Memory-enhanced spatial-temporal encoding framework for industrial anomaly detection system

作     者：Liu, Yang Ju, Bobo Yang, Dingkang Peng, Liyuan Li, Di Sun, Peng Li, Chengfang Yang, Hao Liu, Jing Song, Liang 

作者机构：Fudan Univ Acad Engn & Technol Shanghai 200433 Peoples R China Univ Toronto Dept Comp Sci Toronto ON M5S 1A1 Canada Duke Kunshan Univ Div Nat & Appl Sci Kunshan 215316 Jiangsu Peoples R China Fudan Univ State Key Lab ASIC & Syst Shanghai 201203 Peoples R China Ningbo Univ Fac Elect Engn & Comp Sci Ningbo 315211 Zhejiang Peoples R China Fudan Univ Sch Informat Sci & Technol Shanghai 200433 Peoples R China Univ British Columbia Dept Elect & Comp Engn Vancouver BC V6T 1Z4 Canada Innovat Platform Academicians Hainan Prov Haikou 570228 Hainan Peoples R China 

出 版 物：《EXPERT SYSTEMS WITH APPLICATIONS》 (Expert Sys Appl)

年 卷 期：2024年第250卷

核心收录：

学科分类：1201[管理学-管理科学与工程(可授管理学、工学学位)] 0808[工学-电气工程] 08[工学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：China Mobile Research Fund of the Chinese Ministry of Education [KEH2310029] Specific Research Fund of the Innovation Platform for Academicians of Hainan Province, China [YSPTZX202314] National Natural Science Foundation of China Shanghai Key Research Lab, China NSAI China Scholarship Council Joint Lab. on Networked AI Edge Computing Fudan University-Changan 

主　　题：Industrial anomaly detection Intelligent information system Spatial-temporal anomaly detection Deep autoencoder Normality learning 

摘      要：The development of modern manufacturing has raised greater demands on the accuracy, response speed, and operating cost of industrial accident warnings. Compared to conventional contact sensors, surveillance cameras can contactlessly capture spatial-temporal information of the open workspace with stable data quality, widely used in industrial process monitoring. However, due to the scarcity of industrial video datasets and the rarity and diversity of abnormal events, existing video -based anomaly detection models perform poorly in manufacturing scenarios. In this regard, we collect two datasets from typical industrial sites and propose a memory -enhanced spatial-temporal encoding (MSTE) framework for automatic industrial anomaly detection. The proposed MSTE framework learns spatial and temporal normality as well as spatial-temporal correlations with parallel structures and simultaneously measures deviations in appearance, motion, and consistency to respond to complex industrial anomalies accurately. Experimental results on public benchmarks and realworld industrial videos show that our method outperforms existing methods and achieves accurate temporal localization of various spatial-temporal anomalies, which helps to improve the safety and reliability of intelligent manufacturing.