论文部分内容阅读
为进一步提高基于一种运动信息可分级模型的可分级视频编码的编码效率,减小运动信息可分级的最低码率限制,对该运动可分级模型的二维多向性进行了具体研究与改进,更新了运动估计的流程,提出了多向上的两个运动可分级维度的运动分级等级一致性原则以及二维数据的渐进式存储存取结构,从而实现了二维多向上的运动可分级。实验测试结果证明该二维多向的运动信息模型优于不可分级的运动信息,能有效提高可分级视频编码系统的效率。
In order to further improve the coding efficiency of scalable video coding based on a scalable model of motion information and reduce the minimum rate limit of the scalable motion information, the two-dimensional multipath of the scalable model of motion is studied and improved , Updated the flow of motion estimation and put forward the principle of motion hierarchy level consistency of multi-directional two motion grading dimensions and the progressive memory access structure of two-dimensional data, so as to realize two-dimensional multi-directional motion grading. Experimental results show that the two-dimensional multi-directional motion information model is superior to the non-scalable motion information, which can effectively improve the efficiency of scalable video coding system.