Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Wu, Wensheng, Lu, Zheming, Lu, Ziqian, He, Zewei, Sun, Xuecheng, Wang, Zhao, Han, Jungong, Yu, Yunlong
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2603.02964
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866911482349879296
author Wu, Wensheng
Lu, Zheming
Lu, Ziqian
He, Zewei
Sun, Xuecheng
Wang, Zhao
Han, Jungong
Yu, Yunlong
author_facet Wu, Wensheng
Lu, Zheming
Lu, Ziqian
He, Zewei
Sun, Xuecheng
Wang, Zhao
Han, Jungong
Yu, Yunlong
contents Industrial anomaly detection faces significant challenges due to the scarcity of anomalous samples and the complexity of real-world anomalies. In this paper, we propose a foundation model-based anomaly synthesis pipeline (FMAS) that generates highly realistic anomalous samples without fine-tuning or class-specific training. Motivated by the distinct frequency-domain characteristics of anomalies, we introduce aWavelet Domain Attention Module (WDAM), which exploits adaptive sub-band processing to enhance anomaly feature extraction. The combination of FMAS and WDAM significantly improves anomaly detection sensitivity while maintaining computational efficiency. Comprehensive experiments on MVTec AD and VisA datasets demonstrate that WDAM, as a plug-and-play module, achieves substantial performance gains against existing baselines.
format Preprint
id arxiv_https___arxiv_org_abs_2603_02964
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Improving Anomaly Detection with Foundation-Model Synthesis and Wavelet-Domain Attention
Wu, Wensheng
Lu, Zheming
Lu, Ziqian
He, Zewei
Sun, Xuecheng
Wang, Zhao
Han, Jungong
Yu, Yunlong
Computer Vision and Pattern Recognition
Industrial anomaly detection faces significant challenges due to the scarcity of anomalous samples and the complexity of real-world anomalies. In this paper, we propose a foundation model-based anomaly synthesis pipeline (FMAS) that generates highly realistic anomalous samples without fine-tuning or class-specific training. Motivated by the distinct frequency-domain characteristics of anomalies, we introduce aWavelet Domain Attention Module (WDAM), which exploits adaptive sub-band processing to enhance anomaly feature extraction. The combination of FMAS and WDAM significantly improves anomaly detection sensitivity while maintaining computational efficiency. Comprehensive experiments on MVTec AD and VisA datasets demonstrate that WDAM, as a plug-and-play module, achieves substantial performance gains against existing baselines.
title Improving Anomaly Detection with Foundation-Model Synthesis and Wavelet-Domain Attention
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2603.02964