计算机科学与探索 ›› 2022, Vol. 16 ›› Issue (5): 991-1007.DOI: 10.3778/j.issn.1673-9418.2110022

• 综述·探索 • 上一篇    下一篇

人体动作识别与评价——区别、联系及研究进展

杨刚1, 张宇姝1, 宋震2,+()   

  1. 1.北京林业大学 信息学院,北京 100083
    2.中央戏剧学院 传统戏剧数字化高精尖研究中心,北京 100710
  • 收稿日期:2021-10-13 修回日期:2022-01-06 出版日期:2022-05-01 发布日期:2022-05-19
  • 通讯作者: + E-mail: songzhen@zhongxi.cn
  • 作者简介:杨刚(1977—),男,山西长治人,博士,副教授,CCF会员,主要研究方向为计算机图形学、虚拟现实等。
    张宇姝(1997—),女,湖北十堰人,硕士研究生,CCF会员,主要研究方向为计算机科学与技术(虚拟现实方向)。
    宋震(1976—),男,北京人,博士,教授,博士生导师,主要研究方向为传统戏剧数字化、戏剧人工智能、数字演员与未来戏剧等。
  • 基金资助:
    北京高校卓越青年科学家计划项目(BJJWZYJH01201910048035)

Human Action Recognition and Evaluation—Differences, Connections and Research Progress

YANG Gang1, ZHANG Yushu1, SONG Zhen2,+()   

  1. 1. School of Information Science and Technology, Beijing Forestry University, Beijing 100083, China
    2. Advanced Research Center for Digitalization of Traditional Drama, The Central Academy of Drama, Beijing 100710, China
  • Received:2021-10-13 Revised:2022-01-06 Online:2022-05-01 Published:2022-05-19
  • About author:YANG Gang, born in 1977, Ph.D., associate professor, member of CCF. His research interests include computer graphics, virtual reality, etc.
    ZHANG Yushu, born in 1997, M.S. candidate, member of CCF. Her research interest is computer science and technology (direction of virtual reality).
    SONG Zhen, born in 1976, Ph.D., professor, Ph.D. supervisor. His research interests include digitalization of traditional drama, drama artificial intelligence, digital actors and future drama, etc.
  • Supported by:
    Beijing Outstanding Young Scientist Program(BJJWZYJH01201910048035)

摘要:

人体动作识别与动作评价是近年来的热点研究问题。两者在数据类型、数据处理、特征描述等方面有许多相通之处。近年来,随着应用需求的显著增长,出现了大量有关动作识别与评价的研究工作,但两者间的区别与联系,以及它们的理论方法和技术路线还未见系统的分析与总结。从应用目的与技术特点等方面出发,探讨了两者的联系,给出了两者较为明确的概念界定。在此基础上,从数据处理流程的角度出发,将动作识别与动作评价归纳到一个统一的技术框架中;依据此框架,对动作识别与评价所涉及到的各个重要环节,包括数据类型、预处理、特征描述、分类方法、评价方法等的研究进展和存在的问题进行了系统阐述。其中,在分类方法环节,将当前动作识别的分类方法划分为基于统计模型的方法和基于深度学习的方法进行论述;而在评价方法环节,则以专家知识介入方式为依据,将当前的动作评价相关工作划分为四类并进行了系统梳理。最后对当前存在的瓶颈及未来研究重点进行了总结与展望。

关键词: 动作识别, 动作评价, 特征描述, 相似性度量

Abstract:

Human action recognition and action evaluation are hot research issues in recent years. Technologies of action recognition and action evaluation share similarities in terms of data sources, data pre-processing, feature description, etc. In conjunction with the significant recent growth of application requirements, numerous studies on action recognition and evaluation appear. However, the differences and connections between human action reco-gnition and evaluation, as well as their theoretical methods and technical routes, have not been systematically analyzed and summarized. Starting from the perspective of application purpose and technical characteristics, the relationship between the two is discussed, and a clearer concept definition is given. On this basis, action recognition and action evaluation are summarized into a unified technical framework from the perspective of data processing flow. Based on this framework, the research progress and existing problems of all important links involved in motion recognition and evaluation, including data types, pre-processing, feature description, classification methods and evaluation methods, are systematically described. Among them, in the classification method section, the current action recognition classification methods are divided into statistical model-based methods and deep learning-based methods for discussion; and in the evaluation method section, based on the intervention method of expert know-ledge, the current action evaluation related work is divided into four categories and systematically sorted out. Finally, the bottlenecks and the focus of future research are summarized and prospected.

Key words: action recognition, action evaluation, feature description, similarity measurement

中图分类号: