基於部件拆解之草書生成

以往進行書法字體學習時，總是需要仰賴字帖的示範，臨摹字帖給予書寫者標準答案練習，然而並非所有風格字體皆有字帖出版，尤其是草書字帖，草書字體在歸納上並不容易，且與楷書有著不同骨架，各個流派及年代的書法家字體更是有太多差異之處，收集到的資料寥寥無幾，加上漢字的發展到現今有太多合體字是古人沒有寫過的。因此我們率先提出了草書字體生成系統，利用影像處理技術生成出符合各書法家獨特風格字體且為正確寫法之草書，輔助書寫者進行草書字體學習，讓草書字體書寫者不再只是專業書法家的專利，進而使草書廣泛普及於大眾中。本論文為首先嘗試將影像處理技術應用於中國傳統藝術-草書字體之研究，透過本系統所生成的草書字體與書法家所寫幾乎無太大差異，能夠直接提供作為使用者的書寫參考，且透過本實驗所設計之使用者圖形介面能夠讓使用者更加方便操作本系統。

關鍵字

草書字體；字帖生成系統；影像處理；電腦視覺；字體美化

並列摘要

In the past, learning calligraphy often relied on model samples provided by copying characters from these models, offering practitioners a standard to practice from. However, not all styles of typefaces have published model samples available, especially for cursive script calligraphy. Cursive script poses challenges in summarization and differs significantly from regular script in its structure. Moreover, various calligraphers across different periods and schools of thought exhibit substantial differences in their writing styles. The available resources are scarce, compounded by the evolution of Chinese characters into many compound forms not documented by ancient writers. To address these challenges, we have pioneered a cursive script font generation system utilizing image processing techniques. This system generates cursive script fonts that emulate the unique styles of different calligraphers while adhering to correct writing techniques. It assists learners in studying cursive script fonts, democratizing the practice beyond the exclusive realm of professional calligraphers. Our paper represents the initial attempt to apply image processing technology to the study of traditional Chinese art – specifically, cursive script fonts. The cursive script fonts generated by this system closely resemble those written by calligraphers, providing direct reference material for users. Additionally, our user-friendly graphical interface designed for this experiment facilitates convenient operation of the system for users.