视觉动态手势识别综述  

Overview on vision-based dynamic gesture recognition

作　　者：田秋红[1] 杨慧敏 梁庆龙 包嘉欣 TIAN Qiuhong;YANG Huimin;LIANG Qinglong;BAO Jiaxin(School of Informatics Science and Technology,Zhejiang Sci-Tech University,Hangzhou 310018,China)

机构地区：[1]浙江理工大学信息学院,杭州310018

出　　处：《浙江理工大学学报（自然科学版）》2020年第4期557-569,共13页Journal of Zhejiang Sci-Tech University(Natural Sciences)

基　　金：国家自然科学基金项目(51405448);浙江理工大学博士科研启动项目(11122932611817);浙江理工大学国家级大学生创新创业训练计划项目(201910338012);浙江省大学生科技成果推广项目(14530031661961)。

摘　　要：基于视觉的动态手势识别是人机交互领域应用较为广泛的技术,其发展对于实现人机自然交互有着重要研究意义。文章介绍并分析了目前常用的手势交互系统。通过对近年来国内外相关文献的领域研究进行梳理,概述了视觉动态手势识别的一般流程,其流程包括检测与分割、追踪、特征提取、分类,并对各流程所涉及的方法及优缺点进行了对比分析;探讨了视觉动态手势识别研究所面临的挑战性问题和未来可能的研究方向,为推动该领域的进一步研究提供参考。Vision-based dynamic gesture recognition is widely used in the field of human-computer interaction,and its development has important research significance for achieving natural human-computer interaction.This paper introduces and analyzes the current common gesture interaction system.By collating domestic and foreign relevant literatures in recent years,this paper summarizes the general process of vision-based dynamic gesture recognition,including detection and segmentation,tracking,feature extraction and classification.The methods,advantages and disadvantages involved in each process.The challenges and possible future research directions on vision0 based dynamic gesture recognition are discussed,which can provide a reference for the further research in this field.

关 键 词：视觉动态手势识别综述 手势分割 手势追踪 特征提取 手势分类 

分 类 号：TS195.644[轻工技术与工程—纺织化学与染整工程]