Dual Memory-Guided Probabilistic Model for Weakly-Supervised Anomaly Detection

Hsiu Hua Chou, Ruyi Xu, Kang Yang Huang, Jhih Ciang Wu, Hong Han Shuai, Wen-Huang Cheng*

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Anomaly detection seeks to identify the patterns of instances distinguishable from normal ones. However, current methods primarily align with either one-class or open-set scenarios, leading to an inadequate exploration of anomalous examples. In this paper, we propose a weakly-supervised approach, the Dual Memory-guided Probabilistic Model (DMPM), to explore the comprehensive knowledge of normal and abnormal instances. Employing such dual memory banks, our model provides opposing guidance for probabilistic models during the denoising procedure. We illustrate the effectiveness of our DMPM in addressing weakly-supervised anomaly detection and conduct extensive experiments on popular industrial benchmarks, i.e., MVTec AD and VisA. Moreover, we highlight the adaptability of the unified DMPM, demonstrating its compatibility with diffusion-based approaches that perform on image or latent space.

Original languageEnglish
Title of host publicationHuman Activity Recognition and Anomaly Detection - 4th International Workshop, DL-HAR 2024, and 1st International Workshop, ADFM 2024, Held in Conjunction with IJCAI 2024, Revised Selected Papers
EditorsKuan-Chuan Peng, Yizhou Wang, Ziyue Li, Zhenghua Chen, Min Wu, Jianfei Yang, Sungho Suh
PublisherSpringer Science and Business Media Deutschland GmbH
Pages50-65
Number of pages16
ISBN (Print)9789819790029
DOIs
StatePublished - 2025
Event4th International Workshop on Deep Learning for Human Activity Recognition, DL-HAR 2024, and 1st International Workshop on Anomaly Detection with Foundation Models, ADFM 2024, Held in Conjunction with the International Joint Conference on AI, IJCAI 2024 - Jeju, Korea, Republic of
Duration: 3 Aug 20249 Aug 2024

Publication series

NameCommunications in Computer and Information Science
Volume2201 CCIS
ISSN (Print)1865-0929
ISSN (Electronic)1865-0937

Conference

Conference4th International Workshop on Deep Learning for Human Activity Recognition, DL-HAR 2024, and 1st International Workshop on Anomaly Detection with Foundation Models, ADFM 2024, Held in Conjunction with the International Joint Conference on AI, IJCAI 2024
Country/TerritoryKorea, Republic of
CityJeju
Period3/08/249/08/24

Keywords

  • Anomaly detection
  • Diffusion model
  • Weakly-supervised learning

Fingerprint

Dive into the research topics of 'Dual Memory-Guided Probabilistic Model for Weakly-Supervised Anomaly Detection'. Together they form a unique fingerprint.

Cite this