基于孤立点的手写数字识别算法研究  

Research of Handwritten Numeral Recognition Algorithm Based on the Outlier

作　　者：朱小艳[1] 张西芝[2] ZHU Xiao-yan, ZHANG Xi-zhi （1.Software College of Henan University, Zhengzhou 450064, China;2. Department of Information Management ,Shengda Economics Trade ＆ Management College of Zhengzhou, Xinzheng 451191 ,China）

机构地区：[1]河南大学软件学院,河南郑州450046 [2]郑州升达经贸管理学院资讯系,河南新郑450001

出　　处：《电脑知识与技术》2012年第6期3969-3972,共4页Computer Knowledge and Technology

摘　　要：在信息化飞速发展的时代，字符识别是信息录入和转化的一个重要内容。其中，手写体数字的识别有着非常广泛的应用，如邮政编码、手机手写等。该文使用一种基于“孤立点”、交叉数和中间数的手写数字识别方法，利用孤立点对数字（0—9）进行首次分类，再利用交叉数和中间数对分类后的数字进行进一步分类和识别。实验结果表明，该算法有效提高了数字识别的效率。In the age of information development, character recognition is an important content of information input and conversion.Hand- written numeral recognition has a very wide range of applications, such as postal code, mobile phone handwriting recognition. This paper presents a handwritten digital recognition method based on the＂ outlier＂, crossing number and middle number, which uses isolated point to digital （0-9） for classification, then uses crossing number and middle number for further classification and recognition. The test results show that, the algorithm can improve the efficiency of digital recognition.

关 键 词：特征提取 孤立点 交叉数 中间数 效率 

分 类 号：TP18[自动化与计算机技术—控制科学与工程;自动化与计算机技术—控制理论与控制工程]