| 基于主成分追踪的视频文本定位 |
| |
| 引用本文： | 刘路路,刘亚楠.基于主成分追踪的视频文本定位[J].合肥师范学院学报,2014,32(6):27-29. |
| |
| 作者姓名： | 刘路路 刘亚楠 |
| |
| 作者单位： | 合肥师范学院计算机科学与技术系,安徽合肥,230039 |
| |
| 基金项目： | 安徽省高校省级优秀青年人才基金重点项目“视频文本提取关键技术研究” |
| |
| 摘 要： | 视频中的大部分文本,特别是标题字幕会与背景有较强的对比,另外,部分文本会做线性运行或不规则运动,针对上述两种情况,提出了基于主成分追踪的视频文本定位方法,将主成分追踪这种矩阵分解算法成功应用于视频文本的定位系统中,主要包含两个步骤,首先进行视频分解,通过主成分追踪算法将视频帧分解成低秩矩阵和稀疏矩阵两部分,稀疏矩阵中包含了大部分文本信息,接下来在稀疏矩阵中进行文本定位.该方法减少了后续的文本提取和识别的难度,算法简单可行,实用性强,具有一定的现实意义.
|
| 关 键 词： | 鲁棒主成分分析 主成分追踪 文本定位 |
Video Text Orientation Based on Principle Component Pursuit |
| |
| LIU Lu-lu,LIU Ya-nan.Video Text Orientation Based on Principle Component Pursuit[J].合肥师范学院学报,2014,32(6):27-29. |
| |
| Authors: | LIU Lu-lu LIU Ya-nan |
| |
| Institution: | HU Bi-Ling , GUO Yu-Tang(1. School of Computer Science, He f ei Normal University, He f ei 14098, China 2. School of Computer Science Engineering, University of Electronic Science and Technology of China, Chengdu10614 , China) |
| |
| Abstract: | In large scale network, in which routing protocol is based on link state, it is a key problem topromptly update the routing tables in terms of the network flow and its topology. The changes for the linkstate are： the increase in link cost, the decrease in link cost, the inactive nodes, new nodes added in. Thispaper firstly reviews all the important algorithms and analyzed their innovative point; Secondly, it pointsout the faults in relevant articles which put forward the algorithms to deal with topological change, andwhat＇s more, proposes a new way to deal with the conditions together. |
| |
| Keywords: | dynamic computing the shortest path tree routing algorithms |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
