影像二值化在許多應用上，例如：光學文字辨識系統，車牌辨識系統以及銀行自動檢測系統等，扮演著重要的步驟。而近年來這些系統已實現到許多可攜式裝置應用上，例如：智慧型手機、相機、筆記型電腦等。而因為可攜式裝置的記憶體空間與CPU效能是受到限制的。因此，減低系統的運算量是一個重要的目標。在這篇論文中，我們提出一個基於智慧型區塊偵測之高效率影像二值化演算法。基於文字的特性，影像可分割為許多不同大小的區塊。之後，在每個區塊中建構一個閥值平面，得到最後的二值化影像。

　　在實做方面，我們將所提出的演算法與四種不同演算法包含Otsu 演算法、Niblack演算法、Sauvola演算法、Gatos演算法相互比較。我們用四種不同類型的影像作為測試圖。四種類型分別由論文，報紙，雜誌與書信文取得。首先，分析比較各演算法在相同解析度與不同解析度的執行時間。再者，比較在相同解析度與不同解析度所得到二值化影像的品質。實驗結果顯示，我們方法可以得到快速的執行時間且二值化的品質也都比其他現存的方法來得好。

　　本篇論文的貢獻在於提出一個降低運算量與增強二值化品質的影像二值化演算法。值得注意的是所提出的演算法可達到與全域演算法一樣快的執行速度，並且比其他區域演算法的二值化品質好。因此，本篇論文所提出的演算法可以毫無困難地應用在記憶體空間與CPU效能受到限制的嵌入式系統上。

Document image binarization plays an important role in many applications such as optical character recognition, automatic bank check processing, and vehicle license recognition. In this master thesis, an efficient binarization algorithm with intelligent block size detection is presented. Based on the image characteristic, the document image is automatically divided into several blocks with various sizes. Then, a threshold surface is constructed to derive the binary image.

In our experiments, various document images that taken from papers, newspapers and magazines are used as the database. The five thresholding algorithms including Otsu's, Niblack's, Sauvola's, Gatos's and the proposed algorithms are implemented for comparison. We firstly analyze the processing time to prove our approach is efficient. Afterwards, we measure the recognition rate to show the proposed method has very high performance. Experimental results show that the proposed method can provide promising binarization result and also considerably faster than other existing methods.

The major contribution is that the proposed approach can increase the binarization quality and reduce the processing time. It is noted that the proposed approach can not only achieve high speed as well as the global method, but high quality over the local methods. Consequently, the proposed algorithm can readily implement in hardware or realtime processing systems.

[1] Abbyy, www.finereader.com.
[2] Hitachi mobile phone with ocr., http://forbes.com/technology/feeds/ infoimaging/2005/03/18/infoimagingasiapulse_2005_03_18_ix 4427-0236-.html.
[3] Pantech&curitel ocr mobile phone., http://www.3g.co.uk/PR/Jan2005/8935.htm.
[4] Quicktextscan ocr software., http://jsscomputing.com/quicktextscan/ind
ex.html.
[5] H. Al-Yousefi and S.S. Udpa, Recognition of arabic characters, IEEE Transactions on Pattern Analysis and Machine Intelligence 14 (1992), no. 8, 853-857.
[6] K. S. Bae, K. K. Kim, Y. G. Chung, and W. P. Yu, Character recognition system for cellular phone with camera, Proceedings of the 29th Annual International Computer Software and Applications Conference, vol. 1, 2005, pp. 539-544.
[7] M. Cheriet, J.N. Said, and C.Y. Suen, A formal model for document processing of business forms, Proceedings of the Third International Conference on Document Analysis and Recognition, vol. 1, 1995, pp. 210-213.
[8] , A recursive thresholding technique for image segmentation, IEEE Trans-
actions on Image Processing 7 (1998), 918-921.
[9] B. Gatos, I. Pratikakis, and S. J. Perantonis, Adaptive degraded document image binarization, Pattern Recognition 39 (2006), 317-327.
[10] Rafael C. Gonzalez and Richard E. Woods, Digital image processing, Prentice Hall, 2002.
[11] H.A. Hegt, R.J. Haye, and N.A. Khan, A high performance license plate recognition system, IEEE International Conference on Systems, Man, and Cybernetics, vol. 5, 1998, pp. 4357-4362.
[12] A. Jain, Fundamentals of digital image processing, Prentice Hall, Englewood Cliffs, 1989.
[13] A.K. Jain and Yu Bin, Automatic text location in images and video frames,
Proceedings Fourteenth International Conference on Pattern Recognition, 1998,
pp. 1497-1499.
[14] G. Johannsen and J. Bille, A threshold selection method using information measures, Proceedings Sixth International Conference Pattern Recognition, 1982,
pp. 140-143.
[15] J. N. Kapur, P. K. Sahoo, and A. K. C.Wong, A new method for gray-level picture thresholding using the entropy of the histogram, Computer Vision, Graphics, and Image Processing 29 (1985), 273-285.
[16] M. Laine and O. S. Nevalainen, A standalone ocr system for mobile camera-
phones, The 17th Annual IEEE International Symposium on Personal, Indoor
and Mobile Radio Communications, 2006.
[17] Y. Liu and S. N. Srihari, Document image binarization based on texture features, IEEE Transactions on Pattern Analysis and Machine Intelligence 19 (1997), 533-540.
[18] W. Niblack, An introduction to digital image processing, Pretice-Hall, Englewood Cliffs, NJ, 1986.
[19] N. Otsu, A threshold selection method from gray-level histograms, IEEE Transactions on Systems, Man, and Cybernetics SMC-9 (1979), no. 1, 62-66.
[20] P. K. Sahoo, S. Soltani, A. K. C.Wong, and Y. C. Chen, A survey of thresholding techniques, Computer Vision, Graphics and Image Processing 41 (1988), 233-260.
[21] J. Sauvola and M. Pietikainen, Adaptive document image binarization, Pattern Recognition 33 (2000), 225-236.
[22] M. Sezgin and B. Sankur, Survey over image thresholding techniques and quantitative performance evaluation, Journal of Electronic Imaging 13 (2004), no. 1,146-165.
[23] C. Y. Suen, L. Lam, and D. Guillevic, Bank check processing system, International Journal of Imaging Systems and Technology 7 (1996), 392-403.
[24] O. D. Trier and A. K. Jain, Goal-directed evaluation of binarization methods, IEEE Transactions on Pattern Analysis And Machine Intelligence 17 (1995), no. 12, 1191-1201.
[25] O. D. Trier and T. Taxt, Evaluation of binarization methods for document
images, IEEE Transactions on Pattern Analysis And Machine Intelligence 17
(1995), no. 3, 312-315.
[26] W. Tsai, Moment-preserving thresholding: A new apporach, Computer Vision,
Graphics, and Image Processing 29 (1985), 377-393.