Proposal of a translation system for Braille such as class materials to support volunteers of blind students

2019 Fiscal Year Final Research Report

• Project/Area Number: 17K01097
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Educational technology
• Research Institution: Kinjo University
• Principal Investigator: Shimomura Yuko 金城大学, 社会福祉学部, 教授 (70171006)
• Project Period (FY): 2017-04-01 – 2020-03-31
• Keywords: 視覚障害学生支援 / 墨字翻訳

Outline of Final Research Achievements

Our aim is to provide Braille students with the necessary Braille books promptly for blind students. For that, we used the Python language to convert it to Braille grammar. The conversion from Japanese to Braille was done using a parsing tool MeCab. However, the morpheme segmentation of MeCab was inaccurate, and we examined other methods that are not syntactic tools. As a result, we decided to create a translation program using AI Deep Learning. The data and analysis results used so far can be used as AI training data. Although the AI system we created is still part, the translation result was 99.86% correct. We will continue our research and aim to complete the system.

Free Research Field

福祉情報工学

Academic Significance and Societal Importance of the Research Achievements

覚障害学生にとって点字教科書や参考書はどうしても必要なものであるが、ソフト開発業者は収益が上がらないソフトウエアは開発しない。間違わずに点訳できるボランティア熟練者は少ない。点字文法は複雑で難しいうえに、時々変更される。熟練者は高齢であり、どんどん減っていく。コンピュータで100％翻訳させることが理想であるが、まだ行われていない。そのため今回は日本語形態素解析ツールやAIのDeep Learningを用いた。最新ITを用いることで、今までより、より100％に近づけることができた。