臺灣博碩士論文加值系統

在這篇論文中，我們利用連接元件的幾何特性為基礎，設計了一個在名片辨識系統上，做為版面分析與產生文字行的系統。對於產生文字行方面，我們針對於塗抹所容易造成的缺點去分析。以連接元件為基礎，用Split-and-Merge的方式去產生文字行。並利用連接元件與文字行的幾何特性去提升文字行產生的正確率。用以降低名片辨識系統因為文字行產生錯誤所造成的辨識與後處理的錯誤。
在版面分析部分，我們利用連接元件的幾何特性為基礎來分析名片。我們利用全形與半形的連接元件大小以及各種語言的書寫方式定義出特徵向量。接著我們利用特徵向量的幾何特性，如向量的長度和方向性，來將名片版面歸類到橫式橫寫、橫式直寫、橫直混合、直式直寫、直式橫寫、直式混合六大類中。並依據分類結果提高文字行產生的正確率。
在文字行產生方面，我們測試了六大類共720張名片。文字行分析正確的有98%。而在版面分析方面，我們利用本論文所提方式，於484張名片中，將82.4%名片的版面正確分析出來。

In this thesis, we design a layout analysis and textline generation system on business card recognition based on geometric properties of connected components. In the textline generation, we analyze the disadvantages of smearing and propose a modified method base on connected components. We use the recursive splitting and merging to generate textlines. Then we use the geometric properties of connected components to increase the accuracy of textline generation, reduce the time of post-processing, and decrease the mistakes which many made by erroneous textline generation from the business card recognition system.
In layout classification, we analyze the business cards based on the geometric properties of connected components. We use the full size and half size of connected components and the writing style of languages to define the characteristic vectors. Then we classify the layout of the business cards based on the geometric properties of characteristic vectors, such like the length and direction of characteristic vector. We classify the business cards into six styles: horizontal writing of horizontal style, vertical writing of horizontal style, mixed writing of horizontal style, horizontal writing of vertical style, vertical writing of vertical style, mixed writing of vertical style, and use the information of style to increase the accuracy of textline generation.
In textline generation, we test 720 business cards in six styles. The correctness of textline generation is 98%. In layout classification, we test 484 business cards in six styles, and the average accuracy of layout classification is 82.4%.

CHAPTER 1. INTRODUCTION 1
1.1 Motivation 1
1.2 Problems Analysis 3
1.3 Assumptions 4
1.4 Tasks 5
1.5 System Description 6
1.5.1.1 Textline Generation 7
1.5.1.2 Textline Adjustment 7
1.5.1.3 Layout Classifications 8
1.6 Survey of Related Research 9
1.7 Thesis Organization 11
CHAPTER 2. TEXTLINE GENERATION BASED ON THE GEOMETRIC PROPERTIES OF CONNECTED COMPONENTS 12
2.1 Introduction 12
2.2 Problems Analysis 12
2.3 Proposed Method 16
2.4 Feature Used 19
2.4.1 The major size of connected components 19
2.4.2 The character height and width in textlines 20
2.4.3 The inter-character distance of connected components 20
2.4.4 The deviation angle from the linking line 20
CHAPTER 3. TEXTLINE ADJUSTMENT BASED ON THE GEMORTIC PROPERTIES OF TEXTLINES 23
3.1 Motivation 23
3.2 Problems Analysis 24
3.3 Proposed Method 27
3.4 Feature Used 30
3.4.1 The direction of each textline in blocks 30
3.4.2 The inter-textline distance in blocks 31
3.4.3 The average width and height of characters in textlines 32
3.4.4 The inter-character distance in textlines 33
CHAPTER 4. COMPONENT-BASED LAYOUT CLASSIFICATION 34
4.1 Introduction 34
4.2 Problem Analysis 38
4.3 Proposed Method 39
4.4 Feature Used 42
4.4.1 Direction of characteristic vectors 43
4.4.2 Length of characteristic vectors 43
4.4.3 Amount of characteristic vectors 43
CHAPTER 5 EXPERIMENTAL RESULTS AND ANALYSIS 44
5.1 Introduction 44
5.2 Textline Generation 44
5.2.1 Result of textline generation 46
5.3 Layout Classification 46
5.3 Layout Classification 47
5.4 Error Analysis 49
5.4.1 Textline Generation 49
5.4.2 Layout Classification 50
CHAPTER 6. CONCLUSION AND FUTURE WORKS 51
6.1 Conclusion 51
6.2 Future Works 52
REFERENCES 53

REFERENCES
[1] W. Niblack, An Introduction to Digital Image Processing, pp.115-116, Prentice Hall, 1986.
[2] E. G. Johnston, “Short note: Printed text discrimination,” Comput. Graphics Image Processing, vol. 3, no. 1, pp. 83-89, 1974.
[3] G. Nagy, “A prelimitary inverstigation of techniques for automated reading of unformatted text,” Comm, ACM, vol 11, no. 7, pp. 480-487. 1968.
[4] L. Abele, F. Wahl, and W. Scheri, “Procedures for an automatic segmentation of text graphic and halftone regions in document,” in Proc. 2nd Scandinavian Conf. Image Analysis, 1981, pp. 177-182.
[5] G. Ciardiello, M. T. Degrandi, M. P. Poccotelli, G. Scafuro, and M. R. Spada, “An experimental system for office document handling and text recognition,” in Proc. 9th Int. Conf. Pattern Recognition, 1988, pp. 739-743.
[6] T. Akiyama and N. Hagita, “Automated entry system for printed documents,” Pattern Recogn., vol. 23, no. 11, pp. 1141-1153, 1990.
[7] J. Kreich, A. Luhn, G. Maderlechner, “Knowledge based interpretation of scanned business letters,” in Proc. IAPR Workshop on CV, 1988, pp. 417-420.
[8] J. Toyoda, Y. Noguchi, and Y. Nishimura, “Study of extracting Japanese newspaper article,” in Proc. 6th Int. Conf. on Pattern Recogn., 1982, pp. 1113-1115.
[9] A. C. Downton and C. G. Leedham, “Preprocessing and presorting of envelope images for automatic sorting using OCR,” Pattern Recogn., vol. 23, no. 3/4, pp. 347-362, 1990.
[10] L. A. Fletcher and R. Kasturi, “ A robust algorithm for text string separation from mixed text/graphics images,” IEEE Trans. Pattern Anal. Machine Intell., vol. 10, no. 6, pp. 910-918, 1988.
[11] H. Kato and S. Inokuchi, “The recognition system for printed piano music using musical knowledge and constraints,” in Proc. SSPR90, 1990, pp. 231-248.
[12] J. H. Boose, “A survey of knowledge acquisition techniques and tools,” Knowledge acquisition, vol. 1, no. 1, pp. 3-37, 1989.
[13] O. Iwaki, H. Kida, and H. Arakawa, “A segmentation method based on office document hierarchical structure.” In Proc. IEEE int. Conf. Syst. Man. Cybernetics, Alexandria, VA, Oct, 1987, pp. 759-763.
[14] G. Nagy, S. C. Seth, and S. D. Stoddard, “Document analysis with an expert system,” in E. S. Gelsema and L. N. Kanal, Eds., Pattern Recogn. Practice II. New York: Elsevier, 1986, pp.149-159.
[15] Y. Tsuji, “Document image analysis for generating syntactic structure description,” in Proc. 9th Int. Conf. Pattern Recogn., 1988, pp. 744-747.
[16] D. Wang and S. N. Srichari, “Classification of newspaper image blocks using texture analysis,” CVGIP, vol. 47, pp. 327-352, 1989.
[17] J. Toyoda, Y. Noguchi, and Y. Nishimura, “Study of extracting Japanese newspaper article,” in Proc. 6th Int. Conf. on Pattern Recogn., 1982, pp. 1113-1115.
[18] P. S. Yeh, S. Antoy, A. Litcher, and A. Rosenfeld, “Address location on envelopes,” Pattern Recogn., vol. 20, no. 2, pp. 213-227, 1987.