset: Advanced Scenarios for Unsupervised Anomaly Detection](https://mlsanity.com/paper/2503.21622/) (2025)
- [Towards Training-free Anomaly Detection with Vision and Language Foundation Models](https://mlsanity.com/paper/2503.18325/) (2025)
- [LogicAD: Explainable Anomaly Detection via VLM-based Text Feature Extraction](https://mlsanity.com/paper/2501.01767/) (2025)
- [Industrial Anomaly Detection with Domain Shift: A Real-world Dataset and Masked Multi-scale Reconstruction](https://mlsanity.com/paper/2304.02216/) (2023)
- [Customizing Visual-Language Foundation Models for Multi-modal Anomaly Detection and Reasoning](https://mlsanity.com/paper/2403.11083/) (2024)
- [LogiCode: an LLM-Driven Framework for Logical Anomaly Detection](https://mlsanity.com/paper/2406.04687/) (2024)
- [CSAD: Unsupervised Component Segmentation for Logical Anomaly Detection](https://mlsanity.com/paper/2408.15628/) (2024)
- [LogicAL: Towards logical anomaly synthesis for unsupervised anomaly localization](https://mlsanity.com/paper/2405.06875/) (2024)
- [Out-of-Distribution Detection Without Class Labels](https://mlsanity.com/paper/2112.07662/) (2021)
- [Adapted-MoE: Mixture of Experts with Test-Time Adaption for Anomaly Detection](https://mlsanity.com/paper/2409.05611/) (2024)
- [Component-aware Unsupervised Logical Anomaly Generation for Industrial Anomaly Detection](https://mlsanity.com/paper/2502.11712/) (2025)
- [SALAD -- Semantics-Aware Logical Anomaly Detection](https://mlsanity.com/paper/2509.02101/) (2025)
- [AnomalyVFM -- Transforming Vision Foundation Models into Zero-Shot Anomaly Detectors](https://mlsanity.com/paper/2601.20524/) (2026)
- [MIRAGE: Model-agnostic Industrial Realistic Anomaly Generation and Evaluation for Visual Anomaly Detection](https://mlsanity.com/paper/2603.13507/) (2026)

## Abstract (arXiv)

Detectors of logical anomalies, such as missing, extra or wrongly combined parts, usually score a test image by its distance from normal (good) images. They achieve near-perfect image-level AUROC on standard industrial anomaly detection benchmarks such as MVTec LOCO, but their robustness to changes in capture conditions remains largely unexamined. Under a heterogeneous perturbation protocol in which each test image is randomly kept unaltered or subjected to a brightness reduction or a zoom-out transformation, the normal-reference methods we tested lose 9.7 to 34.1 AUROC points, and two of the strongest, CSAD and SALAD, fall from 97.7% and 96.4% to 68.1% and 62.3%. We propose a framework in which a classifier, trained on normal images and on synthetic anomalies, scores a test image without comparing it with normal data. From a short description of each product, a language model derives anomaly types and editing instructions, an image editor applies them to normal images, and the classifier learns to separate the edited images from automatically discovered normal sub-modes. With its full training recipe, the framework reaches 94.9% AUROC on clean images, 1.5 to 2.8 points below SALAD and CSAD, and 94.1% under the perturbation protocol. Trained without brightness or zoom-out augmentation, so that neither test perturbation occurs in training, it still reaches 92.5% (93.2% clean). These results suggest that clean benchmark scores can hide a strong sensitivity to capture conditions, and that learning the decision from both normal images and rule violations may reduce it.