基于矢量量化的嵌入式零树小波改进编码方法

基于矢量量化的SOFM算法和嵌入式零树小波算法(EZW)如今已经广泛的被使用于图像压缩领域,均被认为是非常有效的压缩编码技术。矢量量化方法压缩比较高,但往往存在分块效应,而EZW算法在高压缩比情况下存在恢复的图像质量较差的问题,提出一种基于矢量量化的嵌入式零树小波方法,它的基本原理是引入差值图像思想,先对原图像做矢量量化,再将原图像与矢量量化的恢复图像求差值图像,差值图像经过小波分解以后会存在大量为0的小波系数,再利用EZW编码,帮助提高重建图像质量。实验表明,相对于EZW和JPEG2000算法,本文算法的压缩比和编码质量均有显著提高。

Improved Embedded Zerotree Wavelet Coding Based on the Vector Quantization

Vector quantization approach high compression ratio,but usually have a domino effect.EZW algorithm in high compression ratio cases recovers images of poor quality.This paper proposes a Improved embedded zerotree wavelet coding method based on the vector quantization.The basic principle is the introduction of the concept of value for the difference image,the difference value image after wavelet decomposition can have a large number of 0 of wavelet coefficients.Using EZW coding will be conducive to improving the quality of reconstructed image.