ITR: Automatic Translation of Scientific Literature to Braille

ITR：科学文献自动翻译为盲文

• 批准号: 0312487
• 负责人: Eitan Gurari
• 金额: $ 35.93万
• 依托单位: Ohio State University Research Foundation -DO NOT USE
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2003
• 资助国家: 美国
• 起止时间: 2003-07-15 至 2007-06-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0312487&HistoricalAwards=false
• 关键词: ITR; Automatic; Translation; Scientific; Literature

This project develops transcribing software to make scientific and technical literature freely and almost instantaneously available to braille readers through fully-automated translation of both MathML and LaTeX to Nemeth braille. The PI's software will be the first braille transcribing utility capable of simultaneous and automated processing of mathematical formulas, scientific text, spatial arrangements and XML-described formats. MathML is an emerging low-level standard markup language for archiving mathematical content, and for communicating and processing mathematics by machines. LaTeX is a high-level language for writing mathematical documents. Scientific word processors can export documents in LaTeX, and many authors also use it directly with text editors. A prototype, which transforms MathML to Nemeth braille mathematics, was built on top of the PI's TeX4ht, a powerful tool for automatic conversion of LaTeX documents to MathML mathematics plus XML text. This two-layer approach avoided the need to readdress numerous difficulties, already resolved in TeX4ht, which had plagued earlier attempts to convert legacy LaTeX to Braille, and also answers the call of the National Federation of the Blind to develop tools for direct translation from MathML into Nemeth Braille. The subject areas of machine translation of MathML code to other formats and of automatic generation of Nemeth braille from XML-tagged source have received little attention. The work will identify problem areas in both fields, provide solutions with respect to the braille transcribing system, and identify significant topics for further investigation.Broader Impacts: At the current going rate, it takes over six months to translate a single textbook at a cost of $5,000 and more. The PI's software will be highly portable across numerous platforms, and will be disseminated free of charge as open source in the public domain, as is already the case for LaTeX and TeX4ht. The ultimate outcome of the project should be increased productivity for the large number of people who must rely on Braille to read, which will in turn lead to increased integration of braille readers into technical professions.

该项目开发了转录软件，通过将MathML和LaTeX全自动翻译成Nemeth盲文，使盲文读者几乎可以立即免费获得科学和技术文献。PI的软件将是第一个盲文转录实用程序，能够同时自动处理数学公式、科学文本、空间安排和XML描述的格式。MathML是一种新兴的低级标准标记语言，用于归档数学内容，以及用于机器通信和处理数学。LaTeX是一种编写数学文档的高级语言。科学文字处理器可以以LaTeX格式导出文档，许多作者也直接在文本编辑器中使用它。一个将MathML转换为Nemeth盲文数学的原型是在PI的TeX4ht上构建的，TeX4ht是一个强大的工具，用于将LaTeX文档自动转换为MathML数学和XML文本。这种双层方法避免了重新考虑在TeX4ht中已经解决的许多困难，这些困难曾困扰着早先将遗留乳胶转换为盲文的尝试，并响应了全国盲人联合会的号召，即开发工具将MathML直接翻译成Nemeth盲文。MathML代码到其他格式的机器翻译以及从带XML标记的源代码自动生成Nemeth盲文的主题领域几乎没有受到关注。这项工作将找出这两个领域的问题领域，提供关于盲文转录系统的解决方案，并确定需要进一步研究的重要课题。广泛影响：按照目前的速度，翻译一本教科书需要六个多月的时间，成本在5,000美元或更多。PI的软件将在许多平台上高度可移植，并将作为开放源代码在公共领域免费分发，就像LaTeX和TeX4ht已经的情况一样。该项目的最终结果应该是提高大量必须依靠盲文阅读的人的生产率，这反过来又将导致盲文读者更多地融入技术专业。