一种基于小波系数方向性的矢量分类方法

One Method of Vector Classifying Based on the Direction of Wavelet Coefficients

针对基于小波变换和矢量量化的图像压缩算法,利用小波变换后系数具有明显方向性的特点以及带内小波系数的相关性,提出一种基于小波系数方向性的矢量分类方法,同时采用一种基于小波系数排序截断的阈值选取方法.实验表明该分类方法原理简单易于实现,提高了算法的整体效率,对不同的图像具有分类效果稳定的优点.

A new method of vector classification, referring to the image compression algorithm based on wavelet transformation and vector quantization, is presented. It utilizes direction and correlation between wavelet coefficients. The threshold selection is based on reordered wavelet coefficients series. Experimental results demonstrate that the principle of this method is simple and available. This method has good stability in classifying different images and improves the efficiency of the image compression algorithm.