【摘要】 通过对Konhonen自组织神经网络设计矢量量化码书的分析研究,提出了一种基于粗糙集理论分类的神经网络图像压缩编码方法。利用粗糙集理论的强大的定性分析能力,结合图像块样本的时域、频域特性对其进行分类,并在分类能力不变的情况下,通过知识约简,去除冗余的属性,导出问题的分类规则,从而得到更好的分类图象压缩编码效果。更多还原

【Abstract】 By analyzing the codebooks of vector quantization designed with Konhonen self-organizing neural network,this paper provides a new approach on image compression based on neural network based on rough sets classification.Rough sets theory has powerful capability of qualitative analysis.Considering characteristics of image blocks in time and frequency domain,we classify the image blocks and reduce classification rules without changing the ability of classification.By eliminate the redundant properties,we educe the classifying rule of the concerned issue so as to get better classifying image compression effect.更多还原