多语标识智能纠错平台的设计与实现  

Design and Implementation of Platform Correcting Multilingual Signage's Wrong Translation

作　　者：郭宁宁[1] 梁野[1,2] 阙颖[1] 

机构地区：[1]北京外国语大学计算机系,北京100089 [2]清华大学信息技术研究院,北京100084

出　　处：《电脑编程技巧与维护》2014年第6期80-82,92,共4页Computer Programming Skills & Maintenance

基　　金：北京高等学校青年英才计划项目;中央高校基本科研业务费专项资金(2010XJ017);清华大学自主科研计划;国家973重点基础研究发展计划(2011CB302302)资助

摘　　要：针对城市公共场所翻译良莠不齐的问题,提出了一个智能化的解决方案。该方案分为手机客户端和后台服务器端两部分,手机端负责对市民拍摄的海量多语言标识图进行预处理和文本识别,并将这些数据传输给服务器;服务器端能够对文本准确性进行识别,并自动纠正错误内容或递交语言专家修改,将正确结果反馈给拍摄者,同时提醒标识主管部门替换。平台的设计与实现将对图像处理、语种识别、文本定位等计算机相关领域,起到一定的示范作用。This paper put forward an intelligent solution to decrease the badly translated signage in our country’s urban public places. This solution is divided into two part： the mobile client part is responsible for preprocessing the mass pictures shot by citizens, recognizing text, as well as delivering these data to the server; the back-end server part can identify the accuracy of the text, correct wrong translation automatically or submit the sign’ s content to language experts for modifying it, and then return the corrected translation back to the photographer and remind competent authorities replacing the wrong sign at the same time. The design and implementation of platform will play a positive role in many computer-related areas, such as image processing, language recognition and text location.

关 键 词：多语标识 图像处理 文本定位 语种识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]