嵌入式零树小波编码算法的优化及仿真实现

Implementation of optimized embedded zero-tree wavelet coding algorithm

嵌入式零树小波编码算法存在扫描过程过于复杂和表示小波系数的符号总数过多两个缺陷,这些缺陷在很大程度影响着算法的扫描速度和编码效率。通过两个方面的优化可以减少这些缺陷造成的影响:一是引入有效子带矩阵来决定是否对某个不重要的小波系数的后代子孙结点进行扫描,这样有效地减少了扫描的时间;二是采用6种符号来表示小波系数,很大程度上减少了表示小波系数的符号总数,也就极大地减少了编码表示小波系数的符号。经过优化后的嵌入式零树小波编码算法不但加快了扫描速度,也提高了编码效率。

In the research of embedded zero-tree wavelet algorithm, we found that the scanning process is too complex and the total symbols which are marked wavelet coefficients are too much. These flaws affect scanning speed and coding efficiency of embedded zero-tree wavelet algorithm to a large extent. To reduce the impact of these flaws, we take two kinds of optimizations ： one was introduce effective subband matrix to determine whether the descendants of a un-significant coefficient could be scanned, the other was use six kinds of symbols to mark wavelet coefficients. The experimental results showed that the first scheme could reduce scanning time consumption effectively, the second scheme reduced the total coding symbols greatly and the total bytes that were encoded these symbols. In a word, these optimizations speeded up scanning speed and enhanced the coding efficiency.