基于改进CPN的快速矢量量化

对向传播神经网络(CPN)可以作为矢量量化器用于图像压缩，但CPN学习算法在进行码书设计时存在两个明显的缺陷。本文对CPN学习算法进行改进，提出了一种新的码书设计算法——快速竞争学习及误差修正算法(FCLECA)和一个基于改进CPN的快速矢量量化器模型，并讨论了FCLECA中的重要步骤和重要参数。仿真实验结果表明，FCLECA在生成高质量码书的同时大幅减少了训练时间，可以有效地实现快速矢量量化。

Fast Vector Quantization Based on a Modified Counterpropagation Network

The Counterpropagation Network(CPN) can be applied to image compression as a vector quantizer. However,the CPN learning algorithm has two obvious disadvantages in codebook designing. In this paper,the CPN learning algorithm is modified. Then a new codebook designing algorithm referred to as the Fast Competitive Learning and Error Correction Algorithm(FCLECA) and a model of fast vector quantizer based on the modified CPN are presented. The key steps and parameters in the FCLECA are also discussed. The results of simulating experiments show that the FCLECA generates high-quality codebooks while the training time is greatly reduced,so it can be used to implement fast VQ effectively.