摘要
利用计算机技术识别《四库全书》文字的研究起步较晚,本文在分析前人研究成果的基础上,基于并行遗传算法,把古体汉字识别问题表达为一类带约束条件的组合优化问题.运用特征属性进行编码,采用并行遗传算法中的适应度函数优化和迁移策略,较好地改善了搜索性能.实验结果表明:并行遗传算法能有效地提高文字识别质量.
The use of computer technology to identify the characters of Si Ku Quan Shu started late.Based on the analysis of previous research results,the problem of ancient Chinese character recognition was expressed as a class of combinatorial optimization problems with constraints by parallel genetic algorithm.Coding with characteristic attributes,the fitness function optimization and migration strategy in parallel genetic algorithm were used to improve the search performance.The experimental results showed that the parallel genetic algorithm could effectively improve the recognition quality.
作者
马海云
张忠林
MA Hai-yun;ZHANG Zhong-lin(School of electronic information and electrical engineering,Tianshui Normal University,Gansu Tianshui 741001,China;School of Electronic and Information Engineering,Lanzhou Jiaotong University,Gansu Lanzhou 730070,China)
出处
《中央民族大学学报（自然科学版）》
2019年第3期9-12,共4页
Journal of Minzu University of China(Natural Sciences Edition)
基金
甘肃省自然科学基金项目(No.18JR3RE245)
关键词
古典文献汉字识别
属性特征
遗传算法
classical Chinese character recognition
attribute characteristics
genetic algorithm
