| 基于LDA和SVM的步态识别 |
| |
| 引用本文： | 许文芳,吴清江,王青力. 基于LDA和SVM的步态识别[J]. 计算机应用与软件, 2008, 25(5): 210-212 |
| |
| 作者姓名： | 许文芳 吴清江 王青力 |
| |
| 作者单位： | 华侨大学信息科学与工程学院,福建,泉州,362021 |
| |
| 摘 要： | 提出了一种新的基于人体中线投影的步态特征提取方法,同时将线性判别分析(LDA)和支持向量机(SVM)相结合进行步态的分类和识别.应用背景差方法分割出运动人体轮廓,对外轮廓沿人体中线投影可以得到前后两个向量,合成1D向量作为步态特征.通过线性判别分析对得到的一维向量进行特征提取和压缩,对得到的识别量应用支持向量机进行步态的分类和识别.应用上述方法在中科院自动化所的步态数据库上进行了实验,实验结果表明该步态识别方法具有较好的识别性能.
|
| 关 键 词： | 步态识别 特征提取 中线投影 |
| 修稿时间： | 2006-08-31 |
GAIT RECOGNITION BASED ON LDA AND SVM |
| |
| Xu Wen-fang,Wu Qing-jiang,Wang Qing-li. GAIT RECOGNITION BASED ON LDA AND SVM[J]. Computer Applications and Software, 2008, 25(5): 210-212 |
| |
| Authors: | Xu Wen-fang Wu Qing-jiang Wang Qing-li |
| |
| Affiliation: | Xu Wenfang Wu Qingjiang Wang Qingli(College of Information Science , Engineering,Huaqiao University,Quanzhou 362021,Fujian,China) |
| |
| Abstract: | A new gait feature extraction method based on body midline projection is presented in this paper,in it the linear discriminant analysis(LDA) and support vector machine(SVM) were combined to classify and recognize gait.Moving body silhouette extraction was achieved by background subtraction.The body silhouette was projected along its midline to obtain the fore-and-aft midline projection vectors,which were then combined into one dimensional data vector as the gait feature.LDA was applied to extract the featur... |
| |
| Keywords: | Gait recognition Feature extraction Midline projection |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
