摘要
提出了一种新的用于视频编码的变尺寸宏块运动估计方法。该方法通过粗尺度 16×16 的当前块和邻域块运动矢量之间的差值 S 判断场景运动的复杂度,再根据差值 S 和当前矢量,决定宏块尺寸。场景运动平稳的部分采用 16×16 的宏块,复杂的部分采用 8×8 或 4×4 的宏块。实验表明,低码率条件下,采用普通运动补偿算法,该变尺寸宏块运动估计方法的预测帧的平均 PSNR 值比常规的固定尺寸块运动估计约提高 1.1-2.7 dB,并显著降低了方块效应。
A new variable size macroblocks motion estimation method used for video coding is proposed. With this method, the complexity of scene motion is estimated by the difference S between 16 × 16 (coarse scale) current block and adjacent block vectors, then the macroblock size is determined according to difference 5 and current vector. The macroblock (16 × 16) is adopted for the parts with smooth scene motion. 8 × 8 or 4 × 4 macroblocks are adopted for the complex scene motion parts. Experiments show that under low bit-rate and adopting a common motion compensation algorithm, the average PSNR value of the predictive frame with variable size macroblock motion estimation method is improved for 1.1 - 2.7 dB better than that of the conventional fixed size block motion estimation, and the block border effect is significantly reduced.
出处
《光电工程》
EI
CAS
CSCD
北大核心
2005年第1期67-70,共4页
Opto-Electronic Engineering
基金
国家 863 计划资助项目
关键词
视频编码
变尺寸宏块
运动图像
运动补偿
Video coding
Various size macroblocks
Motion picture
Motion compensation
