基于自组织特征映射的图象矢量量化

Sofm-based Image Vector Quantization

本文就基于自组织特征映射的图象矢量量化编码做了初步的探讨,得出一些结论。在矢量量化中,码本性能的好坏对重建的图像有直接的影响。我们利用自组织特征映射(SOFM)网络进行聚类,实现了图像矢量码本的生成,然后再根据矢量量化(VQ)编码原理将图像重建。该方法可以达到较高的压缩比,实现了图像压缩。并且,就不同条件下的图像作了对比。

In this paper, SOFM-based image vector qoantization is done some research , and we make some conclusions. During the vector quantization ,the performance of code table is vital to the reconstructed image. Here SOFM is adopted for clustering in order to make a good code table ,and then images are reconstructed. We can get more high compression ratio by this method and draw a comparisons between different images.