# Training-Free Logical and Structural Anomaly Detection via Calibrated Fusion

arXiv: https://arxiv.org/abs/2609.05091 · HTML: https://mlsanity.com/paper/2609.05091/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 12 of its 16 references that are in the map.
- First posted on arXiv 4 Sep 2026.

## Builds on (most cited first, 12 of 12)

- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Wide Residual Networks](https://mlsanity.com/paper/1605.07146/) (2016)
- [Towards Total Recall in Industrial Anomaly Detection](https://mlsanity.com/paper/2106.08265/) (2021)
- [WinCLIP: Zero-/Few-Shot Anomaly Classification and Segmentation](https://mlsanity.com/paper/2303.14814/) (2023)
- [EfficientAD: Accurate Visual Anomaly Detection at Millisecond-Level Latencies](https://mlsanity.com/paper/2303.14535/) (2023)
- [Component-aware anomaly detection framework for adjustable and logical industrial visual inspection](https://mlsanity.com/paper/2305.08509/) (2023)
- [Few Shot Part Segmentation Reveals Compositional Logic for Industrial Anomaly Detection](https://mlsanity.com/paper/2312.13783/) (2023)
- [SAM-LAD: Segment Anything Model Meets Zero-Shot Logic Anomaly Detection](https://mlsanity.com/paper/2406.00625/) (2024)
- [CSAD: Unsupervised Component Segmentation for Logical Anomaly Detection](https://mlsanity.com/paper/2408.15628/) (2024)
- [PUAD: Frustratingly Simple Method for Robust Anomaly Detection](https://mlsanity.com/paper/2402.15143/) (2024)
- [Set Features for Anomaly Detection](https://mlsanity.com/paper/2311.14773/) (2023)
- [UniSLAD: A Unified Framework for Structural and Logical Industrial Visual Anomaly Detection](https://mlsanity.com/paper/2606.20768/) (2026)

## Abstract (arXiv)

Industrial anomaly detection must handle two distinct defect families: structural anomalies, which manifest as local texture corruptions, and logical anomalies, which violate global rules on object count, composition, or arrangement. Existing detectors typically favor one family at the expense of the other. In particular, training-free methods effectively exploit frozen representations but lack an explicit notion of object count, while methods that reason about counts usually rely on category-specific component modeling. We show that counting ability can be introduced into training-free anomaly detection without additional training or part-level supervision. Our key idea is a normal-set calibration that aligns heterogeneous anomaly cues using statistics from normal images, enabling their direct fusion within a unified training-free framework. Built upon this calibration, our detector combines complementary frozen cues to address both logical and structural anomalies. On MVTec-LOCO, our method achieves image-level AUROCs of 89.0 and 95.9 on logical and structural anomalies, respectively, yielding a 92.5 average---the best among training-free detectors in our comparison. It remains competitive with methods requiring network training or part annotations, while its structural variant matches PatchCore on MVTec-AD (99.1 image-AUROC), suggesting that the proposed calibration generalizes beyond logical anomaly detection.