多小波变换零树图像编码方法研究

Study of Multiwavelet Zerotree Image Coding

研究多小波变换零树图像编码的性能 .在标量零树方面 ,通过对图像多小波系数的重新排列 ,提出了新的多小波系数树的定义 ,并由实验证明它比直接采用单小波零树的组织更为有效 ;在矢量零树方面 ,根据多小波变换处理的是矢量信号这一特点 ,提出了矢量四叉树的数据结构 ,并从最优率失真意义上设计了矢量量化的多小波零树编码方案 ,结果表明 ,在该方案下 ,多小波比优选的

We investigate the performance of multiwavelet zerotree image coding. By rearranging the multiwavelet coefficients, a scalar zerotree coding algorithm can achieve better performance than the reported methods, which has been verified by experimental results. Furthermore, we propose a new vector quadtree structure and the corresponding optimal vector zerotree coding scheme in rate distortion sense. This scheme can match the feature of mutiwavelet transform efficiently. By using this scheme, the coding efficiency of multiwavelet outperforms that of the 9 7 scalar wavelet.