首页 | 本学科首页   官方微博 | 高级检索  
     

中文文本图像倒置快速检测算法
引用本文:曾凡锋,张国锋,陈侃.中文文本图像倒置快速检测算法[J].计算机工程与设计,2012,33(9):3512-3516.
作者姓名:曾凡锋  张国锋  陈侃
作者单位:1. 北方工业大学信息工程学院,北京,100144
2. 华南理工大学自动化科学与工程学院,广东广州,510641
基金项目:十一五国家科技支撑平台重点基金项目(2009BA171B02);北京市属高等学校人才强教计划基金项目(PHR201007121)
摘    要:针对图像处理(如OCR技术)对图像方向要求十分严格,文本图像方向具有不确定性的问题,提出了中文文本图像倒置快速检测算法.利用投影技术定位出文本字符,结合中文字符及标点符号结构特征,筛选出文本图像中的标点符号,根据标点符号像素分布特点判断出类型,结合标点符号的使用习惯,采用统计的方法判断中文文本图像是否倒置.实验结果表明,投影方法可以不用基于内容达到高效快速的要求,利用统计方法可以保证判别率,该方法可用于OCR预处理过程.

关 键 词:文本图像  字符结构  投影算法  文本定位  图像倒置

Chinese text image inversion fast detection algorithm
ZENG Fan-feng , ZHANG Guo-feng , CHEN Kan.Chinese text image inversion fast detection algorithm[J].Computer Engineering and Design,2012,33(9):3512-3516.
Authors:ZENG Fan-feng  ZHANG Guo-feng  CHEN Kan
Affiliation:1.College of Computer Sciences,North China University of Technology,Beijing 100144,China; 2.College of Automation Science and Engineering,South China University of Technology,Guangzhou 510641,China)
Abstract:Given the facts that the image processing technology(such as OCR) is strict with image direction and text image direction is uncertain,this paper put forward a rapid inversion detection algorithm for Chinese text image.Firstly,locate the text characters by using the projection technique.Select the punctuation marks in the text image considering Chinese characters and punctuation structure features.Then,judge the type of punctuation marks according to their pixel distribution characteristics and using habits.Finally,decide inversion of Chinese text image by using statistical method.The experimental results show that the projection technique can achieve fast and efficient requirements without content-based processing,and the statistical method can guarantee the discrimination rate for inversion detection.This method can be used in OCR preprocessing.
Keywords:document image  character structure  projection algorithm  text location  image inversion
本文献已被 CNKI 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号