FO-PaDiM is proposed, a feature configuration optimization framework built on the PaDiM pipeline that keeps the original statistical modeling unchanged while exhaustively searching feature-layer combinations and embedding dimensionality, and provides evidence that multi-level feature configurations can be beneficial but that their advantage is category-dependent and small in magnitude.
Abstract
Industrial anomaly detection is a critical component of machine-vision-based intelligent manufacturing, especially when defective samples are scarce and pixel-level annotations are expensive to obtain. PaDiM is a widely used unsupervised anomaly detection method that models patch-level feature distributions extracted from pretrained convolutional neural networks; its practical performance depends on the selected feature layers and the dimensionality of patch embeddings. This study proposes FO-PaDiM, a feature configuration optimization framework built on the PaDiM pipeline that keeps the original statistical modeling unchanged while exhaustively searching feature-layer combinations and embedding dimensionality. Unlike a previous exploratory version of this study, all configuration selection is performed under a leakage-free development protocol—a deterministic 80/20 split of training normals combined with multi-type synthetic anomalies—so that the MVTec AD test set is used exactly once, with a locked configuration and frozen thresholds. On eight MVTec AD categories (bottle, cable, capsule, hazelnut, metal_nut, pill, screw, transistor) with 14 configurations per category (seven non-empty ResNet-18 layer combinations × n_features in {32, 64}), the selection is found to be category-dependent: single-stage configurations (layer2, layer3 or layer1) are preferred in six categories and multi-stage combinations in two, while the previously reported universal preference for the three-layer combination does not replicate once test-set leakage is removed. Under the four evaluated weighting schemes, the layer combination remains unchanged for 62.5% of categories and n_features for 87.5%. Final locked test performance reaches a mean Image AUROC of 0.8676, Image F1 of 0.8079, Pixel AUROC of 0.9644, and Pixel F1 of 0.4582. The results provide evidence that multi-level feature configurations can be beneficial but that their advantage is category-dependent and small in magnitude, and they confirm the persistent difficulty of threshold-dependent binary localization (Pixel F1) despite strong ranking quality (Pixel AUROC).
Under industry 4.0, automated quality inspection, aiming to automatically identify surface defects on product, is vital for intelligent manufacturing. In the real-world industrial scenarios, the lack of defective samples and the high cost of manual annotation pose significant obstacles to the effective application of s...
Zhe-Xin Huang· Technological Pathways and D...· 0 citations
A lightweight real-time deep learning framework is proposed for industrial defect inspection, centering on an improved RT-DETR model that incorporates a large kernel attention backbone and a generalized layer aggregation network. This architecture addresses two issues: few-shot learning scenarios and stringent industri...
Jian-Rong Tian· International Conference on...· 0 citations
Unsupervised industrial anomaly detection uses normal training images to identify and localize defects. The challenge addressed here is localizing small defects amid normal appearance variation without requiring different feature layers to share one normal reference. We propose dual-layer DINOv3 anomaly-score fusion (D...
Industrial anomaly detection is increasingly required in real-world inspection settings. Although recent methods achieve strong performance on standard benchmarks, many of them still rely on architectures that are difficult to efficiently deploy in real-world edge scenarios. Thus, we present TinySimpleNet, a compact re...
Luigi Capogrosso, Enrico Fraccaroli, Michele Magno et al.· IEEE/IFIP International Conf...· 0 citations
We study unsupervised industrial anomaly detection in the single-class setting, where a detector learns only from normal images of one product category. Reconstruction-based detectors model normal features from a frozen encoder, but their target layers are usually chosen by architectural convention. We instead ask what...
Unsupervised anomaly detection is critical for industrial surface inspection, especially for resin cut-off wheels where defect samples are scarce and morphologically diverse. GANomaly serves as a classic reconstruction-based baseline using latent vector discrepancy, yet its single-score mechanism conflates pixel-level...
Xiao-Na Song, Run-Qing Zhang, Kai-Xuan Lv et al.· Italian National Conference...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.