自工業革命後，各式自動化機器逐漸取代人力，大量生產之物品雖然價廉，惟使用完全相同外型之產品卻無法表現個人獨特風格，如何適切地於數量、單價及獨特性間取得較佳權衡組合，為一值得探討之課題。
以電腦所配置之標準字型做為溝通媒介並無法如手寫信箋般傳達通訊者獨有之特徵，雖可搭配有限表情符號及圖片供使用者選用，俾利於字裡行間適時表達情緒，但制式化之符號或圖案並不能表現如手寫文字般獨特之多元特徵，無法兼顧字型輸入、顯示、傳送之便利性及個人化之獨特風格。本論文針對前揭使用電腦標準字型缺失，經由擷取手寫字元特徵，據以變更標準字型筆畫長度、斜率及字根長寬比例，產生獨一無二個人化之字型供使用者使用。
中文字型之個人化處理共可細分為標準字元之處理、手寫字元之分析及特徵對應三步驟，先透過字元、字根及筆畫所組成之階層結構，將電腦配置之標準中文字集中每一字元均拆解為組成字元之字根、構成字根之筆畫線段與形成筆畫之輪廓數目、控制點數目及控制點相對關係，再於使用者書寫預設數目之指定中文字元後，擷取手寫字元之各組成字根長寬比例、構成字根之筆畫長度及斜率等特徵，最後依所擷取手寫字元特徵，據以修改標準中文字集中組成每一字元之字根長寬比例、構成字根之筆畫長度及斜率，將標準中文字集之字元轉換成具有個人手寫特徵之字元，產生獨一無二之個人化中文字型。

With the advent of industrial revolution, machines replace the traditional human power. Even though both the quantity of merchandise produced increases and per unit cost decreases dramatically, the stereotypical outlook of mass-produced goods drives the human innate quest for recognizable traits to differentiate among otherwise homogeneous products. A font personalizing system is proposed to transform standard Chinese vector font, according to a writer’s handwritten characteristics, into a personalized font possessing writer’s discernible features.
In the proposed font personalizing system, a radicals-contours-line segments-control points four-level hierarchical approach is taken in analyzing a Chinese character. Both standard and handwritten characters are decomposed into constituent radicals first. The number of contours of every radical, the number of connected line segments making up each contour and the control points forming each line segment are then derived. The resulting ensemble of information is later used for finding the radical correspondence between standard characters and handwritten ones. Once the correspondence is established, the height-width ratio of handwritten radicals, together with the length and slope of line segments are utilized to modify the matching counterparts in all characters of a standard Chinese font set. In this manner, the representing features extracted from only a limited number of handwritten characters can be used to automatically modify a full set of Chinese fonts. The personalized font set obtained can efficiently replace existing standard fonts to exhibit unique handwritten features.

第一章 序論 - 1 -
1.1前言 - 1 -
1.2個性化商品 - 2 -
1.3字型發展 - 9 -
1.4相關研究 - 10 -
1.5研究簡介 - 16 -
第二章 理論基礎 - 18 -
2.1正交投影法 - 18 -
2.1.1理論 - 18 -
2.2 Bezier curve - 19 -
2.3 True Type Font 簡介- 21 -
第三章 研究方法與步驟 - 30 -
3.1前置處理 - 30 -
3.1.1字元分類 - 30 -
3.1.2筆畫線段分類 - 32 -
3.1.3 字元拆解 - 35 -
3.1.4 字元拆解特性 - 37 -
3.2 特徵擷取 - 80 -
3.3 特徵套用 - 84 -
第四章 研究結果 - 85 -
第五章 結論 - 100 -
參考文獻 - 101 -

[1]陳水淋,“中文字型個性化系統之研究,”國立台灣大學碩士論文,1995.
[2]張皓傑,“個性化字型之筆劃特徵抽取系統,” 國立台灣大學碩士論文,1996.
[3]F. Lin and X. Tang, “Offline handwritten Chinese character stroke extraction,” Proc. Int. Conf. Pattern Recognition, pp. 249-252, 2002.
[4]X. Gao, L. Jin and J. Yin, “A New Stroke-based Directional Feature Extraction Approach for Handwritten Chinese Character Recognition,” Proc. 6th Int. Conf. Document Analysis and Recognition, pp.635-639, 2001.
[5]H. Chien, T.C. Chuang and F. Kuo, “Italic detection and rectification,” Proc. Int. Conf. Digital Object Identifier, pp. 530-533, 2005.
[6]S. Watanabe, T. Furuhashi, K. Obata and Y. Uchikawa, “A study on feature extraction using a fuzzy net for off-line signature recognition,” Proc. Int. Conf. Digital Object Identifier, pp. 2857-2860, 1993.
[7]S. Yeung and S. Fong, “An off-line Approach to Extract Natural Strokes from Handwritten Chinese Characters,” Proc. Int. Conf. Man and Cybernetics, Vol.2, pp.1933-1938, 1995.
[8]H.P. Chiu, D.C. Tseng, “A feature-preserved thinning algorithm for handwritten Chinese characters,” Proc. Int.. Conf. Pattern Recognition, pp. 235-239,1996
[9]H. Lee and C. Huang, “Fuzzy Feature Extraction on Handwritten Chinese Characters,” Proc. 3th Int. Conf. on Fuzzy Systems, pp.1809-1814, 1994.
[10]Y. Liu, L. Zhang and J. Tai, “A new approach to on-line handwritten Chinese character recognition,” Proc. Int. Conf. Document Analysis Recognition, pp.192-195, 1993.
[11]X. Li, W. Oh, J. Hong and W. Gao, “Recognizing Components of Handwritten Characters by Attributed Relational Graphs with Stable Features,” Proc. 4th Int. Conf. Document Analysis and Recognition, pp.616-620, 1997.
[12]R. Cao and C.L. Tan, “A model of stroke extraction from Chinese character images,” Proc. Int. Conf. Pattern Recognition, pp.368-371,2000.
[13]T. Qu and A. Adler, “Dynamic Signature Verification System Using Stroked Based Features,” Proc. Int. Conf. Audio and Visual Environments and Their Applications, pp.83 - 88, 2003.
[14]C. Hsu, F. Chen, C. Chang and S. Jeng, “On-line Chinese signature verification based on multi-expert strategy,” Proc. Int. Conf. Security Technology, pp. 169-173, 1998.
[15]J. Lin and G. Li,“Off-Line Chinese Signature Verification,”Proc. Int. Conf. Image Processing , pp. 205-207, 1996.
[16]H. Cha and S. Lee, “Establishing Handwriting Individuality Using Pattern Recognition Techniques, ”Proc. Int. Conf. Document Analysis and Recognition, pp. 1195-1204,2001.
[17]Computer font, http://en.wikipedia.org/wiki/Outline_font
[18]The OpenType Font File, http://www.microsoft.com/typography/OTSPEC/otff.htm