论文部分内容阅读
在 MPEG- 4编码系统框架中 ,为支持基于内容的编码性能 ,视频自动、半自动分割成为关键技术之一 .文中提出了一种利用以一组帧为参考帧将视频序列自动分割成静止背景和运动对象的算法 .这种方法考虑了全局的运动 ,并且提出一种新的运动区域检测算法 .算法主要思想是先对帧进行运动区域的简化 ,通过简化后的帧与原帧的帧差获取运动区域 .将以一组帧为参考所检测出的运动区域合成一个二进制掩模 ,对这一掩模进行运动分析和形态滤波 ,从而获得最终的运动对象 .实验结果表明 ,利用本文所提出的方法能够较精确地分割出运动的视频对象 .
In the framework of MPEG-4 coding system, automatic and semi-automatic video segmentation is one of the key technologies in order to support the content-based coding performance.In this paper, a method is proposed to automatically segment a video sequence into a still background and This method takes global motion into account and proposes a new motion region detection algorithm.The main idea of the algorithm is to simplify the motion region of the frame first and obtain the frame difference between the simplified frame and the original frame Motion region. A binary mask is synthesized from the motion region detected by using a set of frames, and the motion analysis and morphological filtering are performed on the mask to obtain the final motion object. The experimental results show that, The method can segment the moving video object more accurately.