一种基于自适应小波变换的矢量量化方法  

A Adaptive Wave-based Vector Quantization Algorithm

作　　者：王雪梅[1] 

机构地区：[1]枣庄学院数学与信息科学系,山东枣庄277160

出　　处：《佳木斯大学学报（自然科学版）》2007年第4期476-477,480,共3页Journal of Jiamusi University：Natural Science Edition

摘　　要：运用小波变换进行图像压缩的算法其核心都是小波变换的多分辨率分析以及对不同尺度的小波系数的量化和编码.本文提出了一种基于能量的自适应双正交小波变换和矢量量化相结合的算法.即在一定的能量准则下,根据子图像的能量大小决定是否进行小波分解,然后给出恰当的小波系数量化.该方法充分利用了不同尺度间小波系数的相关性并采用自组织特征映射神经网络进行矢量量化.实验表明,该方法获得较高的编码效率和重构图像质量.Image compression based on wavelet transform, which gives emphasis to multi - resolution analysis and quantization and coding of the different scaling wavelet coefficients, this paper puts forward a compression algorithm based on vector quantization and adaptive biorthogonal wavelet transform based on energy. That sub - images are whether to be compressed or not is decided by their energy that is defined by certain criterion. Then the quantization and coding of some wavelet coefficients is given. The correlation of the wavelet coefficients among different scales is fully utilized and the Self- Organizing feature map neural network is used. The results of experiments show that high coding efficiency and reconstructed image quality are both obtained.

关 键 词：自适应小波变换 矢量量化 自组织特征映射神经网络 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]