节点文献
基于混合运动模型的边信息生成算法
Algorithm of side information generation based on mixed motion model
【摘要】 为提高不规则运动视频序列的边信息质量,提出基于混合运动模型的边信息生成算法。对曲线运动假设理论进行研究,分析二次函数和指数函数的理论原理,在此基础上设定阈值,将对称运动矢量和采用二次函数和指数函数的曲线运动估计相结合,建立采用混合运动模型的边信息生成算法的步骤和流程。利用3个视频序列进行实验测试仿真,实验结果表明,该算法特别适用于运动剧烈的视频序列,得到的边信息更准确,所需码率明显减少。
【Abstract】 To improve the quality of side information of the irregular motion video sequences,the algorithm of side information generation based on mixed motion model was presented.The curvilinear motion hypothesis theory was studied,the principles of the quadratic and exponential functions were analyzed,and the threshold was set.Symmetrical motion vector and curvilinear motion estimate based on quadratic and exponential function were combined,and the steps and processes of side information generation algorithm using mixed motion model were established.The test with three video sequences was done.The test results show that the algorithm is particularly suitable for severe motion video sequences,more accurate side information is obtained,and the desired bit rate is significantly reduced.
【Key words】 side information; mixed motion model; symmetrical motion vector; curvilinear motion; rate;
- 【文献出处】 计算机工程与设计 ,Computer Engineering and Design , 编辑部邮箱 ,2016年10期
- 【分类号】TN919.81
- 【被引频次】2
- 【下载频次】38