基于分形理论的字符切分方法研究  

Study of character segmentation based on fractal theory

作　　者：窦育强 范海菊[1] 

机构地区：[1]河南师范大学,453007

出　　处：《中国科技信息》2008年第21期41-41,43,共2页China Science and Technology Information

基　　金：河南省自然科学基金资助(0511011500)

摘　　要：文字图像由于受噪声、随机污点的干扰,以及图像边框的影响,其字符很难精确分割,直接影响后续的特征提取以及识别工作。为解决上述问题,本文提出采用分形理论进行字符切分,仿真结果证明此方法和传统的基于上下轮廓凹凸特征的分割方法相比,可以有效抑制噪声特别是图像边框干扰,提高了切分的正确率。Because of the influences on character image such as noise、random spot and image frame,it was difficult to divide characters accurately.So,followed work of parameter extraction and character recognition were influenced directly.In order to solve aforementioned problems,this paper brought forward a method which introduced fractal theory to divide characters.Simulation results proved that when comparing with the traditional way based on parameters of top and bottom concavo-convex outline,this method could effectively suppress noises especially for image frame influencing.Consequently,the accuracy of character segmentation was improved.

关 键 词：图像边框 分形理论 字符切分 

分 类 号：TP391.43[自动化与计算机技术—计算机应用技术] O18[自动化与计算机技术—计算机科学与技术]