| 结合人体运动特征的行为识别 |
| |
| 引用本文： | 李宁,须德,傅晓英,袁玲.结合人体运动特征的行为识别[J].北方交通大学学报,2009(2):6-10. |
| |
| 作者姓名： | 李宁 须德 傅晓英 袁玲 |
| |
| 作者单位： | 北京交通大学计算机与信息技术学院,北京100044 |
| |
| 基金项目： | 国家“863计划”项目资助（2007AA01Z168）;北京交通大学科技基金资助项目（2007XM008） |
| |
| 摘 要： | 人体运动具有马尔可夫性质,即当前状态只受前一状态的影响.目前为止,用于人体行为识别的隐马尔可夫模型（HMM）大多使用的是全连接结构（Full-Connected structure）,并且没有把状态数目的选取和状态转移条件与人体运动特性间的关系作为研究重点.本文针对这种关系提出了基于“从左到右三状态半连接HMM”的人体行为识别方法,为每个状态的输出概率引入了权重的概念.实验表明,该方法能够在降低运算复杂度的同时,提高行为识别率,从而证明了人体运动特性分析在HAR领域中的应用价值.
|
| 关 键 词： | 行为识别 前中后三状态半连接HMM 人体运动特征 星状骨架特征 半连接HMM 全连接HMM |
Action Recognition Combined with Human Action Property |
| |
| LI Ning,XU De,FU Xiaoying,YUAN Ling.Action Recognition Combined with Human Action Property[J].Journal of Northern Jiaotong University,2009(2):6-10. |
| |
| Authors: | LI Ning XU De FU Xiaoying YUAN Ling |
| |
| Institution: | (School of Computer and Information Technology, Beijing Jiaotong University, Beijing 100044, China) |
| |
| Abstract: | Hidden Markov Model （HMM） based human action recognition （HAR） has been broadly adopted by HAR community. However, existing works do not pay attention to the relationship between the layout of the model and the property of human action. In this paper, a novel HAR method is proposed based on the assumption that human action can be essentially recognized by three key postures located around the initial, middle and terminal action period. Rested on this hypothesis, we improve the HMM-Based HAR method and propose a left-to-right three-state HMM. Experiments show the approach constitutes a suggestive plausibility proof for the close relationship between action property and the design of HMM for HAR task. |
| |
| Keywords: | action recognition begin-middle-end semi-connected HM M human action property starskeleton feature semi-connected HMM （SCHMM） full-connected HMM （FCHMM） |
| 本文献已被 维普 等数据库收录！ |
