@article{oai:ipsj.ixsq.nii.ac.jp:00232412, author = {守岡, 知彦 and Tomohiko, Morioka}, issue = {2}, journal = {情報処理学会論文誌}, month = {Feb}, note = {甲骨文字や小篆,隷変以降の歴史的な楷書等の歴史的漢字字体と現代の漢字字体を対象にした複数の包摂粒度を許容する漢字構造記述を行うための手法について述べる.漢字は文字数が多く形状の歴史的変遷もあり部品や構造が認定しにくかったり一意に決めにくいものも存在する.そこで,複数の漢字の圏の間の自然変換に基づく部品と漢字構造の整理と,漢字構造の自動変換に基づく手法を試み,CHISE文字オントロジに適用した., This paper describes a method for describing Hanzi structures that allows multiple glyph-granularities for historical glyphs, such as, Oracle-Bone script, Small-Seal script and historical Kaishu styles, as well as modern Hanzi glyphs. Hanzi has a large number of characters and their glyphs have changed over time, making the components and structures of some characters difficult to recognize or ambiguous. Therefore, we attempted a model to organize components and Hanzi structures based on natural transformations between multiple Hanzi categories, introduced automatic conversion of Hanzi structures, and applied them to the CHISE character ontology.}, pages = {340--347}, title = {通時的な多粒度漢字構造情報記述の試み}, volume = {65}, year = {2024} }