摘要： 简要介绍了基于自组织特征映射（SOFM）神经网络的图像压缩的传统算法。通过对传统方法的优缺点分析，提出了一种新的简单的矢量量化压缩方法。新算法采用分类码书设计和残留编码，大大提高了图像的客观指标和主观视觉效果。实验表明此方法明显优于传统的SOFM算法，而且易于硬件实现。

关键词: 矢量量化, 自组织特征映射, 神经网络, 分类码书

Abstract: The traditional algorithm of image compression based on the self-organizing feature map neural network is introduced. A new and simple compression algorithm based on the vector quantization is put forward by analyzing of traditional algorithm. The new algorithm, including the sorting codebook design and the remains coding, greatly improves the objective level and the subjective visual impression. Experimental results show that the new method is better than the traditional SOFM algorithm and can be realized easily by hardware.

中图分类号: 

• TN911.73