计算机工程与应用2025，Vol.61Issue(14)：65-87,23.DOI:10.3778/j.issn.1002-8331.2410-0291
步态识别研究综述
Review of Gait Recognition Research
摘要
Abstract
Gait recognition is a biometric technology,which is used to identify people by analyzing their walking patterns through cameras.It has the advantages of long-distance operation,being difficult to spoof,and working at low resolution.In recent years,with the rapid development of deep learning technology,gait recognition technology has been significantly developed in the integration with deep learning technology,especially the application of convolutional neural networks,circular neural networks,self-encoder,and generative confrontation networks in gait recognition has greatly improved the efficiency and accuracy of gait recognition.This paper summarizes the development of gait recognition technology,including its unique characteristics,development history,and literature review in recent years.The paper lists different data sets and discusses their different characteristics,and proposes a classification method of gait recognition technology,including body representation-based and template-based methods.In addition,it discusses the application of gait recogni-tion in some scenes and the existing problems,including the defects of easy loss of time and fine-grained spatial infor-mation,and further discusses the future development direction of gait recognition.关键词
步态识别/深度学习/身体表征/步态模板Key words
gait recognition/deep learning/physical representation/gait template分类
信息技术与安全科学引用本文复制引用
史晓国,云静,张钰莹,刘雪颖..步态识别研究综述[J].计算机工程与应用,2025,61(14):65-87,23.基金项目
国家自然科学基金(62062055) （62062055）
内蒙古高校青年科技英才项目(NJYT24061) （NJYT24061）
内蒙古自治区直属高校基本科研业务费项目(JY20220249). （JY20220249）
