广义高斯信源的自适应编码研究

Adaptive Coding for Generalized Gaussian Sources in Image Coding

首先介绍了当前最新的几种视频编码标准并且进行了比较,熵编码是每一个视频编码标准必须认真研究的课题,为了减少由于失配所带来的效率损失,本文提出了一种自适应编码技术:自适应指数哥伦布码,并与自适应算术编码进行了比较。分析和仿真都表明即使信源特性在大范围内发生变化,自适应指数哥伦布码对于保持高的编码效率是足够稳健的(90%以上的情况),同时保持了指数哥伦布码和哥伦布-莱期码的简洁性。

Several latest video coding standards are introduced and compared, Since entropy coding is one of the most important topics that need serious study in the fields of video coding standards, the main work of this paper focuses on adaptive entropy coding. In this paper, an adaptive coding technique is proposed, namely, adaptive Exp-Golomb codes （AEG）, in order to mitigate the efficiency loss due to mismatch. Analysis and simulations show that the adaptive techniques are robust enough to keep high coding efficiency （constantly about 90%） even if the sources undergo big changes.