摘要
随着多媒体资源的飞速增长,多媒体检索技术成为当前检索技术的一个热点,该文将介绍文档图像这类特殊的多媒体资源检索技术中的特征提取方法.根据汉字特点,从几个角度,介绍了几种汉字图像块的特征提取方法及相应的相似度算法,并进行了试验测试,试验结果表明几种方法可行.
As the multimedia resources increasing rapidly, multimedia retrieval technology has become the focus of present retrieval technique. This article will introduce the feature-extracting of Document Image which is a special kind of multimedia. From a few aspects, this article introduces the approaches of Chinese character image feature-extracted and the corresponding similarity algorithms. All the methods were tested, the experimental results show that all the methods are feasible.
出处
《江西师范大学学报（自然科学版）》
CAS
北大核心
2008年第2期138-141,共4页
Journal of Jiangxi Normal University(Natural Science Edition)
基金
863计划重点项目(2006AA010105)
国家自然科学基金项目(60772081)
北京市属市管高校人才强教计划项目(PXM2007-014224-044677,PXM2007-014224-044676)
北京市教委科技发展计划项目(KM200710772010)
关键词
文档图像
汉字特征提取
特征的相似度计算
document-image
Chinese character feature-extracting
features similarity computation
