摘要
不同的视频应用对视频对象的分割和跟踪的速度和精确度具有不同的要求。提出了一种视频对象的分级分割和跟踪框架,基于视频对象特征描述子算法可以实时地分割和跟踪视频对象,基于区域特征描述子算法对分割的视频对象进一步细化,提高空域准确性。该框架可以满足各种视频应用。
Different video applications have various speed and accuracy requirements for segmentation and tracking of video object. This paper brings out a scalable segmentation and tracking framework of video objects. The algorithm based on video object characteristic descriptors can real time segment and track video objects, and the algorithm based on region descriptors can further improve the spatial accuracy of the segmented video objects. The results of the algorithm can meet different requirements for various video applications.
出处
《计算机应用研究》
CSCD
北大核心
2005年第5期154-155,158,共3页
Application Research of Computers
关键词
视频对象
分割
跟踪
特征描述子
Video Object
Segmentation
Tracking
Characteristic Descriptors
