改进的MPEG-4静态纹理BQ模式编码方法

Improved MPEG-4 VTC BQ Mode Coding Algorithm

从零树符号的信息含义及其与编码层目标图像的一致性出发 ,研究MPEG 4VTC工具中BQ模式编码 (即PEZW方法 )的改进问题 提出基于符号分解的PEZW改进方法———SP PEZW方法 提出零树符号的分解表示方法 ,将零树符号中包含的两部分信息用两个分解符号分别表示 ,并通过删除编码层中与目标图像“不一致”的分解符号来提高编码层的压缩比 实验表明 ,SP PEZW方法与PEZW方法相比 ,低分辨率编码层有更高的压缩比 ,低分辨率空间层以指定码率解码的图像质量有明显提高 ;高分辨率编码层的压缩比并没有明显下降 ,而高分辨率空间层以指定码率解码的图像质量略有提高 更重要的是 ,SP PEZW方法的实现只是在PEZW方法中增加一个零树符号分解编码环节 ,同时还继承了PEZW方法的随机读取、有容错性、可实现局部编码。

Problem of scalable image coding is investigated from the viewpoint of consistency of zerotree symbol with the objective image of bitstream layer. An improved PEZW algorithm, SP-PEZW, is presented, in which a zerotree symbol is denoted by two decomposing symbols separately, and compression rates of the bitstream layers are improved by deleting the inconsistent decomposing symbols of the bitstream layers. Experiment results show that, comparing with PEZW algorithm, the bitrate of low resolution bitstream layer of SP-PEZW algorithm is lower, while bitrate of the highest resolution bitstream layer is the same, and image of the highest resolution spatial layer which is decoded at the special bitrate is better. Furthermore, SP-PEZW is implemented by adding a zerotree symbol decomposing coding module to the PEZW algorithm, and preserves the features such as random access, error resilience, localized coding, and low memory requirement.