Research On Chinese Braille Translation System Based On Out-of-Vocabulary Ary Detection

With the development of Internet information,blind people should break the shackles of the Braille books and enjoy rich and colorful information.However,at present,the degree of Chinese Braille information processing is low,and the reading materials which blind people can access are deficient.In order to help blind people to get information as early conveniently,this paper designs and implements a Chinese Braille translation system with automatic detection of Out-of-Vocabulary(OOV).The system includes modules such as OOV detection,Chinese word segmentation,word-tone conversion,pinyin-ASCII conversion,and Braille sequence generation and so on.This system provides two sets of translation modes:current Chinese Braille and the National Common Braille.It also supports decoding and translating standard BRF files,which can help users to quickly and easily obtain Braille texts.Using with Braille display devices greatly expand the information channels of blind people.This article mainly contains the following parts:According to the rules of the Chinese Braille word segmentation,the analysis found that the most influential problem is the OOV problem,so the research work on the automatic detection of OOV is carried out.Combining the degree of internal cohesion and the degree of external freedom,this paper proposes an unsupervised OOV detection method based on improved PMI algorithm and minimum branch entropy combining strategy.Through theoretical and experimental analysis,the algorithm can generate a personalized OOV dictionary for different texts,and does not require long-term learning and training to adjust parameters,and has a certain improvement in the accuracy and recall rate of detection.After fully analyzing the characteristics of the traditional matching word segmentation and the Jieba system,the independent words detected by the algorithm are formed into a personalized OOV dictionary,which is used to improve the existing Jieba system,and the improved Jieba system is tested.The performance of word segmentation has been improved,which also proves the effectiveness of the OOV detection algorithm.By loading the personalized OOV dictionary,the effect of Chinese Braille word segmentation can be improved.The process of converting Chinese characters into pinyin firstly uses the Pypinyin module to initially obtain the pinyin strings of Chinese characters,and to segment the initials,finals and tones of the pinyin strings.After fully analyzing the difference between the Braille spelling system and the Chinese character pinyin system,the corresponding processing algorithms are designed for the special pronunciation.After completing the pinyin correction processing for the Chinese braille translation,the obtained initials,finals,and tones are respectively matched with the Braille ASCII code,and finally the ASCII code is converted into the Braille sequence.The above algorithms are developed and implemented separately,and the functions are integrated to form a complete Chinese Braille translation system.The design and development of the corresponding human-computer interaction interface is convenient for users.Finally,the multi-dimensional testing of the Chinese Braille translation system is carried out,including system function testing,Chinese Braille translation algorithm performance testing,and application testing of translation result files.The translation results of this system are compared with the translation results of the Braille Online Platform of China to verify that the system has high translation accuracy.The translation time is counted,and calculated that the system has a fast translation speed,the overall has good practicability.After the Braille expert check,the translation result file also has a good performance on the Braille display devices.