Digital Document Processing

Cover -- Contents -- Preface -- Contributors -- 1 Reading Systems: An Introduction to Digital Document Processing -- 1.1 Introduction -- 1.2 Text Sensing -- 1.3 Sensor Scope -- 1.4 Sensor Grid -- 1.5 Pre-processing -- 1.6 Invariance to Affine Transforms -- 1.7 Invariance to Ink-Trace Thickness -- 1.8 Shape Features -- 1.9 Processing Type -- 1.10 Computing Architecture -- 1.11 Computing Strategy -- 1.12 Knowledge Base -- 1.13 Cognitive Reliability -- 1.14 Response in Case of Difficult Input -- 1.15 Classification Accuracy -- 1.16 Energy and Mental Concentration -- 1.17 Processing Speed -- 1.18 Volume Processing -- 1.19 Summary of Human Versus Machine Reading -- 1.20 Conclusion -- References -- 2 Document Structure and Layout Analysis -- 2.1 Introduction -- 2.1.1 Physical Layout and Logical Structure -- 2.2 Pre-processing -- 2.2.1 Noise Removal -- 2.2.2 Foreground Detection -- 2.2.3 Skew Correction -- 2.3 Representing Document Structure and Layout -- 2.4 Document Layout Analysis -- 2.4.1 Approaches to Layout Analysis -- 2.5 Understanding Document Structure -- 2.6 Performance Evaluation -- 2.7 Handwritten Document Analysis -- 2.7.1 On-Line Documents -- 2.8 Summary -- References -- 3 OCR Technologies for Machine Printed and Hand Printed Japanese Text -- 3.1 Introduction -- 3.2 Pre-Processing -- 3.2.1 Text Line Segmentation and Skew Correction -- 3.2.2 Character Segmentation of Japanese and English Mixed Text -- 3.2.3 Nonlinear Normalization of Character Size -- 3.2.4 Slant Correction -- 3.2.5 Segmentation-Recognition of Handwritten Japanese Document -- 3.3 Feature Extraction -- 3.3.1 Chain Code Histogram -- 3.3.2 Gradient Feature [16] -- 3.4 Classification -- 3.4.1 Linear Classifiers for High-Speed Pre-Classification -- 3.4.2 Quadratic Classifiers -- 3.5 Dimension Reduction -- 3.5.1 Discriminant Analysis -- 3.5.2 Principal Component Analysis -- 3.6 Performance Evaluation of OCR Technologies -- 3.6.1 Used Database -- 3.6.2 Dimension Reduction -- 3.6.3 Comparison on Pre-Classification -- 3.6.4 Comparison on Main Classification -- 3.7 Learning Algorithms -- 3.7.1 GLVQ -- 3.7.2 Performance Evaluation -- 3.8 Conclusion -- References -- 4 Multi-Font Printed Tibetan OCR -- 4.1 Introduction -- 4.2 Properties of Tibetan Characters and Scripts -- 4.3 Isolated Tibetan Character Recognition -- 4.3.1 Character Normalization -- 4.3.2 Feature Formulation -- 4.3.3 Classifier Design -- 4.4 Tibetan Document Segmentation -- 4.4.1 Text Line Separation -- 4.4.2 Character Segmentation -- 4.5 Experiment Results -- 4.5.1 Performance of Isolated Character Recognition Algorithm -- 4.5.2 Performance of Segmentation Algorithm -- 4.6 Summary -- References -- 5 On OCR of a Printed Indian Script -- 5.1 Introduction -- 5.2 Origin and Properties of Indian Scripts -- 5.3 Document Pre-Processing -- 5.3.1 Binarization -- 5.3.