H.264宏块选择的自适应算法研究

在视频编码的过程中,运动估计占据了举足轻重的地位.其性能优劣会在一定程度
上决定了码流质量.采用16×16 宏块能够提高压缩比,但容易产生方块效应,使用4×4
宏块可以提高图像质量,但降低了压缩比。H.264 视频压缩标准采用多模式运动估计,
可以有效减少块匹配预测误差,但随着模式的增多,算法计算量成倍增加.为了克服这个
困难,本文提出一种新的自适应的宏块划分和运动估计算法。这种分块算法在综合考虑
图像本身特性在编码过程中,根据图像的各个部分运动程度不同而采用不同的分快策
略,兼顾图像质量和处理开销。
关键词: H.264; 运动估计; 块匹配
Abstract: Motion estimation plays an important role in video encoding. Its performance
affects the quality of encoded bit stream greatly. Using 16×16 macro block leads to serious
mosaic affection, Although it may be help to enhance the compression ratio. Using 4×4
macro block can improve picture quality easily comparing with previous method, but it
sacrifice the high compression ratio. Through the using of multi block pattern, H.264 gets
the balance between quality and compression ratio. With the increasing of macro block
patterns, computing complexity increases sharply. In order to avoid this dilemma, this
paper provides a new algorithm to accomplish this goal. According to the moving character
of the picture, the algorithm adopts diverse policy to get the balance between quality and
complexity