基于小波近似的分形矢量量化图像编码方法

本文把小波分解、分形、矢量量化有机地结合起来,提出基于小波近似的分形矢量量化图像编码方法(WA-FVQ-IC)。对原图进行小波分解及四叉树分割,形成小波近似的粗糙图像和分割块的统计平均值,小波近似的粗糙图像经变换构成变换的图像,并据此提取误差信号的码书,而后对误差信号进行分形矢量量化。本方法充分利用小波的多分辨率特性,构成差值信号分形矢量量化的码书,不需要外部训练,解码时不需要迭代,充分发挥了分形和矢量量化的优点。

Fractal Vector Quantization Image Coding Based on Wavelet Approximated

in this paper,we propose a Image encoding method on wavelet-approximated Fractal VectorQuantization. first, we perform Wavelet analysis and Quandtree partition on the original Image toWavelet to obtain Wavelet -approximated coarsely Image and statistics average value of partitionblock. these data are transmitted to receiver. then transform Image can be obtained by Wavelet-ap-proximated coarsely Image at transmitting and receiving side,meantime,the codebook of error Image,can be generated. finally, we perform Fractal Vector Quantization on error Image. this 'method, thecharacteristic of Multi-resolution of Wavelet is used to generate Fractal Vector Quantization code-book of error Image,it's unnecessary to do outside training and iterative processing when We decodethe Image,in fact we make full use of advantages of Fractal and Vector Quantization,the result of ex-periment show the excellent performance of this encoding method.