摘要
本文推广了矢量量化中的乘积码为类似线性码的形式,从而提高了数据压缩的精度。本文的主要结果是证明了算法的收敛性.并给出了应用于图象压缩的一个例子.
In this paper we use linear-like codes to expand the product codes to a more general form forvector quantization, and give the learning algorithm Finally we prove that this algorithm converges in finite stepsand the classification is local optimal. We also give an example of image compression by using this algorithm, theresult shows better quality of restored image comparing with block codes method.
出处
《信号处理》
CSCD
北大核心
1993年第2期115-120,共6页
Journal of Signal Processing
