Scholarlink

The global rise in manga’s popularity has fueled interest in automated manga translation, but a lack of English-translated datasets limits research in text recognition, machine translation, and creative domain language processing. To address this, an English-translated manga dataset was developed using a third-party API to translate Japanese text into English. The process involved text extraction, preprocessing and re-integrating translations into panels with contextual accuracy. Translation quality, assessed by using BERTScore for semantic similarity, showed strong alignment with human-translated benchmarks. This dataset supports advancements in automated manga translation, cross-lingual language modeling, and digital localization.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Automated Translation of Manga Comic Strips: Applying Deep Learning Techniques to Bridge Language Barriers in Japanese Comics

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Aditya Gupta,
  • Ayushi Gupta,
  • Yajas Arora,
  • Ali Imam Abidi

摘要

The global rise in manga’s popularity has fueled interest in automated manga translation, but a lack of English-translated datasets limits research in text recognition, machine translation, and creative domain language processing. To address this, an English-translated manga dataset was developed using a third-party API to translate Japanese text into English. The process involved text extraction, preprocessing and re-integrating translations into panels with contextual accuracy. Translation quality, assessed by using BERTScore for semantic similarity, showed strong alignment with human-translated benchmarks. This dataset supports advancements in automated manga translation, cross-lingual language modeling, and digital localization.