北京航空航天大学学报 ›› 2015, Vol. 41 ›› Issue (6): 1122-1127.doi: 10.13700/j.bh.1001-5965.2014.0414

• 论文 • 上一篇    下一篇

基于局域性约束线性编码的人体动作识别

白琛, 孙军华   

  1. 北京航空航天大学 仪器科学与光电工程学院, 北京 100191
  • 收稿日期:2014-07-10 出版日期:2015-06-20 发布日期:2015-07-30
  • 通讯作者: 孙军华(1975—),男,湖北荆门人,教授,sjh@buaa.edu.cn,主要研究方向为视觉测量、图像分析与识别. E-mail:sjh@buaa.edu.cn
  • 作者简介:白琛(1990—),男,天津人,硕士研究生,chenbai@aspe.buaa.edu.cn

Human action recognition based on locality-constrained linear coding

BAI Chen, SUN Junhua   

  1. School of Instrumentation Science and Opto-electronics Engineering, Beijing University of Aeronautics and Astronautics, Beijing 100191, China
  • Received:2014-07-10 Online:2015-06-20 Published:2015-07-30

摘要: 针对动作特征类内差异较大,导致动作分类识别率较低的问题,以及当前算法在计算复杂度和扩展可识别动作类别方面的不足,提出一种基于局域性约束线性编码(LLC)的人体动作识别方法.算法将人体关节的位置、速度和加速度作为局部动作特征;采用局域性约束线性编码对局部动作特征求解稀疏表达,从而减小特征的类内差异,增强区别力;由于编码方法具有解析解,方法处理视频速度可达760帧/s;词典由K均值法分别对每类数据学习得到的子词典组成,使算法在扩展可识别动作类别时无需全局优化.此外,为避免了词典较大情况下分类器的过拟合现象,利用词典元素类别对编码系数进行降维.在使用深度摄像机获得的MSR-Action3D数据库上对所提出的方法进行验证,取得了85.7%的识别率.

关键词: 动作识别, 局域性约束线性编码, 词典学习, 时间金字塔匹配, 深度图像

Abstract: Large intra-class variations of action features lead to low classification accuracy of action recognition, on the other hand, current algorithms exist drawbacks in computational complexity and extension of recognizable action classes. A method based on locality-constrained linear coding (LLC) for action recognition from depth images was proposed. In order to reduce the intra-class variations and increase classification accuracy, joints' positions, velocities and acceleration features were concatenated to form local action features, then LLC was used to calculate sparse representations of local action features. Analytical solution of LLC ensures computational speed of our method is up to 760 frames per second. Dictionary is composed by sub-dictionaries learned by K-means from features of each class separately, so global optimization is avoided during extending recognizable action classes. Moreover, to avoid classifier to be over-fitting, a dimensionality reduction method based on labels of dictionary items was proposed. The proposed method was evaluated on MSR-Action3D dataset captured by depth cameras. The experimental results show that the proposed approach achieves classification accuracy of 85.7%.

Key words: action recognition, locality-constrained linear coding, dictionary learning, temporal pyramid matching, depth images

中图分类号: 


版权所有 © 《北京航空航天大学学报》编辑部
通讯地址:北京市海淀区学院路37号 北京航空航天大学学报编辑部 邮编:100191 E-mail:jbuaa@buaa.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发