基于小波变换在强干扰条件下印刷体汉字处理研究  

RESEARCH ON PROCESS OF INTENSIVELY DISTURBED TYPOGRAPHY CHINESE CHARACTER BASED ON WAVELET TRANSFORM

作　　者：唐亮[1] 胡运发[2] 张文龙[1] 

机构地区：[1]上海师范大学计算机科学与工程系 [2]复旦大学计算机与信息技术系

出　　处：《计算机应用与软件》2005年第8期46-47,111,共3页Computer Applications and Software

摘　　要：本文提出了在有相当强烈的干扰条件下,基于Haar小波变换的印刷体汉字处理的一种新方法。此方法只对反映原始字符信号概貌的小波变换低频部分保留,同时忽略包含大量干扰信号的高频部分,由此达到剔除大部分干扰信号的效果。实验结果表明本文提出的方法具有良好的去干扰特性,为汉字处理提供了新的思路。In this paper,a novel approach based on Haar wavelet transform is proposed to process typography Chinese characters which are intensively disturbed, This approach addresses eliminating mostly of signals disturbing characters by only reserving low-frequency portion of result of wavelet transform which carries general picture of original character, at the same time, ignoring high-frequency portion of result of wavelet transform which carries mostly of disturbing signals. The result of experiment shows that the approach proposed by this paper has the feature eliminating intensively-disturbing easily and effectively, and it provides Chinese character processing with new idea.

关 键 词：印刷体汉字处理 小波变换 字符处理 MALLAT算法 干扰信号 

分 类 号：TP391.12[自动化与计算机技术—计算机应用技术]