A 63-Way Character Recognition / (Record no. 8809)
[ view plain ]
| 000 -LEADER | |
|---|---|
| fixed length control field | 07178nam a22002537a 4500 |
| 008 - FIXED-LENGTH DATA ELEMENTS--GENERAL INFORMATION | |
| fixed length control field | 210112b2014 a|||f mb|| 00| 0 eng d |
| 040 ## - CATALOGING SOURCE | |
| Original cataloging agency | EG-CaNU |
| Transcribing agency | EG-CaNU |
| 041 0# - Language Code | |
| Language code of text | eng |
| Language code of abstract | eng |
| 082 ## - DEWEY DECIMAL CLASSIFICATION NUMBER | |
| Classification number | 610 |
| 100 0# - MAIN ENTRY--PERSONAL NAME | |
| Personal name | Ismail Mohamed Sobhy |
| 245 1# - TITLE STATEMENT | |
| Title | A 63-Way Character Recognition / |
| Statement of responsibility, etc. | Ismail Mohamed Sobhy |
| 260 ## - PUBLICATION, DISTRIBUTION, ETC. | |
| Date of publication, distribution, etc. | 2014 |
| 300 ## - PHYSICAL DESCRIPTION | |
| Extent | 96 p. |
| Other physical details | ill. |
| Dimensions | 21 cm. |
| 500 ## - GENERAL NOTE | |
| Materials specified | Supervisor: Mohamed A. El-Helw |
| 502 ## - Dissertation Note | |
| Dissertation type | Thesis (M.A.)—Nile University, Egypt, 2014 . |
| 504 ## - Bibliography | |
| Bibliography | "Includes bibliographical references" |
| 505 0# - Contents | |
| Formatted contents note | Contents:<br/>1 Introduction 1<br/>1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1<br/>1.1.1 Advantages Of Acquiring Text . . . . . . . . . . . . . . . . 1<br/>1.1.2 Types of Text . . . . . . . . . . . . . . . . . . . . . . . . . 2<br/>1.1.3 Text Extraction in Videos and Images . . . . . . . . . . . 2<br/>1.2 Aim of the Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . 3<br/>1.3 Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3<br/>1.4 Thesis Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4<br/>2 Background and Related Work 5<br/>2.1 Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5<br/>2.2 Natural Scene Text . . . . . . . . . . . . . . . . . . . . . . . . . . 5<br/>2.3 Text Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6<br/>2.4 The Traditional Text Extraction Pipeline . . . . . . . . . . . . . . 7<br/>2.4.1 Text Detection and Localization . . . . . . . . . . . . . . . 8<br/>2.4.1.1 Region Based Text Detection and Localization<br/>Technique: Stroke Width Transform [1] . . . . . 8<br/>2.4.1.2 Texture Based Detection: A Laplacian Method<br/>for Video Text Detection [2] . . . . . . . . . . . . 12<br/>v<br/>CONTENTS<br/>2.4.2 Text Frame Selection/Classication . . . . . . . . . . . . . 13<br/>2.4.2.1 Text Tracking: An Eective Video Text Tracking<br/>Algorithm Based on SIFT Feature and Geometric<br/>Constraint [3] . . . . . . . . . . . . . . . . . . . . 14<br/>2.4.3 Text Enhancement . . . . . . . . . . . . . . . . . . . . . . 18<br/>2.4.3.1 Text Enhancement: Edge based Binarization for<br/>Video Text Images[4] . . . . . . . . . . . . . . . . 19<br/>2.4.4 Character Recognition . . . . . . . . . . . . . . . . . . . . 22<br/>2.4.4.1 Tesseract . . . . . . . . . . . . . . . . . . . . . . 22<br/>2.5 Alternative Methods . . . . . . . . . . . . . . . . . . . . . . . . . 23<br/>2.5.1 Character Detection . . . . . . . . . . . . . . . . . . . . . 24<br/>2.5.2 Non-Maximal Suppression . . . . . . . . . . . . . . . . . . 25<br/>2.5.3 Character Recognition . . . . . . . . . . . . . . . . . . . . 26<br/>2.6 Dictionary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26<br/>2.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27<br/>3 Algorithms Behind The Proposed Character and Background<br/>Recognition 29<br/>3.1 Perface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29<br/>3.2 Algorithms For Feature Extraction . . . . . . . . . . . . . . . . . 31<br/>3.2.1 Histogram of Oriented Gradients (HOG) . . . . . . . . . . 31<br/>3.2.2 Felzenszwalb's HOG[5] . . . . . . . . . . . . . . . . . . . . 34<br/>3.3 Used Classier: Support Vector Machines(SVMs) . . . . . . . . . 37<br/>3.3.1 Hard Margin . . . . . . . . . . . . . . . . . . . . . . . . . 38<br/>3.3.1.1 Primal Form . . . . . . . . . . . . . . . . . . . . 39<br/>3.3.1.2 Dual Form . . . . . . . . . . . . . . . . . . . . . 39<br/>3.3.2 Kernels . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41<br/>3.3.2.1 Linear Kernel . . . . . . . . . . . . . . . . . . . . 41<br/>3.3.2.2 Radial Basis Function . . . . . . . . . . . . . . . 42<br/>3.3.3 Soft Margin . . . . . . . . . . . . . . . . . . . . . . . . . . 43<br/>3.3.3.1 Updated Primal Form . . . . . . . . . . . . . . . 43<br/>3.3.3.2 Updated Dual Form . . . . . . . . . . . . . . . . 44<br/>3.4 Grid Search . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45<br/>vi<br/>CONTENTS<br/>3.5 Data Synthesization: Synthetic Minority<br/>Over-sampling TEchnique (SMOTE)[6] . . . . . . . . . . . . . . . 46<br/>3.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50<br/>4 Results and Discussion 51<br/>4.1 Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51<br/>4.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51<br/>4.2.1 Histogram of Oriented Gradients (HOG) and Felzenzwalb's<br/>HOG Conguration . . . . . . . . . . . . . . . . . . . . . . 51<br/>4.2.2 Grid Search Conguration . . . . . . . . . . . . . . . . . . 52<br/>4.2.3 Support Vector Machines (SVMs) Conguration . . . . . . 53<br/>4.2.4 Available Datasets . . . . . . . . . . . . . . . . . . . . . . 53<br/>4.2.5 Datasets' Requirements . . . . . . . . . . . . . . . . . . . 55<br/>4.2.6 Datasets Combinations . . . . . . . . . . . . . . . . . . . . 57<br/>4.2.6.1 First Combination . . . . . . . . . . . . . . . . . 57<br/>4.2.6.2 Second Combination . . . . . . . . . . . . . . . . 58<br/>4.3 Evaluation Metrics . . . . . . . . . . . . . . . . . . . . . . . . . . 59<br/>4.3.1 Confusion Matrix . . . . . . . . . . . . . . . . . . . . . . . 59<br/>4.3.2 F-score . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61<br/>4.4 Experiments & Results . . . . . . . . . . . . . . . . . . . . . . . . 63<br/>4.4.1 63 Classes with HOG . . . . . . . . . . . . . . . . . . . . . 63<br/>4.4.2 63 Classes with FHOG . . . . . . . . . . . . . . . . . . . . 65<br/>4.4.3 62 Classes . . . . . . . . . . . . . . . . . . . . . . . . . . . 67<br/>4.4.4 62 and 63 Classes vs [7] . . . . . . . . . . . . . . . . . . . 67<br/>4.4.5 56 Classes with HOG . . . . . . . . . . . . . . . . . . . . . 68<br/>5 Conclusion and Future work 71<br/>5.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71<br/>5.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73<br/>Bibliography |
| 520 3# - Abstract | |
| Abstract | Abstract:<br/>Text extraction from documents is an essential task which proved its<br/>contribution to many applications. The same benets can be gained<br/>when it comes for images and videos. However, text extraction from<br/>images and videos is not as advanced as in the case of documents. It<br/>is trailing behind because of the characteristics of natural scene text.<br/>One of the most important operations in any text extraction pipeline<br/>is the character recognition module. Character recognition is a substantial<br/>phase which needs to be performed in a better way so that it<br/>is improved and its dependence on other phases decrease. Its renement<br/>means better results for the whole text extraction process. With<br/>many advancements in object detection methods, it is an opportunity<br/>to introduce the same methodologies used for objects on text that<br/>appears in images and videos. The thesis aims to have a combined<br/>63-way character recognition to deal with characters and background<br/>at same time. Moreover, the thesis deals with preparing datasets and<br/>synthesizing samples to train character recognition. The nal results<br/>will be compared versus normal 62-way character recognition which<br/>only works for characters. |
| 546 ## - Language Note | |
| Language Note | Text in English, abstracts in English. |
| 650 #4 - Subject | |
| Subject | Informatics-IFM |
| 655 #7 - Index Term-Genre/Form | |
| Source of term | NULIB |
| focus term | Dissertation, Academic |
| 690 ## - Subject | |
| School | Informatics-IFM |
| 942 ## - ADDED ENTRY ELEMENTS (KOHA) | |
| Source of classification or shelving scheme | Dewey Decimal Classification |
| Koha item type | Thesis |
| 650 #4 - Subject | |
| -- | 266 |
| 655 #7 - Index Term-Genre/Form | |
| -- | 187 |
| 690 ## - Subject | |
| -- | 266 |
| Withdrawn status | Lost status | Source of classification or shelving scheme | Damaged status | Not for loan | Home library | Current library | Date acquired | Total Checkouts | Full call number | Date last seen | Price effective from | Koha item type |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Dewey Decimal Classification | Not For Loan | Main library | Main library | 01/12/2021 | 610/ IS.A 2014 | 01/12/2021 | 01/12/2021 | Thesis |