視覚障害者のための生成AIを用いた図表認識に関する一考察

Read full text Summarise with VIRAT
Authors: Yashio Uemura, Takenori Noguchi, Kaede Ishihara
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Advances in accessible e-books have addressed many barriers to text-based reading for the visually impaired, yet the interpretation of diagrams continues to be a hurdle due to the lack of implemented alternative text. Consequently, users are increasingly turning to Generative AI as a solution. This research clarifies the actual usage patterns of the visually impaired through a survey and evaluates the performance of Generative AI in describing various diagrams. We propose a structured approach for the indep endent reading of diagram-heavy documents. While the potential of Generative AI is substantial, the study identifies key limitations and highlights the necessity of developing tutorials to enhance user skills.

Fields of Study & Subjects

Digital Accessibility for Disabilities Assistive Technology in Communication and Mobility Tactile and Sensory Interactions

Publication details

Journal / source
Institutional Repositories DataBase (IRDB)
Volume
106
Pages
1–7
Publication date
31 January 2026
Language
JA