Publications
Detailed Information
Text Localization in Natural Images Using Multiple Feature Fusion : 다양한 Feature의 조합을 통하여 자연 영상에서 글자를 찾는 방법
DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | 유석인 | - |
dc.contributor.author | 김재석 | - |
dc.date.accessioned | 2017-07-14T02:58:17Z | - |
dc.date.available | 2017-07-14T02:58:17Z | - |
dc.date.issued | 2015-02 | - |
dc.identifier.other | 000000024743 | - |
dc.identifier.uri | https://hdl.handle.net/10371/123115 | - |
dc.description | 학위논문 (석사)-- 서울대학교 대학원 : 전기·컴퓨터공학부, 2015. 2. 유석인. | - |
dc.description.abstract | Text localization in natural scene images is an important first step to analyze content-based images. In this thesis, we propose an accurate method for detecting texts in natural scene images by adapting multiple feature combinations. Firstly, various color spaces are used to extract Maximally Stable Extremal Regions (MSERs) as character candidates. K-means clustering, image gradients, and Lucy and Richardson de-convolution (LRD) method are used to emphasize significant feature points and to compensate noisy images for character candidates. Secondly, important character candidates will be accentuated by applying enhanced canny edge detector to grayscale image gradients and our cumulative MSERs image. Thirdly, character candidates will be merged into text candidates by using clustering algorithm and geometric information such that texts usually appear in a linear form. Finally, inconsequential text candidates will be eliminated by using stroke width, text region height, width, and parallel edge features.
The method was evaluated on two benchmark datasets: International Conference on Document Analysis and Recognition (ICDAR) 2013 and Street View Text (SVT) from Google Maps. Experimental results on respective datasets show that the proposed algorithm works successfully with not only the normal text images, but also highlighted, transparent, small, and blurred texts. | - |
dc.description.tableofcontents | Abstract i
Contents ii List of Figures iii List of Tables iv Chapter 1 Introduction 1 Chapter 2 Previous Work 4 2.1 SWT . . . . . . . . . . . . . . . . 5 2.2 MSERs . . . . . . . . . . . . . . . 7 2.3 Mean-Shift Clustering . . . . . . . . . . . 9 2.4 Object Approach. . . . . . . . . . . . . 9 Chapter 3 Proposed Approach 10 3.1 Character Candidate Extraction . . . . . . . . . 11 3.1.1 MSERs . . . . . . . . . . . . . 11 3.1.2 Edge Enhanced Map . . . . . . . . . 12 3.1.3 K-Means Clustering . . . . . . . . . . 14 3.2 Non-text Filtering . . . . . . . . . . . 17 3.2.1 Geometric Filtering . . . . . . . . . . 17 3.2.2 Stroke Width Filtering . . . . . . . . . 17 Chapter 4 Experiments 20 4.1 Environment. . . . . . . . . . . . . . 20 4.2 Evaluation Metrics . . . . . . . . . . . . 21 4.3 Experimental Results . . . . . . . . . . . 23 4.4 Performance Evaluation . . . . . . . . . 24 4.5 Runtime Evaluation . . . . . . . . . . . 24 Chapter 5 Conclusion 29 5.1 Summary of the Work. . . . . . . . . . . 29 5.2 Future Work . . . . . . . . . . . . . . 30 Bibliography 31 Abstract in Korean 35 | - |
dc.format | application/pdf | - |
dc.format.extent | 1765342 bytes | - |
dc.format.medium | application/pdf | - |
dc.language.iso | en | - |
dc.publisher | 서울대학교 대학원 | - |
dc.subject | Text localization | - |
dc.subject | feature fusion | - |
dc.subject | maximally stable extremal regions | - |
dc.subject | k-means clustering | - |
dc.subject | Lucy and Richardson deconvolution | - |
dc.subject.ddc | 621 | - |
dc.title | Text Localization in Natural Images Using Multiple Feature Fusion | - |
dc.title.alternative | 다양한 Feature의 조합을 통하여 자연 영상에서 글자를 찾는 방법 | - |
dc.type | Thesis | - |
dc.description.degree | Master | - |
dc.citation.pages | vii, 36 | - |
dc.contributor.affiliation | 공과대학 전기·컴퓨터공학부 | - |
dc.date.awarded | 2015-02 | - |
- Appears in Collections:
- Files in This Item:
Item View & Download Count
Items in S-Space are protected by copyright, with all rights reserved, unless otherwise indicated.