Digital text archiving for cursive writing documents by using the reading voice of which the experts read aloud.

2023 Fiscal Year Final Research Report

• Project/Area Number: 21K18372
• Research Category: Grant-in-Aid for Challenging Research (Exploratory)
• Allocation Type: Multi-year Fund
• Review Section: Medium-sized Section 2:Literature, linguistics, and related fields
• Research Institution: Ritsumeikan University
• Principal Investigator: Akama Ryo 立命館大学, 文学部, 教授 (70212412)
• Co-Investigator(Kenkyū-buntansha): 西浦 敬信 立命館大学, 情報理工学部, 教授 (70343275) ; 山路 正憲 立命館大学, 衣笠総合研究機構, 研究員 (00899049)
• Project Period (FY): 2021-07-09 – 2024-03-31
• Keywords: 形態素解析 / UNIDIC / 古典籍 / 古文書 / 翻刻 / 翻刻支援

Outline of Final Research Achievements

While automated transcription technology using AI image recognition is advancing in the digital archiving of transcribed historical books and documents, many of the people who have the ability to decipher ancient documents are elderly and often have difficulty with keyboard input. In this study, we developed a system to record the "manuscript" of the transcribed text by having a person with deciphering ability read the document aloud. We were able to incorporate the system using the ARC's old book Portal database as the system infrastructure, enabling the system to record and store the read-out data. As with image recognition, however, the rate of correct answers as a text mixed with kanji and kana is not high, and at present the system serves only as a "rough draft" and is only practical as a manuscript for proofreading.

Free Research Field

日本文学

Academic Significance and Societal Importance of the Research Achievements

本来日本語で書かれている古典籍や古文書を解読しようというモチベーションは、多くの人が持っており、そうしたモチベーションを活用したSNS型翻刻サイトとして「みんなで翻刻」は、著名なシステムとなっている。一方、オンラインでの翻刻作業を苦手とするが、長年の経験により解読能力が極めて高い人材を、古い文献の解読プロジェクトの協力メンバーとして参加してもらうことで、翻刻効率は大幅に向上できる。コンピュータシステムを使って翻刻本文アーカイブを試みる若手研究者と高い能力を有する高齢者とのコラボレーションを有効に機能させるシステムができたことは大きな意義があろう。