首页> 外国专利> Method to automatically recognize the Arabic text

Method to automatically recognize the Arabic text

机译:自动识别阿拉伯文字的方法

摘要

PROBLEM TO BE SOLVED: To provide a method for automatically recognizing an Arabic text.;SOLUTION: The method includes the steps of: constructing an Arabic corpus including Arabic text files written in various styles and a ground truth corresponding to each of the Arabic text files; associating a style index with the Arabic text file and storing the style index; digitizing a line of an Arabic character to form an array of pixels; dividing the line of the Arabic character into line images; forming a text feature vector on the basis of the line images; using the Arabic text files and the ground truth in the Arabic corpus according to the style index to train a hidden Markov model; and supplying the text feature vector to the hidden Markov model to recognize the line of the Arabic character.;COPYRIGHT: (C)2015,JPO&INPIT
机译:解决的问题:提供一种自动识别阿拉伯文本的方法;解决方案:该方法包括以下步骤:构建阿拉伯语语料库,其中包括以各种样式编写的阿拉伯语文本文件以及与每个阿拉伯语文本文件相对应的基本事实;将样式索引与阿拉伯文本文件相关联并存储样式索引;数字化阿拉伯字符的一行以形成像素阵列;将阿拉伯字符的线条划分为线条图像;根据线条图像形成文本特征向量;根据样式索引,使用阿拉伯文本文件和阿拉伯语料库中的地面实况来训练隐马尔可夫模型;并将文本特征向量提供给隐藏的Markov模型以识别阿拉伯字符的行。;版权所有:(C)2015,JPO&INPIT

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号