一种基于零树量化的小波变换图像压缩方法

Image Compression Algorithm Based on Zerotree of Wavelet Coefficients

在对零树编码方法研究的基础上 ,针对零树编码方法存在编码时间较长的缺点 ,提出了一种改进的零树编码方法 .主要包括 :对最低频子带系数单独编码 ;设立标志矩阵 ;以扫描单元为单位进行主表扫描 ;实验表明这种改进的零树编码方法 ,加快了编码速度 .另外 ,该编码方法也能够产生嵌入式码流 。

On the basis of EZW algorithm, a primary shortcoming in EZW algorithm making encoding time short is found. Several ways is present to improve its performance. Two flag matrix is set to avoid redundant scanning and replicate encoding of significant wavelet coefficients. The turn of scanning is rearranged according to scan units based on the tree structure. The new zerotree quantization algorithm not only decreases the time consuming of encoding, but also produces embed stream and supports multirate image decoding.