受損漢字的修復之研究

A Study on the Reconstruction of Damaged Chinese Characters

陳東儒(Tung-Ju Chen)

指導教授 ： 施仁忠

國立交通大學/資訊學院/多媒體工程研究所/碩士(2018年)

摘要

書法是中國幾千年來的傳統藝術之一，與其有關的作品包含書信、銘文、屏風、室內裝飾等。大多數具有歷史的石碑與拓文在數千間，都有遭受到或多或少的風化或人為破壞。在本文中，我們提出了一個兩階段方法來處理掃描這些歷史作品所得到的影像。此方法可以將漢字形狀，從這些受到損壞的掃描影像中重建出來。

對於包含損壞漢字字符的影像，我們首先從含有雜訊的輸入影像中，將與漢字字符相關的部分提取出來。然後從中找出斷裂的筆劃，並將它們連接在一起以重建該字符。最後產生一張具有重建後漢字字符的二元影像。我們將提出的方法應用於五種不同銘文或拓文上的各個字符，以展示我們的結果。

關鍵字

影像處理 ； 文字提取 ； 文字修復

並列摘要

Calligraphy is one of the traditional arts in China for thousands of years. These works are in the form of letters, inscriptions, folder screens, indoor decorations, etc. For thousands of years, most historical stone inscriptions and stone rubbings are suffered from weathering or man-made destruction. In this thesis, we propose a two-stage approach to process the images scanned from these works. It reconstructs the shape of a Chinese character from these images which keep the damage.

Given an image containing a damaged Chinese character, we first extract the Chinese character part from the noisy input image. Then, we find broken strokes and connect them together, in order to reconstruct the character. Finally, a binary image containing the reconstructed Chinese character is generated. We demonstrate our results by applying our approach to characters of five different inscriptions or stone rubbings.

並列關鍵字

Image processing ； Character extraction ； Character reconstruction

參考文獻

[1] A. Ben-Hur, D. Horn, H. T. Siegelmann, and V. Vapnik. Support vector clustering. J.Mach. Learn. Res., 2:125–137, Mar. 2002. ISSN 1532-4435. URL http://dl.acm.org/citation.cfm?id=944790.944807.

[2] D. Bradley and G. Roth. Adaptive thresholding using the integral image. Journal ofgraphics tools, 12(2):13–21, 2007.

[3] J. E. Bresenham. Algorithm for computer control of a digital plotter. IBM Systems Journal,4(1):25–30, 1965. ISSN 0018-8670. doi: 10.1147/sj.41.0025.

[4] J. Canny. A computational approach to edge detection. IEEE Transactions on PatternAnalysis and Machine Intelligence, PAMI-8(6):679–698, Nov 1986. ISSN 0162-8828.doi: 10.1109/TPAMI.1986.4767851.

[5] R. Cao and C. L. Tan. A model of stroke extraction from chinese character images. InProceedings 15th International Conference on Pattern Recognition. ICPR-2000, volume 4,pages 368–371 vol.4, 2000. doi: 10.1109/ICPR.2000.902935.