Research and Developing in Teaching Environment using Board Writing for the Visually Impaired

首先，我们开发了手写字符识别系统，该系统在Windows个人电脑的虚拟黑板上工作。然而，该系统的输入区域设置在个人电脑的显示器上，它工作在一个真正的黑/白电路板上，附着一个传感元件。该系统可识别汉字和数学表达式，用于高等数学/体育课程。识别速度足以满足实时使用。然而，日语文本是横向和纵向书写的，本系统只分析横向书写。而且每个日文字符都要写在一个固定的框架内，数学表达式也要写在同一个框架内。所以当我们写同样的数学表达式时，有一些限制。其次，开发了教科书、教辅书、试卷和自然科学区域版面分析的文档模型描述方法。每个区域包含数学表达式。研究了文本、图形、表格、数学表达式和文档逻辑结构等版面元素，并开发了版面识别系统。考虑到其他应用程序，我们选择XML文件作为内部文件结构和输出文件。在PDF文件日益普及的今天，我们开发了面向视障人士的科学PDF文件文本信息提取系统。在不久的将来，该技术将被应用于从PDF中提取图形，提取的图形是视障人士翻译的触觉数据。然而，前两项是研究和开发，我们与国际委员会定义了国际标准的盲文代码，称为UEBC。UEBC是在2004年4月改编的，我们考虑了与我们开发的系统和UEBC的一致性，我们的内部格式应用了一个新的代码定义。

First, we developed the hand-write character recognition system, which works on a virtual black board on the Windows personal computer. However the input area of this system set on the display of the personal computer, it works on a real black/white board which is attached a sensing element. This system recognizes Chinese character called Kanji and mathematical expressions, which is used in higher mathematic/physic education courses. The recognition speed is sufficient for real time use. However Japanese text is written in horizontal and vertical direction, this system analyzes only horizontal writing. Also each Japanese character should be written a fixed frame, the mathematical expressions are written in a same frame. So there are little restrictions when we write same mathematical expressions.Second, we developed document model description method for layout analysis about schoolbook, study-aid book and examination papers and natural science area. Each area contains mathematical expressions. We researched layout elements (text, figure, table, mathematical expressions, and logical structure of document) for layout analysis, and developed its recognizing system. We choose XML file as an internal file structure and an output file considering other application programs. Nowadays, PDF is becoming a popular digital file, so we developed extracting text information system from scientific PDF file for visually impaired. Near the future, this technology is applied for extracting figures from PDF, and extracted figures are translated tactile data for visually impaired.However, previous two items are research and developing, we defined international standard braille code called UEBC with international committee. The UEBC is adapted in April 2004, we considered consistency with our developed systems and UEBC, a new code definition is applied our internal format.