摘要
在分析传统的手写汉字投影切分算法的基础上,提出了改进算法。算法初切分阶段通过对局部图像进行水平投影法切分来确定行切分点,从而得到当行字符,之后针对单行字符,通过多种策略进行单字切分,包括针对字间间距、标点符号等信息做出的单字切分策略。实验结果表明,该算法能有效的减少传统垂直投影分割法的误切分现象,且分割速度快,实现容易。该算法经过大量实验证明了其在准确率方面的优越性。
Based on the analysis of traditional projection segmentation algorithm for handwritten chinese characters proposed the improved algorithm.Early stages,taking advantage of the local level projection to determine the line of cut-off point of of the partial image to obtain a line of characters;Then making use of a variety of strategies to obtain words,including for the spacing between words,punctuation,etc.Experimental results show that the algorithm can effectively reduce errors of the traditional vertical projection segmentation algorithm;simultaneously,the improved algorithm splits fastly and easy to implement.After a large number of experiments certify that the algorithm is better quality in accuracy rate.
作者
王序哲
曹卫
WANG Xu-zhe,CAO Wei（School of Computer Science,South-Central University for Nationalities,Wuhan 430074,China）
出处
《电脑知识与技术》
2011年第9期6205-6206,共2页
Computer Knowledge and Technology
关键词
局部图像
切分点
行和数组
列和数组
误切分
local image
segmentation point
line and array
column and array
mistakenly segmentation
