Automatic Braille Transformation System of Tex documents by computer.

1.在本研究项目中，我们制作了一个日文LaTeX文档（包括数学公式）到盲文代码的自动转换系统，作为一个Windows 95应用程序。该系统包括以下三个步骤。(1)将句子切分为词类，并将汉字转换为日文假名(2)将其转换为盲文代码和盲文文档格式(3)对转换结果进行盲文编辑器校对。为了提高上述第一步(1)的精度，我们还制作了一个包含约7000个数学专业词汇的词典，其中包括阅读、分词位置和语素编码。为了减轻盲文转换工作中数学公式输入的负担，我们研究了数学公式的光学识别系统，并表明在我们的实验系统中可以以较高的精度识别TeX打印的文档。我们成功地将这种数学公式识别算法应用于从Dvi文件中重建TeX源文件的问题。这使我们能够将我们的盲文转换系统应用于使用不同版本的各种宏包和各种用户宏命令编写的TeX文档。

1.In This research program, we made an automatic transformation system of japanese LaTeX docuents including mathematical formulas into braille codes, as a Windows 95 application. The system consists of the following three steps.(1) Segmentation of the sentences into the parts of speech and transformation of chinese characters into japanese KANA(2) Transformation into braille codes and braille document formats(3) Braille editor for proofreading of the transformation results.We alse made a dictionary of about 7000 mathematical technical words with readings, segmentation positions and morpheme codes inorder to improve the precision of the first step (1) above.2.In order to relieve the load of imputs of mathematical formulas in braille transformation works, we investigated the optical recognition system of mathematical formulas, and showd that the recognition of the documents printed by TeX is possible with high accuracy in our experimental system.We applied this recognition algorithm of mathematical formulas to the problem of reconstruction of TeX source files from Dvi files successfully.This enables us to apply our braille transformation system to TeX documents written using various macro pacages of diffenent versions and various user macro commands.