透過您的圖書館登入
IP:3.17.70.182
  • 學位論文

水墨畫模擬:渲染部位清淡描跡與留白之處理

Rendering of Slight Detailing and Blanking over Washed Areas in Chinese Brush Paintings

指導教授 : 林正中

摘要


傳統水墨畫乍看只有黑與白,但墨色的濃淡乾濕,以及隨筆勾勒的獨有神韻,卻可以讓觀者對於畫者想要傳達的意境,產生不同程度的體會。黑白水墨的韻味與彩墨的繽紛都是觀看水墨畫的樂趣。 本論文提出一個將影像轉換為水墨畫風格的模擬系統,輸入為一張彩色參考圖像,輸出為黑白水墨結果及彩色水墨結果。主要步驟如下: 首先,計算參考圖像的相對亮暗點與梯度,透過這些資訊將參考圖像渲染平滑化,並再度壓縮其間距達到去除細節的目的。 接著,為了讓結果影像不至於沉悶黯淡,並貼近水墨畫多處留白的特色。利用兩種不同高低門檻的灰階值獲得不同程度留白的影像,將高門檻低留白的影像進行白描、將低門檻高留白的影像進行描跡。 最後,將前者再度留白的影像加入後者加上的線條,再將前者灰區以後者白區為依據再度留白獲得最終影像。

並列摘要


At the first glance, traditional Chinese brush paintings may appear to be merely a drawing art form in straight black and white presentation. There are, however, far more than it may look. The shades and forms of inking manipulation and line sketching of shapes, generally revealing a taste of simplicity, spiritual unwinding, and perhaps inexactness, all in fact reflect the mind and skill mastery of the artist when creating the art work. It allows viewers to have various aspects of appreciating the artist's art works. The charms of black and white inking and the glamour of chromatic inking bring about the pleasures of enjoying Chinese brush paintings. This paper proposes an image-based simulation system for generating Chinese-brush-painting-like output. Taking a color image as input, the system generates two outputs: a black-and-white ink painting and a chromatic one. The process consists of four operational stages as follows. [1] Deriving from the input a grey-level counterpart, the map Mdiffuse is generated wherein certain initial ink-diffusing effects that is unique in the Chinese brush painting could be found; [2] further process with grey-scale compression and mean lifting on Mdiffuse results in Mink-seed, perceptually a brighter, contrast-lowered and detail-obscured version of Mdiffuse and thus revealing more flavor of ink-diffusion; [3] a dual-process for recovering the details lost during the ink-diffusing stages is underway, by which high-contrasted details in bright shades and low-contrasted features in dark shades could be brought back, and thus enhances the vividness of the painting amazingly, as would be observed in the map Mfinal; [4] by exploiting the very same process respectively on the R-/G-/B-components of a colored input image, three maps Mred_final, Mgreen_final, and Mblue_final will then be available for comprising a Chinese-brush-painting-like output in color accordingly. In all, the outcome of the simulation system is deemed as visually satisfactory in almost all test cases, with the following qualities: [1] effects of ink-diffusion appear wherever deemed as appropriate in the output; [2] restorations of the details undergo only at key positions in such a way that life-like sensations can be created with just a few final touches in the work; [3] the use of colors and shades may deviate from the input to some degrees, sometimes beyond one’s expect but causing no visual discomfort or dissonance; [4] the sizes of subjects in the painting may vary a bit without causing any visual attention. Together with all these, the output of the proposed simulation system looks just “Chinese-brush- painting-like”.

參考文獻


[1] 陳建發,「淺析『黑白概念』在傳統中國繪畫中的表現」,視覺藝術論壇,第五期,86~103頁,2010。
[2] 趙陽,胡國強,梁玉珊,兒童水墨畫初階(1)──基本篇,台灣珠海出版有限公司,1992。
[3] 陳弘逸,「以真實影像為基礎之水墨畫風格模擬」,國立交通大學,碩士論文,2017。
[4] C.-M. Wang and R.-J. Wang, “Image-Based Color Ink Diffusion Rendering,” IEEE Transactions on Visualization and Computer Graphics, vol. 13, no. 2, pp.235-246, 2007.
[5] L. Liang and L. Jin, “Image-Based Rendering for Ink Painting,” IEEE International Conference on Systems, Man, and Cybernetics, Manchester, pp. 3950-3954, 2013.

延伸閱讀