| 一种基于优化预处理的均值漂移算法 |
| |
| 引用本文： | 史以桢,黎宁.一种基于优化预处理的均值漂移算法[J].信息通信,2009(3):40-43,59. |
| |
| 作者姓名： | 史以桢 黎宁 |
| |
| 作者单位： | 南京航空航天大学,信息科学与技术学院,江苏,南京·210016 |
| |
| 摘 要： | 对传统的Mean Shift算法进行了改进.加入了前期预处理.包括特征选择与分类.提高了复杂场景中目标的鲁棒性.传统的Mean Shift算法固定的使用一个或两个特征(比如颜色)对目标进行跟踪.当跟踪场景发生变化.容易跟踪失败.我们通过选择目标与背景区别度最大的特征同时定义显著特征与非显着特征,实现了对复杂变化场景中目标更为鲁棒的跟踪.通过对多组数据进行实验.证实了该算法的可靠性.
|
| 关 键 词： | 目标跟踪 特征选择与分类 显著特征 |
A mean-shift algorithm based on pretreatment |
| |
| SHI Yi-zhen,LI Ning.A mean-shift algorithm based on pretreatment[J].Information & Communications,2009(3):40-43,59. |
| |
| Authors: | SHI Yi-zhen LI Ning |
| |
| Affiliation: | Department of Communication and Information System;Nanjing University of Aeronautics & Astronautics;Nanjing 210016;China |
| |
| Abstract: | We improve the standard mean shift tracking algorithm and add the pretreatment, including selection and classifi cation of features, then we improve the robustness of the object tracking in the complicated circumstance. One or two features (such as the color ) will be selected stably for object tracking in the standard mean shift tracking algorithm , but the tracking of object maybe fail when the circumstance change . The features which distinguish the object from the background mostly will be selected in o... |
| |
| Keywords: | Mean Shift |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
