水墨畫之影像修補技術

溫宗興
With the rise of Chinese worldwide, traditional Chinese art works have become much more valued than before. Many works of renowned painters have been steadily poured into the world’s major auctions, and that their values often hit a new record high. Nonetheless, when we take a look at the recent Chinese history, those frequent wars and invasions from superpower countries have made preservation of these paintings quite a challenge. Although many precious artistic works have been luckily survived from such a disturbing period; the marks left on them by age and wars, however, cannot be easily erased.

To effectively pass down and also to document those valuable images, scholars and collectors have been using scientific methods to preserve them digitally. This thesis has chosen Chinese paintings as the subject of the study and proposed a two-layer digital image inpainting technology that offers an automatic restoration and completion for damaged images. This thesis has chosen Chinese paintings the subject and proposed a quick and effective image inpainting algorithm. This algorithm will break down an inpainting region into two layers; the skeleton layer and the texture layer. For the skeleton layer, this thesis proposes a new skeleton inpainting algorithm which uses an edge map to restore the skeleton part of the target region. For the texture layer, this thesis adds the result from the skeleton layer to partition or add constraints to the target region and then apply the image inpainting. The change of target region results in modification of searching region, such that fewer errors would occur in the inpainting image, and there would be a higher speed for the inpainting process as well. Last, the previously added skeleton will be removed and filled-in. Experimental results have demonstrated that the proposed method has a good performance in image completion for Chinese paintings.

隨著全球華人的崛起，傳統的中國藝術品也逐漸受到重視，在近年來各大世界級的拍賣會場上，不難發現有許多著名畫家的真跡不斷被釋出，其成交金額更是屢創新高；不過回首中國近代歷史，在長年的戰爭與列強的侵略下，讓這些畫作的保存增添了許多的困難。因此，許多珍貴的作品雖幸運的留存下來，但卻無法擺脫歲月與戰火的摧殘，為了有效傳承並紀錄這些珍貴的圖像，現今的學者與收藏家利用科學方法將其作數位的典藏；本論文以水墨畫數位影像為目標，提出二階段式的數位影像修補技術，能將破損的部份自動予以修補成完整的影像。本論文中我們針對水墨畫影像修補實施研究，提出一個快速且有效的影像修補演算法，本演算法主要將修補步驟區分成skeleton layer與texture layer兩階段；在skeleton layer中，提出一個新的結構修補演算法，運用edge map重建目標區中的結構部份。在texture layer中，首先加入skeleton layer的結果，使目標區內各物件的區塊能先加以分割或限制，接著運用範例式影像修補技術實施影像修補，在此步驟時，本論文改良了原演算法中的搜尋範圍，使修補結果能減少錯誤發生並加快修補速度，最後再將加入的結構移除並予以填補。從實驗結果中可證明，本論文所提出的方法擁有強健的影像重建及修補能力，對受損的水墨畫影像可有效實施修復。