基于最优矢量量化零树编码的多小波图像压缩

Multiwavelet Image Compression Based on Optimal Vector Zerotree Coding

提出一种基于多小波变换的矢量量化零树编码方案。利用多小波变换后各个子块之间的相关性,把图像的多小波分解看成是矢量四叉树结构,通过对多小波系数重排来保持小波系数的零树特征。对图像进行压缩仿真试验,结果表明多小波变换的图像压缩结果优于传统的小波变换结果。

We propose an optimal vector zerotree coding scheme based on multiwavelet transform. According to the correlation among each block, we bring forward a vector quadtree structure. We keep the zerotree feature by rearranging the multiwavelet coefficients. Simulation shows that the coding efficiency of multiwavelet under this scheme outperforms that of the traditional wavelet transform.