Improved Deep Support Vector Data Description Model Using Feature Patching for Industrial Anomaly Detection
Sensors · 26.12.2024 · 25 (1) · 67
DOI: 10.3390/s25010067 ↗Kurz gefasst
Fachartikel in der Zeitschrift Sensors, online erschienen am 26.12.2024. Verfasst von Wei Huang und 4 weiteren Autor:innen. Beteiligte Einrichtung: College of Computer Science, Zhejiang University of Technology, Hangzhou 310023, China. Die Arbeit wurde bisher 7-mal zitiert (Crossref) und nennt 38 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Key R&D Program of China, "Pioneer" R&D Program of Zhejiang, Zhejiang Provincial Natural Science Foundation.
Abstract (Originalsprache)
In industrial contexts, anomaly detection is crucial for ensuring quality control and maintaining operational efficiency in manufacturing processes. Leveraging high-level features extracted from ImageNet-trained networks and the robust capabilities of the Deep Support Vector Data Description (SVDD) model for anomaly detection, this paper proposes an improved Deep SVDD model, termed Feature-Patching SVDD (FPSVDD), designed for unsupervised anomaly detection in industrial applications. This model integrates a feature-patching technique with the Deep SVDD framework. Features are extracted from a pre-trained backbone network on ImageNet, and each extracted feature is split into multiple small patches of appropriate size. This approach effectively captures both macro-structural information and fine-grained local information from the extracted features, enhancing the model’s sensitivity to anomalies. The feature patches are then aggregated and concatenated for further training with the Deep SVDD model. Experimental results on both the MvTec AD and CIFAR-10 datasets demonstrate that our model outperforms current mainstream approaches and provides significant improvements in anomaly detection performance, which is vital for industrial quality assurance and defect detection in real-time manufacturing scenarios.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- College of Computer Science, Zhejiang University of Technology, Hangzhou 310023, China.
- Zhejiang HOUDAR Intelligent Technology Co., Ltd., Hangzhou 310023, China.
ORCID-Profile: Yongjie Li, Zhaonan Xu, Xinwei Yao
Förderung
- National Key R&D Program of China
- "Pioneer" R&D Program of Zhejiang
- Zhejiang Provincial Natural Science Foundation
- National Key R&D Program of China
- “Pioneer” and “Leading Goose” R&D Program of Zhejiang
Zitieren
APA
Huang, W., Li, Y., Xu, Z., Yao, X., & Wan, R. (2024). Improved Deep Support Vector Data Description Model Using Feature Patching for Industrial Anomaly Detection. Sensors, 25(1), 67. https://doi.org/10.3390/s25010067
Vancouver
Huang W, Li Y, Xu Z, Yao X, Wan R. Improved Deep Support Vector Data Description Model Using Feature Patching for Industrial Anomaly Detection. Sensors. 2024;25(1):67. doi:10.3390/s25010067
BibTeX
@article{Huang2024,
author = {Huang, Wei and Li, Yongjie and Xu, Zhaonan and Yao, Xinwei and Wan, Rongchun},
title = {Improved Deep Support Vector Data Description Model Using Feature Patching for Industrial Anomaly Detection},
journal = {Sensors},
year = {2024},
volume = {25},
number = {1},
pages = {67},
doi = {10.3390/s25010067}
}
EUVIMED