中國清朝皇帝時常在閱覽過不少宮廷收藏之水墨畫作後，在畫作空白處用毛筆寫作題文 (墨字)、以及蓋上自己的御用印章 (紅字)，用來表達自身的感想，但此舉大大影響原作的藝術價值、以及後人觀賞原作的不適觀感。為了能保留畫作內容的完整性，本研究在不同色彩空間中使用2種影像修復演算法來自動移除水墨畫上的部分圖案，結合Python 程式語言與 OpenCV 函式庫，去偵測中式水墨畫上面的特定位置，並用不同色彩空間 (CIELAB 和YCrCb) 的提取方式去做比較，提取後的結果當作遮罩，最後進行繪畫影像的還原與修復 (移除紅色印章或墨色毛筆字，以及青綠山形的提取)，使用的修復演算法包含運用逐像素填充法的「快速匹配法 (Fast Marching Method)」和利用逐區塊填充法的「基於範例影像修復法 (Exemplar-Based Image Inpainting)」。本研究設計成自動提取圖畫的紅色、墨色或青綠色區域，把提取的結果當作影像修復用的遮罩，並利用兩種色彩空間去比較較為合適的遮罩，來證明不管是什麼修復方式都需要採用較佳的色彩空間，才能讓色彩取樣達到較好效果。

It is observered that the emperors of Qing Dynasty often wrote inscriptions and put their own royal seal on the blank spaces of the famous Chinese brush paintings in order to express their personal feelings. However, this kind of behavior greatly affected the artistic value of the original work and the uncomfortable perception of the original work for future generations. In order to preserve the integrity of the content of the paintings, this research evaluated the performances of two image inpainting algorithms (Fast Marching Method and Exemplar-Based Image Inpainting) applying in two color spaces (CIELAB and YCrCb) to remove some specific patterns on the paintings. By using Python + OpenCV program to detect specific position on the ancient paintings in CIELAB or YCrCb color spaces, and the extracted results was used as the inpainting mask. Finally, the painting images were restored and repaired (including the removals of the red seal or ink brush writing, and the extraction of the green mountain shape). The test restoration algorithms include Fast Marching Method using pixel-based filling approach and Exemplar-Based Image Inpainting using patch-based filling approach. This research proved that no matter what the restoration methods were adopted, the better color space was needed.

[1] M. Bertalmio, G. Sapiro, V. Caselles, and C. Ballester, “Image inpainting,” in Proc. ACM Conf. Comp. Graphics (SIGGRAPH), New Orleans, LA, pp. 417-424, July 2000.
[2] T. F. Chan, J. Shen, “Non-texture inpainting by curvature–driven diffusions,” J.Vis. Comm. Image Rep., vol. 4, no. 12, pp. 436-449, 2001.
[3] C. A. Z. Barcelos, M. A. Batista, “Image restoration using digital inpainting and noise removal,” Image and Vision Computing, vol. 25, no. 1, pp. 61-69, Jan. 2007.
[4] A. Telea, “An image inpainting technique based on the Fast Marching Method,” Journal of Graphics Tools, vol. 9, no. 1, pp. 23-34, 2004.
[5] T. F. Chan, J. Shen, “Mathematical models for local nontexture inpaintings,” SIAM Journal on Applied Mathematics, vol. 62, no. 3, pp. 1019-1043, 2002.
[6] J. Shen, S. H. Kang, and T. F. Chan, “Euler’s elastica and curvature-based inpainting,” SIAM Journal on Applied Mathematics, vol. 63, no. 2, pp. 564-592, 2003.
[7] A. Tsai, A. Yezzi, and A. S. Willsky, “Curve evolution implementation of the Mumford-Shah functional for image segmentation, denoising, interpolation, and magnification,” IEEE Trans. Image Process., vol. 10, no. 8, pp. 1169-1186, 2001.
[8] S. Esedoglu, “Digital inpainting based on the Mumford-Shah-Euler image model,” European Journal of Applied Mathematics, vol.13, no. 4, pp. 353-370, 2003.
[9] R. Bornard, E. Lecan, L. Laborelli, and J. H. Chenot, “Missing data correction in still images and image sequences,” in ACM Multimedia, France, Dec. 2002.
[10] A. Criminisi, P. Perez, and K. Toyama, “Region filling and object removal by Exemplar-Based image inpainting,” IEEE Trans. Image Process., vol. 13, no. 9, pp. 1200-1212, Sept. 2004.
[11] M. Bertalmio, L. Vese, G. Sapiro, and S. Osher, “Simultaneous structure and texture image inpainting,” IEEE Trans. Image Process., vol. 12, no. 8, pp. 882-889, Aug. 2003.
[12] X. Shao, Z. Liu, and H. Li, “An image inpainting approach based on the Poisson equation,” Second International Conference on Document Image Analysis for Libraries, DIAL, art. no. 1612979, pp. 368-372, 2006.
[13] L. L. Zhao, L. Shen, and R. C. Hong, “Survey on image inpainting research progress,” Computer Science, vol. 48, no. 3, Mar. 2021.
[14] N. Otsu, “A threshold selection method from gray-level histograms,” IEEE Trans. Man and Cybernetics, vol. 9, no.1, pp. 377-393, 1979.
[15] T. Jiang, S. Y. Chen, Q. Li, K. L. Zhang, and Y. Chang, “Image inpainting of wind tunnel test based on Otsu method and FMM,” Acta Aeronautica et Astronautica Sinica, vol. 41, no. 2, 2020.
[16] H. S. Chen, S. H. Chen, Y. H. Chao, M. R. Luo, and P. L. Sun, “Applying image-based color palette for achieving high image quality of displays,” Color Research and Application, vol. 39, no. 2, pp. 154-168, 2014.
[17] 李立宗 編著，科班出身的 AI 人必修課：OpenCV 影像處理,使用 python，深智數位股份有限公司，臺北市，2019。
[18] 黃日鋒、詹文鑫、陳鴻興、胡國瑞、徐道義、孫沛立、羅梅君 合著，陳鴻 興 編審，顯示色彩工程學，全華圖書股份有限公司，新北市，2011。
[19] 魏碩廷、陳鴻興、徐明景、李文淵、謝翠如、吳瑞卿、孫沛立 合著，陳鴻 興 總編輯，色彩新論- 從心理設計到科學應用，五南圖書出版股份有限公司，臺北市，2019。
[20] 陳鴻興、蕭琇霙 合著，圖解現代色彩學，五南圖書出版股份有限公司，臺北市，2018。
[21] 繆鵬 編著，CV+深度學習: AI最完整的跨套件Python人工智慧電腦視覺，深智數位股份有限公司，臺北市，2019。
[22] H. S. Chen, S. M. Chen, C. Y. Jiang, Y. C. Zhang, C. Y. Lin, C. E. Lin, and J. A. Lee, “Computational tongue color simulation in tongue diagnosis,” Color Research and Application, 2021.
[23] OpenCV Color conversions, https://docs.opencv.org/3.4/de/d25/imgproc_color_conversions.html, 2021.
[24] Confidence Ellipse, http://www.real-statistics.com/, 2021.
[25] 繪製二維數據集的置信橢圓, https://hg95.github.io/posts/3b0a2e38/, 2021.
[26] J. Howse, “OpenCV Computer Vision with Python,” Packt Publishing, 2013.
[27] 謝梁、魯穎、勞虹嵐 合著，廖信彥 審校，Python深度學習實作：Keras快速上手，博碩文化股份有限公司，新北市，2018。
[28] 張元翔 編著，數位影像處理 : Python程式實作，全華圖書股份有限公司，新北市，2019。
[29] 黃日鉦 編著，人工智慧與深度學習：理論與Python實踐，碁峰資訊股份有限公司，臺北市，2020。

[30] J. Minichino, J. Howse 合著，劉波、苗貝貝、史斌 編譯，OpenCV 3計算機視覺 : Python語言實現，機械工業出版社，2016。
[31] EIZO White Paper: Color matching between sRGB monitors and wide color gamut monitors, www.eizo.com.tw/, 2021.
[32] 故宮書畫典藏資料檢索系統, https://painting.npm.gov.tw/, 2021.
[33] J. A. Sethian, “A fast marching level set method for monotonically advancing fronts,” Proc. Nat. Acad. Sci. U.S.A., vol. 93, no. 4, pp. 1591-1595, 1996.
[34] R. Catanuto, G. Morabito, and S. Toumpis, “Optical routing in massively dense networks: Practical issues and dynamic programming interpretation,” in Wireless Communication Systems. ISWCS '06. 3rd International Symposium on, pp. 83-87, 2006.
[35] M. Kalantari, M. Shayman, “Routing in wireless Ad Hoc networks by analogy to electrostatic theory,” IEEE International Conference on Communications, Paris, France, vol. 7, pp.4028-4033, 2004.
[36] Image Thresholding, https://docs.opencv.org/4.x/d7/d4d/tutorial_py_thresholding.html, 2021.
[37] A. A. Efros, T. K. Leung, “Texture synthesis by nonparametric sampling,” in IEEE Int. Conf. Computer Vision, Corfu, Greece, pp. 1033-1038, Sept. 1999.
[38] A. Arash, G. R. Raquel, “Color Conversion in Deep Autoencoders,” Society for Imaging Science and Technology, pp. 89-98, Oct. 2021.
[39] S. C. Pei, Y. C. Zeng, and C. H. Chang, “Virtual restoration of ancient Chinese paintings using color contrast enhancement and lacuna texture synthesis,” IEEE Trans. Image Process., vol. 13, no. 3, pp. 416-429, Mar. 2004.
[40] S. C. Pei, Y. M. Chiu, “Background adjustment and saturation enhancement in ancient Chinese paintings,” IEEE Trans. Image Process., vol. 15, no. 10, pp. 3230-3234, Oct. 2006.
[41] J. Long, E. Shelhamer, and T. Darrell, “Fully convolutional networks for semantic segmentation,” IEEE Conf. Computer Vision and Pattern Recognition, pp.3431-3440, 2015.
[42] Image segmentation, https://en.wikipedia.org/wiki/Image_segmentation, 2021.