| 一种正则粒子滤波的人脸跟踪算法 |
| |
| 引用本文： | 张瞳.一种正则粒子滤波的人脸跟踪算法[J].科技广场,2011(7):43-45. |
| |
| 作者姓名： | 张瞳 |
| |
| 作者单位： | 宝鸡文理学院物理与信息技术系,陕西宝鸡,721007 |
| |
| 基金项目： | 宝鸡文理学院重点项目(NO.ZK09132)资助 |
| |
| 摘 要： | 本文针对人脸运动的非线性非高斯的特点,引入正则粒子滤波算法来进行运动预测估计,且根据预测精度对预测过程中目标运动速度和过程噪声方差进行自适应更新。实验结果表明,在人脸的旋转、肤色和部分遮挡影响下跟踪精度较高,抵抗光照环境变化,以及人脸大小变化等的鲁棒性较强。
|
| 关 键 词： | 人脸轮廓跟踪 粒子滤波 非线性滤波 Bhattacharyya距离 |
Human Face Tracking Algorithm Based on Particle Filter |
| |
| Zhang Tong.Human Face Tracking Algorithm Based on Particle Filter[J].Science Mosaic,2011(7):43-45. |
| |
| Authors: | Zhang Tong |
| |
| Institution: | Zhang Tong(Department of Physics and Information Technology,Baoji University of Arts & Science,Shaanxi Baoji 721007) |
| |
| Abstract: | Human tracking technique is one of core method in intelligence surveillance system.The human face moving is non-linear and non-gauss,particle filter algorithm is proposed for motion estimation,which is efficient even in occlusion situation.According to the face structure,we advanced a blocking color histogram to describe the human face character. |
| |
| Keywords: | Human Face Estimation Particle Filter Non-linear Bhattacharyya |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
