Robust Distribution Alignment for Industrial Anomaly Detection under Distribution Shift

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Liao, Jingyi, Xu, Xun, Su, Yongyi, Tu, Rong-Cheng, Liu, Yifan, Tao, Dacheng, Yang, Xulei
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866908274877530112
author Liao, Jingyi
Xu, Xun
Su, Yongyi
Tu, Rong-Cheng
Liu, Yifan
Tao, Dacheng
Yang, Xulei
author_facet Liao, Jingyi
Xu, Xun
Su, Yongyi
Tu, Rong-Cheng
Liu, Yifan
Tao, Dacheng
Yang, Xulei
contents Anomaly detection plays a crucial role in quality control for industrial applications. However, ensuring robustness under unseen domain shifts such as lighting variations or sensor drift remains a significant challenge. Existing methods attempt to address domain shifts by training generalizable models but often rely on prior knowledge of target distributions and can hardly generalise to backbones designed for other data modalities. To overcome these limitations, we build upon memory-bank-based anomaly detection methods, optimizing a robust Sinkhorn distance on limited target training data to enhance generalization to unseen target domains. We evaluate the effectiveness on both 2D and 3D anomaly detection benchmarks with simulated distribution shifts. Our proposed method demonstrates superior results compared with state-of-the-art anomaly detection and domain adaptation methods.
format Preprint
id arxiv_https___arxiv_org_abs_2503_14910
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Robust Distribution Alignment for Industrial Anomaly Detection under Distribution Shift
Liao, Jingyi
Xu, Xun
Su, Yongyi
Tu, Rong-Cheng
Liu, Yifan
Tao, Dacheng
Yang, Xulei
Computer Vision and Pattern Recognition
Anomaly detection plays a crucial role in quality control for industrial applications. However, ensuring robustness under unseen domain shifts such as lighting variations or sensor drift remains a significant challenge. Existing methods attempt to address domain shifts by training generalizable models but often rely on prior knowledge of target distributions and can hardly generalise to backbones designed for other data modalities. To overcome these limitations, we build upon memory-bank-based anomaly detection methods, optimizing a robust Sinkhorn distance on limited target training data to enhance generalization to unseen target domains. We evaluate the effectiveness on both 2D and 3D anomaly detection benchmarks with simulated distribution shifts. Our proposed method demonstrates superior results compared with state-of-the-art anomaly detection and domain adaptation methods.
title Robust Distribution Alignment for Industrial Anomaly Detection under Distribution Shift
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2503.14910