一种利用FMO特性实现ROI可伸缩编码的方法  

An Improved ROI Scalable Video Coding Scheme Based on FMO

作　　者：陈刚[1] 戴礼荣[1] 宋彦[1] 赖伟[1] 

机构地区：[1]中国科学技术大学电子工程与信息科学系,安徽合肥230027

出　　处：《计算机仿真》2007年第1期101-106,共6页Computer Simulation

摘　　要：为了在限定带宽下得到更好的编码效率和视频质量,在对现有的感兴趣区域(Regions O f Interest,RO I)可伸缩编码进行分析的基础上,提出了一种利用H.264框架下的灵活宏块排序(Flexib leMacrob lock O rdering,FMO)特性实现基于RO I的分层可伸缩编码方法。该方法通过改善低带宽下RO I优先编码传输方式及时空域码流分配方案,克服了现有的RO I编码方法带来的边界效应和运动错位现象,在增强视频的整体主观质量的同时提高了码流的可伸缩性。仿真和统计试验结果表明,在较低带宽下,所提出的方法在编码性能和主观视觉效果上有较大的改善。This paper proposes a novel scalable videcr coding scheme based on the analysis of recent methods on ROI （Region of Interesting） Scalable Video Coding, which takes advantage of the FMO （Flexible Macroblock Ordering） property in H. 264 framework to obtain enhanced coding efficiency and better subjective video quality under limited bandwidth. In this scheme, by improving the bit allocation scheme on temporal and spatial scalable coding and transferring ROI priority, this method overcomes the boundary effect and distortion between ROI and background scene and enhances subjective visual quality and scalability of bit streams. Simulation and experimental results indicate that, under the same bandwidth, the proposed scheme is ＂superior to the methods reported in literature recently in term of subjective visual quality and coding performance such as video recovery and scalability.

关 键 词：可伸缩编码 码流分配 感兴趣区域 灵活宏块排序 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]