UniFlow: Unified Normalizing Flow for Unsupervised Multi-Class Anomaly Detection

Multi-class anomaly detection is more efficient and less resource-consuming in industrial anomaly detection scenes that involve multiple categories or exhibit large intra-class diversity. However, most industrial image anomaly detection methods are developed for one-class anomaly detection, which ty...

Full description

Saved in:
Bibliographic Details
Main Authors: Jianmei Zhong, Yanzhi Song
Format: Article
Language:English
Published: MDPI AG 2024-12-01
Series:Information
Subjects:
Online Access:https://www.mdpi.com/2078-2489/15/12/791
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850042221713162240
author Jianmei Zhong
Yanzhi Song
author_facet Jianmei Zhong
Yanzhi Song
author_sort Jianmei Zhong
collection DOAJ
description Multi-class anomaly detection is more efficient and less resource-consuming in industrial anomaly detection scenes that involve multiple categories or exhibit large intra-class diversity. However, most industrial image anomaly detection methods are developed for one-class anomaly detection, which typically suffer significant performance drops in multi-class scenarios. Research specifically targeting multi-class anomaly detection remains relatively limited. In this work, we propose a powerful unified normalizing flow for multi-class anomaly detection, which we call UniFlow. A multi-cognitive visual adapter (Mona) is employed in our method as the feature adaptation layer to adapt image features for both the multi-class anomaly detection task and the normalizing flow model, facilitating the learning of general knowledge of normal images across multiple categories. We adopt multi-cognitive convolutional networks with high capacity to construct the coupling layers within the normalizing flow model for more effective multi-class distribution modeling. In addition, we employ a multi-scale feature fusion module to aggregate features from various levels, thereby obtaining fused features with enhanced expressive capabilities. UniFlow achieves a class-average image-level AUROC of 99.1% and a class-average pixel-level AUROC of 98.0% on MVTec AD, outperforming the SOTA multi-class anomaly detection methods. Extensive experiments on three benchmark datasets, MVTec AD, VisA, and BTAD, demonstrate the efficacy and superiority of our unified normalizing flow in multi-class anomaly detection.
format Article
id doaj-art-8decdfe9357f4f7aae6c7fc508a6846c
institution DOAJ
issn 2078-2489
language English
publishDate 2024-12-01
publisher MDPI AG
record_format Article
series Information
spelling doaj-art-8decdfe9357f4f7aae6c7fc508a6846c2025-08-20T02:55:38ZengMDPI AGInformation2078-24892024-12-01151279110.3390/info15120791UniFlow: Unified Normalizing Flow for Unsupervised Multi-Class Anomaly DetectionJianmei Zhong0Yanzhi Song1School of Artificial Intelligence and Data Science, University of Science and Technology of China, Hefei 230026, ChinaSchool of Artificial Intelligence and Data Science, University of Science and Technology of China, Hefei 230026, ChinaMulti-class anomaly detection is more efficient and less resource-consuming in industrial anomaly detection scenes that involve multiple categories or exhibit large intra-class diversity. However, most industrial image anomaly detection methods are developed for one-class anomaly detection, which typically suffer significant performance drops in multi-class scenarios. Research specifically targeting multi-class anomaly detection remains relatively limited. In this work, we propose a powerful unified normalizing flow for multi-class anomaly detection, which we call UniFlow. A multi-cognitive visual adapter (Mona) is employed in our method as the feature adaptation layer to adapt image features for both the multi-class anomaly detection task and the normalizing flow model, facilitating the learning of general knowledge of normal images across multiple categories. We adopt multi-cognitive convolutional networks with high capacity to construct the coupling layers within the normalizing flow model for more effective multi-class distribution modeling. In addition, we employ a multi-scale feature fusion module to aggregate features from various levels, thereby obtaining fused features with enhanced expressive capabilities. UniFlow achieves a class-average image-level AUROC of 99.1% and a class-average pixel-level AUROC of 98.0% on MVTec AD, outperforming the SOTA multi-class anomaly detection methods. Extensive experiments on three benchmark datasets, MVTec AD, VisA, and BTAD, demonstrate the efficacy and superiority of our unified normalizing flow in multi-class anomaly detection.https://www.mdpi.com/2078-2489/15/12/791multi-class anomaly detectionanomaly localizationdefect detectionnormalizing flow
spellingShingle Jianmei Zhong
Yanzhi Song
UniFlow: Unified Normalizing Flow for Unsupervised Multi-Class Anomaly Detection
Information
multi-class anomaly detection
anomaly localization
defect detection
normalizing flow
title UniFlow: Unified Normalizing Flow for Unsupervised Multi-Class Anomaly Detection
title_full UniFlow: Unified Normalizing Flow for Unsupervised Multi-Class Anomaly Detection
title_fullStr UniFlow: Unified Normalizing Flow for Unsupervised Multi-Class Anomaly Detection
title_full_unstemmed UniFlow: Unified Normalizing Flow for Unsupervised Multi-Class Anomaly Detection
title_short UniFlow: Unified Normalizing Flow for Unsupervised Multi-Class Anomaly Detection
title_sort uniflow unified normalizing flow for unsupervised multi class anomaly detection
topic multi-class anomaly detection
anomaly localization
defect detection
normalizing flow
url https://www.mdpi.com/2078-2489/15/12/791
work_keys_str_mv AT jianmeizhong uniflowunifiednormalizingflowforunsupervisedmulticlassanomalydetection
AT yanzhisong uniflowunifiednormalizingflowforunsupervisedmulticlassanomalydetection