| 基于CHMM的行为识别系统设计 |
| |
| 引用本文： | 丁然,林春丽,夏余.基于CHMM的行为识别系统设计[J].鞍山钢铁学院学报,2010(5):474-477. |
| |
| 作者姓名： | 丁然 林春丽 夏余 |
| |
| 作者单位： | [1]辽宁科技大学电子与信息工程学院,辽宁鞍山114051 [2]辽宁科技大学高等职业技术学院,辽宁鞍山114051 |
| |
| 摘 要： | 阐述了行为识别的基本过程。将前景序列进行归一化后,采用二维主成分分析方法（2DPCA）降维进行特征提取,利用连续隐马尔科夫模型（CHMM）建模,对行为有效分类和识别。实验结果表明,该设计方案的识别率可以达到95%。
|
| 关 键 词： | 行为识别 二维主成分分析 连续隐马尔科夫模型 |
Design of behavior recognition system with CHMM |
| |
| DING Ran,LIN Chun-Li,XIA Yu.Design of behavior recognition system with CHMM[J].Journal of Anshan Institute of Iron and Steel Technology,2010(5):474-477. |
| |
| Authors: | DING Ran LIN Chun-Li XIA Yu |
| |
| Affiliation: | 1.School of Electric and Information Engineering,University of Science and Technology Liaoning,Anshan 114051,China;2.School of Higher Vocational Education,University of Science and Technology Liaoning,Anshan 114051,China) |
| |
| Abstract: | The process of the behavior recognition is described in detail.The foreground sequences is normalized,and a two-dimensional principle component analysis（2DPCA） is adopted to reduce the dimension and extract features directly,then a continuous hidden Markov model（CHMM） is used to classify the behavior.Experimental results show that the recognition rate can reach 95%. |
| |
| Keywords: | behavior recognition 2DPCA CHMM |
| 本文献已被 维普 等数据库收录！ |
|
