基于运动矢量的摄像机运动定性分类方法
A QUALITATIVE CAMERA MOTION CLASSIFICATION METHOD BASED ON MOTION VECTOR
-
摘要: 在视频序列中 ,摄像机运动在某种程度上反映了当前视频的部分语义信息 ,知道了摄像机运动将能够更好地实现对视频的浏览以及检索 .针对现有的依据光流分布来求摄像机运动参数算法中的不足 ,给出了一种定性分析相机运动的方法 ,该方法能够在相机聚焦中心—— FOE(focus- of- expansion)不在成像平面中心时检测出给定类型的摄像机运动 .实验结果表明该方法对于从视频序列中分析出给定类型的摄像机运动具有较好的效果Abstract: Motion of camera can reflect certain kinds of semantic information in video sequence. It is an active research direction in content based video retrieval or browse system. According to the shortcoming of the present methods based on optical flow distribution, a new method is presented to extract certain kinds of camera motion parameters even when the FOE is not in the center of the image. Experiments show that this method can provide satisfactory results in camera motion classification of video sequence.
下载: