节点文献

基于对象的视频分割技术研究

Research on Video Segmentation for Object-Based Application

【作者】 初广丽

【导师】 崔广才;

【作者基本信息】 长春理工大学 , 计算机软件与理论, 2007, 硕士

【摘要】 随着信息技术和计算机网络的飞速发展,多媒体信息已经成为人类获得信息的主要载体,而且日益丰富。海量数据要求必须提供高效的压缩和处理方法来对信息进行存储、传输和检索等各种处理。基于对象的视频处理技术是有效措施之一,视频对象分割是视频技术的基础而重要的工作,视频分割质量的好坏直接影响后期视频处理工作。 本文在阅读大量国内外最新文献以及对多种分割算法进行深入研究的基础上提出了一种快速有效的基于背景重构和变化检测的视频分割算法,该算法利用统计累积的方法得到包含遮挡域的完整背景信息,利用当前帧和所得到的背景图像相减得到视频对象,成功的克服了遮挡问题,在静止的背景条件下,对于单一运动目标能够得到比较准确的分割结果。

【Abstract】 With the rapid development of information technology and Internet, multimedia has become the main media where people get information from. Abundant data force people to look for an efficient and high impact way to deal with so much information. Such as storing, transmitting and searching. Well, the technology of video object-based is one of important means. Video object segmentation is the base of video technology. The quality of segmentation will affect the subsequent work.After reading a lot of the latest articles and researching all kinds of algorithms, I brought forward an auto segmentation algorithm based on rebuilding background and change detection in the article. It takes advantage of statistic and addition to get the integrated background. Then subtract the background from the current frame. It can get rid of the effect of shelter. We can get more accurate result by using this algorithm under the condition of still background and single motion object.

  • 【分类号】TP391.41
  • 【被引频次】4
  • 【下载频次】230
节点文献中: 

本文链接的文献网络图示:

本文的引文网络