细化算法在手写体字符识别中的应用  

Application of Thinning Algorithms in Handwriting Character Recognition

作　　者：李永江[1] 潘保昌[1] 郑胜林[1] 

机构地区：[1]广东工业大学信息工程学院,广东广州510006

出　　处：《现代电子技术》2008年第12期91-92,100,共3页Modern Electronics Technique

摘　　要：在手写体字符识别的研究中,在提取特征的过程中需要提取字符的骨架信息,所以要对每一个字符进行细化处理。在研究ZS和LW细化算法的基础上,采用串行细化与并行细化相结合的方法,并引入一些模板改进算法使其能够有效去除多余枝杈和冗余像素的问题;通过对手写体数字、字母的文字进行测试,表明该细化效果较好,并有一定的实用价值。In the study of handwriting recognition, getting the skeleton information is needed in the process of extracting features,so we should thin every character before that. Based on the study of ZS and LW thinning algorithms, combine the serial thinning with the parallel thinning methods and bring some template to improve the thinning algorithm and make it successfully solve the problems such as needless trees and useless pixel. After testing in handwriting number and letters,it shows that the algorithm has better effect and is worth for application.

关 键 词：骨架 连通 模板 串行细化 并行细化 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]