中國水墨人物畫之合成

Abstract

相對於傳統西畫的「寫實」，中國水墨畫重「寫意」，所謂「筆簡意遠」，就是以最少的筆墨表達最多的意境。而水墨人物畫更是以線寫形，種種輕重剛柔的線條組成一幅人物畫，運用線描來表達「畫意」，因此，如何讓人物畫的各種筆法，透過電腦來表達，就是本篇論文的重點：我們提出模組化合成水墨人物畫的方法。經由對現有水墨人物畫筆法、風格之分析，建構一個各種物件模組的筆法資料庫，再參考使用者的構圖，或是現有畫作、照片的佈局，由電腦組合各種模組快速合成一幅水墨人物畫；如此，使用者便可以充分利用已知的各種筆法，節省實際作畫的繁瑣與時間，創造出更多種可能性的作品。

In contrast with the Western paintings, Chinese ink paintings lay stress on the “implicit meanings”. In other words, painters try to use as few strokes as possible to express their deepest feelings. However, ink figure paintings even more rely on various linear strokes to convey the meaning of paintings. In this thesis, we try to express ink figure paintings of different styles by modularized synthesis. We analyze conventionalized strokes of traditional ink figure paintings and set up a database of various components with different styles. Then, according to a reference image or figure, we can synthesis an ink figure painting. By allowing users to reuse all kinds of conventionalized strokes, we can save the tedium and time consumed by real ink painting, and also make the creation with more and more possibility.