手写识别建模数学方法研究  

A Modeling Method of Chinese's Handwriting Character Online Recognition

作　　者：黄弋石[1] 梁艳 

机构地区：[1]南通大学公共卫生学院实验中心,江苏南通226019 [2]南通浪潮通用软件有限公司,江苏南通226000

出　　处：《软件》2013年第8期13-15,共3页Software

基　　金：南通大学校级基金(03041184)

摘　　要：为了解决汉字手写识别,提出最简单的建模方法。运用独特思路,建立了拓扑学的二维图形定义的模型。在网格背景中,使用二十七种点的相互的拓扑关系定义,识别汉字基本笔画的六种分类,然后识别了汉字的每种单一连续的手写笔画,从而识别了汉字。规避了常见的二维图形学中的复杂的线条计算与判定。从而,使得手写识别变得十分轻松,而且机理简单、清晰、易懂,使得编程工作量降低到最低点。本文的方法,具有通用性,可以机动灵活的移植到其它任何民族的文字。In order to solve the handwriting recognition of Chinese characters, put forward the simplest modeling method. Using the unique way of creation, we established the deifnition of 2 d graphics model with the topology. In the grid-background, set up 27 kinds of point deifnition, used them on the mutual topological relationship of basic Chinese characters stroke of six categories, and then identiifed each single continuous handwritten strokes, to recognize the Chinese characters. Pass by the common complex 2 d graphics computing and decision. Make handwriting recognition become very easy. The mechanism is simple, clear, simple, making the programming work to the lowest cost. The method works versatile, can be lfexible to transfer for any other nation characters.

关 键 词：汉字手写识别 英文手写识别 联机识别 连笔识别 手写识别 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]