Text extraction in digital news video using morphology

Hyeran Byun; Inyoung Jang; Yeongwoo Choi

doi:10.1007/3-540-45869-7_39

Text extraction in digital news video using morphology

Hyeran Byun, Inyoung Jang, Yeongwoo Choi

Department of Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

2 Citations (Scopus)

Abstract

In this paper, a new method is presented to extract both superimposed and embedded scene texts in digital news videos. The algorithm is summarized in the following three steps: preprocessing, extracting candidate regions, and filtering candidate regions. For the first preprocessing step, a color image is converted into a gray-level image and a modified local adaptive thresholding is applied to the contrast-stretched image. In the second step, various morphological operations and Geo-correction method are applied to remove non-text components while retaining the text components. In the third filtering step, non-text components are removed based on the characteristics of each candidate component such as the number of pixels and the bounding box of each connected component Acceptable results have been obtained using the proposed method on 300 domestic news images with a recognition rate of 93.6%. Also, the proposed method gives good performance on the various kinds of images such as foreign news and film videos.

Original language	English
Title of host publication	Document Analysis Systems V - 5th International Workshop, DAS 2002, Proceedings
Editors	Daniel Lopresti, Jianying Hu, Ramanujan Kashi
Publisher	Springer Verlag
Pages	341-352
Number of pages	12
ISBN (Print)	3540440682, 9783540440680
DOIs	https://doi.org/10.1007/3-540-45869-7_39
Publication status	Published - 2002
Event	5th International Workshop on Document Analysis Systems, DAS 2002 - Princeton, United States Duration: 2002 Aug 19 → 2002 Aug 21

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	2423
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Other

Other	5th International Workshop on Document Analysis Systems, DAS 2002
Country	United States
City	Princeton
Period	02/8/19 → 02/8/21

All Science Journal Classification (ASJC) codes

Theoretical Computer Science
Computer Science(all)

Access to Document

10.1007/3-540-45869-7_39

Fingerprint Dive into the research topics of 'Text extraction in digital news video using morphology'. Together they form a unique fingerprint.

Cite this

Byun, H., Jang, I., & Choi, Y. (2002). Text extraction in digital news video using morphology. In D. Lopresti, J. Hu, & R. Kashi (Eds.), Document Analysis Systems V - 5th International Workshop, DAS 2002, Proceedings (pp. 341-352). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 2423). Springer Verlag. https://doi.org/10.1007/3-540-45869-7_39

Byun, Hyeran ; Jang, Inyoung ; Choi, Yeongwoo. / Text extraction in digital news video using morphology. Document Analysis Systems V - 5th International Workshop, DAS 2002, Proceedings. editor / Daniel Lopresti ; Jianying Hu ; Ramanujan Kashi. Springer Verlag, 2002. pp. 341-352 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{aa3945487b734d729d2d5374fa860a96,

title = "Text extraction in digital news video using morphology",

abstract = "In this paper, a new method is presented to extract both superimposed and embedded scene texts in digital news videos. The algorithm is summarized in the following three steps: preprocessing, extracting candidate regions, and filtering candidate regions. For the first preprocessing step, a color image is converted into a gray-level image and a modified local adaptive thresholding is applied to the contrast-stretched image. In the second step, various morphological operations and Geo-correction method are applied to remove non-text components while retaining the text components. In the third filtering step, non-text components are removed based on the characteristics of each candidate component such as the number of pixels and the bounding box of each connected component Acceptable results have been obtained using the proposed method on 300 domestic news images with a recognition rate of 93.6%. Also, the proposed method gives good performance on the various kinds of images such as foreign news and film videos.",

author = "Hyeran Byun and Inyoung Jang and Yeongwoo Choi",

note = "Publisher Copyright: {\textcopyright} Springer-Verlag Berlin Heidelberg 2002. Copyright: Copyright 2020 Elsevier B.V., All rights reserved.; 5th International Workshop on Document Analysis Systems, DAS 2002 ; Conference date: 19-08-2002 Through 21-08-2002",

year = "2002",

doi = "10.1007/3-540-45869-7_39",

language = "English",

isbn = "3540440682",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "341--352",

editor = "Daniel Lopresti and Jianying Hu and Ramanujan Kashi",

booktitle = "Document Analysis Systems V - 5th International Workshop, DAS 2002, Proceedings",

address = "Germany",

}

Byun, H, Jang, I & Choi, Y 2002, Text extraction in digital news video using morphology. in D Lopresti, J Hu & R Kashi (eds), Document Analysis Systems V - 5th International Workshop, DAS 2002, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 2423, Springer Verlag, pp. 341-352, 5th International Workshop on Document Analysis Systems, DAS 2002, Princeton, United States, 02/8/19. https://doi.org/10.1007/3-540-45869-7_39

Text extraction in digital news video using morphology. / Byun, Hyeran; Jang, Inyoung; Choi, Yeongwoo.

Document Analysis Systems V - 5th International Workshop, DAS 2002, Proceedings. ed. / Daniel Lopresti; Jianying Hu; Ramanujan Kashi. Springer Verlag, 2002. p. 341-352 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 2423).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Text extraction in digital news video using morphology

AU - Byun, Hyeran

AU - Jang, Inyoung

AU - Choi, Yeongwoo

PY - 2002

Y1 - 2002

N2 - In this paper, a new method is presented to extract both superimposed and embedded scene texts in digital news videos. The algorithm is summarized in the following three steps: preprocessing, extracting candidate regions, and filtering candidate regions. For the first preprocessing step, a color image is converted into a gray-level image and a modified local adaptive thresholding is applied to the contrast-stretched image. In the second step, various morphological operations and Geo-correction method are applied to remove non-text components while retaining the text components. In the third filtering step, non-text components are removed based on the characteristics of each candidate component such as the number of pixels and the bounding box of each connected component Acceptable results have been obtained using the proposed method on 300 domestic news images with a recognition rate of 93.6%. Also, the proposed method gives good performance on the various kinds of images such as foreign news and film videos.

AB - In this paper, a new method is presented to extract both superimposed and embedded scene texts in digital news videos. The algorithm is summarized in the following three steps: preprocessing, extracting candidate regions, and filtering candidate regions. For the first preprocessing step, a color image is converted into a gray-level image and a modified local adaptive thresholding is applied to the contrast-stretched image. In the second step, various morphological operations and Geo-correction method are applied to remove non-text components while retaining the text components. In the third filtering step, non-text components are removed based on the characteristics of each candidate component such as the number of pixels and the bounding box of each connected component Acceptable results have been obtained using the proposed method on 300 domestic news images with a recognition rate of 93.6%. Also, the proposed method gives good performance on the various kinds of images such as foreign news and film videos.

UR - http://www.scopus.com/inward/record.url?scp=33847234482&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33847234482&partnerID=8YFLogxK

U2 - 10.1007/3-540-45869-7_39

DO - 10.1007/3-540-45869-7_39

M3 - Conference contribution

AN - SCOPUS:33847234482

SN - 3540440682

SN - 9783540440680

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 341

EP - 352

BT - Document Analysis Systems V - 5th International Workshop, DAS 2002, Proceedings

A2 - Lopresti, Daniel

A2 - Hu, Jianying

A2 - Kashi, Ramanujan

PB - Springer Verlag

T2 - 5th International Workshop on Document Analysis Systems, DAS 2002

Y2 - 19 August 2002 through 21 August 2002

ER -

Byun H, Jang I, Choi Y. Text extraction in digital news video using morphology. In Lopresti D, Hu J, Kashi R, editors, Document Analysis Systems V - 5th International Workshop, DAS 2002, Proceedings. Springer Verlag. 2002. p. 341-352. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/3-540-45869-7_39