本論文係有關中文扭曲文件影像之還原與文字切割，相關的研究工作包含四大部份：
第一部分為數位文件影像二值化演算法之軟體設計。
第二部分為二值化中文扭曲文件影像還原演算法之軟體設計，將扭曲或傾斜的中文文件影像予以校正還原，它包含了偵測本文行與字、偵測文字基線與歪斜校正等相關演算法之軟體設計。
第三部分為中文文件影像切割之軟體設計，利用物件連通標記演算法將文字做切割，以便於做後續的字體正規化與文件影像分析，並產生相關文字切割座標檔。
第四部分為中文扭曲文件影像還原與文字切割之可靠度評估與執行效能，實驗證明在各種文件影像中，扭曲文件影像還原與文字切割都能運作良好。
　　整體而言，本論文係完成一個中文扭曲文件影像還原與文字切割相關之演算法與軟體設計，並經由各種不同的文件影像實驗，證實本論文所發展的演算法有極佳的校正效果，此可以改善後續的文件影像分析系統之辨識正確率。

This thesis is about recovery and character segmentation of Chinese warped document images. The research work includes four parts：
The first part is about software design of the binarization algorithm for digital document images.
The second part is about software design of recovery from warped document images. To restore the distorted or skewed Chinese document images this software includes the following algorithms such as text line and word detection, word baseline estimation, and skew correction.
The third part is about software design for segmenting the Chinese document images. The characters are segmented by using the connected component labeling algorithm. This would make character normalization and document image analysis easier. Finally, a coordinate file for individual characters will be generated.
The fourth part is about the evaluation of reliability and run-time performance of recovery and character segmentation of Chinese warped document images. Experiments over various kinds of Chinese warped document images have shown that recovery and character segmentation can be operated very well.
On the whole, this thesis has accomplished the related algorithms and software design of recovery and character segmentation of Chinese warped document images. After being verified by various kinds of document images, the algorithm developed in this thesis has shown very good performance in skew correction and can improve the recognition rate of the document image analysis system used later.

[1] 徐千洋著,Linux函式庫參考手冊,旗標出版股份有限公司,民國九十年四月初版。

[2] 翁智賢,影像處理演算處理器之軟/硬體整合設計與實現,國立台灣科技大學碩士學位論文, 民國九十六年。

[3] 許志豪,影像二值化演算處理器之軟/硬整合設計與實現,國立台灣科技大學碩士學位論文,民國九十八年。

[4] 李皇達,中文文件影像分析系統之設計與實現, 國立台灣科技大學碩士學位論文,民國九十八年。

[5] 李俊毅,文件傾斜校正系統之設計與實現, 國立台灣科技大學碩士學位論文,民國九十九年。

[6] 黃耀陞,文件影像旋轉演算處理器之軟/硬整合設計與實現,國立台灣科技大學碩士學位論文,民國九十九年。
[7] B. Fu, M.Wu, R. Li,W. Li, and Z. Xu, “A model-based book de-warping method using text line detection,” In 2nd Int. Workshop on Camera-based Document Analysis and Recognition, Curitiba, Brazil, 2007.

[8] B. Gatos, I. Pratikakis, and K. Ntirogiannis, “Segmentation Based Recovery of Arbitrarily Warped Document Images,” Ninth International Conference on Document Analysis and Recognition, Volume 2, 23-26 Sept. 2007, pp. 989 – 993

[9] Linda G. Shapiro and George C. Stockman, Computer Vision, Prentice Hall, 2001.

[10] U.V. Marti and H. Bunke, “Using a statistical language model to improve the performance of an HMM-based cursive handwriting recognition system”, Int. Journal of Pattern Recognition and Artifical Intelligence, 15(1), 2001, pp. 65– 90.

[11] A. Masalovitch and L. Mestetskiy, “Usage of continuous skeletal image representation for document images de-warping,” In 2nd Int. Workshop on Camera-based Document Analysis and Recognition, Curitiba, Brazil, 2007.

[12] W. Niblack, An Introduction to Image Processing, Prentice-Hall, En-
glewood Cliffs, NJ, 1986, pp. 115-116.

[13] Rafael C. Gonzalez and Richard E. Woods﹒Digital Image Processing, 2nd ed., Addison-Wesley Longman Publishing Co., Inc. Boston, MA, USA. 2001.

[14] Robert M. Haralick and Linda G. Shapiro, Computer and Robot Vision, Vol.1, Addision Wesley Publishing Company, 1992.

[15] J. Sauvola and M. Pietikainen, ”Adaptive document Image binarizat-
ion,” Pattern Recognition, 33, 2000, pp. 225-236.

[16] F. Shafait and T. M. Breuel, “Document Image Dewarping Contest,” In 2nd Int. Workshop on Camera-based Document Analysis and Recognition, Brazil, 2007, pp. 181-188.