Fine-tuning Large Language Models for Automatic Font Skeleton Generation: A First Attempt

• Item type: SIG Technical Reports(1)
• 公開日: 2024-05-08
• タイトル: Fine-tuning Large Language Models for Automatic Font Skeleton Generation: A First Attempt
• タイトル / 言語: en
• タイトル: Fine-tuning Large Language Models for Automatic Font Skeleton Generation: A First Attempt
• 言語: eng
• キーワード / 主題Scheme: Other
• キーワード / 主題: セッション1(PRMU)
• 資源タイプ / 資源タイプ識別子: http://purl.org/coar/resource_type/c_18gh
• 資源タイプ: technical report
• 著者所属: Tokyo Institute of Technology
• 著者所属: Google Research
• 著者所属: Tokyo Institute of Technology
• 著者所属: Tokyo Institute of Technology
• 著者所属(英): en
• 著者所属(英): Tokyo Institute of Technology
• 著者所属(英): en
• 著者所属(英): Google Research
• 著者所属(英): en
• 著者所属(英): Tokyo Institute of Technology
• 著者所属(英): en
• 著者所属(英): Tokyo Institute of Technology
• 著者名: Yuxuan, Liu; Yasuhisa, Fujii; Xinru, Zhu; Kayoko, Nohara
• 著者名(英): Yuxuan, Liu; Yasuhisa, Fujii; Xinru, Zhu; Kayoko, Nohara
• 論文抄録 / 内容記述タイプ: Other
• 論文抄録 / 内容記述: Despite the pivotal role that font skeletons could play in the typeface research and font design, the availability of font skeleton data is sparse and limited. This research explores the possibility of using Large Language Models (LLMs) to generate font skeleton data based on font outline data. By fine-tuning GPT-3.5 with a dataset of 8,213 font outlines and corresponding skeletons, a preliminary LLM model for font skeleton generation has been achieved. Both quantitative and qualitative evaluations were performed to assess the effectiveness of the fine-tuned model, including Rasterized Pixel Distance, Chamfer Distance and visual analysis. This research has established a foundation for the automatic generation of font skeletons using LLMs, setting the stage for future work on automatic skeleton generation and the wider application of font skeletons in typography.
• 論文抄録(英) / 内容記述タイプ: Other
• 論文抄録(英) / 内容記述: Despite the pivotal role that font skeletons could play in the typeface research and font design, the availability of font skeleton data is sparse and limited. This research explores the possibility of using Large Language Models (LLMs) to generate font skeleton data based on font outline data. By fine-tuning GPT-3.5 with a dataset of 8,213 font outlines and corresponding skeletons, a preliminary LLM model for font skeleton generation has been achieved. Both quantitative and qualitative evaluations were performed to assess the effectiveness of the fine-tuned model, including Rasterized Pixel Distance, Chamfer Distance and visual analysis. This research has established a foundation for the automatic generation of font skeletons using LLMs, setting the stage for future work on automatic skeleton generation and the wider application of font skeletons in typography.
• 書誌レコードID / 収録物識別子タイプ: NCID
• 書誌レコードID / 収録物識別子: AA11131797
• 書誌情報: 研究報告コンピュータビジョンとイメージメディア（CVIM）; 巻 2024-CVIM-238, 号 4, p. 1-6, 発行日 2024-05-08
• ISSN / 収録物識別子タイプ: ISSN
• ISSN / 収録物識別子: 2188-8701
• Notice: SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc.
• 出版者 / 言語: ja
• 出版者: 情報処理学会