| 基于自分裂竞争学习算法的关键帧提取 |
| |
| 引用本文： | 夏利民,邓克捷.基于自分裂竞争学习算法的关键帧提取[J].计算机工程与应用,2011,47(2):146-148. |
| |
| 作者姓名： | 夏利民 邓克捷 |
| |
| 作者单位： | 中南大学,信息与工程学院,长沙,410075 |
| |
| 基金项目： | 中国博士点基金，湖南省自然科学基金 |
| |
| 摘 要： | 模糊C均值算法在进行关键帧提取时难以取得全局最优值,导致所提取的关键帧无法完整地描述镜头信息。提出一种基于自分裂竞争学习(SSCL)的关键帧提取方法,根据SSCL的分裂机制确定全局最优类数目的特点来确定关键帧的数量,同时根据SSCL的竞争学习机制有效确定类中心的特点来确定准确的帧图像作为视频的关键帧。实验证明基于SSCL的关键帧提取的方法比基于模糊C均值关键帧提取的方法能够更好地描述镜头内容。
|
| 关 键 词： | 自分裂竞争学习 关键帧 聚类数目 聚类中心 |
| 收稿时间： | 2009-8-11 |
| 修稿时间： | 2009-10-20
|
Key frames extraction based on self-splitting competitive learning |
| |
| XIA Limin,DENG Kejie.Key frames extraction based on self-splitting competitive learning[J].Computer Engineering and Applications,2011,47(2):146-148. |
| |
| Authors: | XIA Limin DENG Kejie |
| |
| Affiliation: | School of Information Science and Engineering,Central South University,Changsha 410075,China |
| |
| Abstract: | Due to the limitation of getting global optimal solution on Fuzzy C Mean（FCM） clustering algorith in key framesetraction,it is difficult to describe the shot fully.An new way of extracting key frame on Slef-Splitting Competitive Learning（SSCL） will be proposed in this paper.It is effectively to calculate numbers of key frames by the characteristic of splittingof SSCL that get global optimal clusting numbers validly.At the same time,exact key frame will be extracted by the characteristic of competitive learning that get clustering center.The experiment will present,the advance which the result of SSCLis better than FCM in shot describing. |
| |
| Keywords: | self-splitting competitive learning key frame clustering number clustering center
|
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《计算机工程与应用》浏览原始摘要信息 |
|
点击此处可从《计算机工程与应用》下载全文 |
|
