一个印刷体汉字识别系统的设计  

Design for a Printed Chinese Character Recognition System

作　　者：于秀云[1] 

机构地区：[1]曲阜师范大学数学与计算机科学系

出　　处：《计算机工程与设计》1998年第5期29-35,共7页Computer Engineering and Design

摘　　要：给出一个印刷体汉字识别系统的设计方案，它主要包括扫描输入、模糊增强与聚类分割、图象数据二值化、通过并行神经网络进行汉字匹配等四个步骤。针对汉字识别的特点，采用了一些人工智能领域的新算法。实验结果表明，系统基本符合实用性要求。In this paper a printed Chinese characters recognition system is proposed It consists of four key parts,namely,scanner inputting,fuzzy enhancement and cluster segmentation,image data binarization, and parallel artificial neural net based pattern matching Considering the specialities of Chinese characters recognition,some new algorithms are employed in the AI field Experimental implementation result shows that the proposed system is practical

关 键 词：模式识别 汉字识别 自然语言处理 印刷体汉字 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]