An efficient method for text detection in video based on stroke width similarity

Viet Cuong Dinh, Soo Chun Seong, Cha Seungwook, Ryu Hanjin, Sanghoon Sull

Research output: Chapter in Book/Report/Conference proceedingConference contribution

18 Citations (Scopus)

Abstract

Text appearing in video provides semantic knowledge and significant information for video indexing and retrieval system. This paper proposes an effective method for text detection in video based on the similarity in stroke width of text (which is defined as the distance between two edges of a stroke). From the observation that text regions can be characterized by a dominant fixed stroke width, edge detection with local adaptive thresholds is first devised to keep text- while reducing background-regions. Second, morphological dilation operator with adaptive structuring element size determined by stroke width value is exploited to roughly localize text regions. Finally, to reduce false alarm and refine text location, a new multi-frame refinement method is applied. Experimental results show that the proposed method is not only robust to different levels of background complexity, but also effective to different fonts (size, color) and languages of text.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages200-209
Number of pages10
Volume4843 LNCS
EditionPART 1
Publication statusPublished - 2007 Dec 1
Event8th Asian Conference on Computer Vision, ACCV 2007 - Tokyo, Japan
Duration: 2007 Nov 182007 Nov 22

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 1
Volume4843 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other8th Asian Conference on Computer Vision, ACCV 2007
CountryJapan
CityTokyo
Period07/11/1807/11/22

Fingerprint

Edge detection
Stroke
Semantics
Color
Dilatation
Language
Video Indexing
Adaptive Threshold
Video Retrieval
Text
Similarity
False Alarm
Edge Detection
Dilation
Refinement
Experimental Results
Operator

ASJC Scopus subject areas

  • Biochemistry, Genetics and Molecular Biology(all)
  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Dinh, V. C., Seong, S. C., Seungwook, C., Hanjin, R., & Sull, S. (2007). An efficient method for text detection in video based on stroke width similarity. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (PART 1 ed., Vol. 4843 LNCS, pp. 200-209). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4843 LNCS, No. PART 1).

An efficient method for text detection in video based on stroke width similarity. / Dinh, Viet Cuong; Seong, Soo Chun; Seungwook, Cha; Hanjin, Ryu; Sull, Sanghoon.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 4843 LNCS PART 1. ed. 2007. p. 200-209 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4843 LNCS, No. PART 1).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Dinh, VC, Seong, SC, Seungwook, C, Hanjin, R & Sull, S 2007, An efficient method for text detection in video based on stroke width similarity. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 1 edn, vol. 4843 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 1, vol. 4843 LNCS, pp. 200-209, 8th Asian Conference on Computer Vision, ACCV 2007, Tokyo, Japan, 07/11/18.
Dinh VC, Seong SC, Seungwook C, Hanjin R, Sull S. An efficient method for text detection in video based on stroke width similarity. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 1 ed. Vol. 4843 LNCS. 2007. p. 200-209. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1).
Dinh, Viet Cuong ; Seong, Soo Chun ; Seungwook, Cha ; Hanjin, Ryu ; Sull, Sanghoon. / An efficient method for text detection in video based on stroke width similarity. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 4843 LNCS PART 1. ed. 2007. pp. 200-209 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1).
@inproceedings{fc45d2059f8f43a1ac277d46f0327447,
title = "An efficient method for text detection in video based on stroke width similarity",
abstract = "Text appearing in video provides semantic knowledge and significant information for video indexing and retrieval system. This paper proposes an effective method for text detection in video based on the similarity in stroke width of text (which is defined as the distance between two edges of a stroke). From the observation that text regions can be characterized by a dominant fixed stroke width, edge detection with local adaptive thresholds is first devised to keep text- while reducing background-regions. Second, morphological dilation operator with adaptive structuring element size determined by stroke width value is exploited to roughly localize text regions. Finally, to reduce false alarm and refine text location, a new multi-frame refinement method is applied. Experimental results show that the proposed method is not only robust to different levels of background complexity, but also effective to different fonts (size, color) and languages of text.",
author = "Dinh, {Viet Cuong} and Seong, {Soo Chun} and Cha Seungwook and Ryu Hanjin and Sanghoon Sull",
year = "2007",
month = "12",
day = "1",
language = "English",
isbn = "9783540763857",
volume = "4843 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
number = "PART 1",
pages = "200--209",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
edition = "PART 1",

}

TY - GEN

T1 - An efficient method for text detection in video based on stroke width similarity

AU - Dinh, Viet Cuong

AU - Seong, Soo Chun

AU - Seungwook, Cha

AU - Hanjin, Ryu

AU - Sull, Sanghoon

PY - 2007/12/1

Y1 - 2007/12/1

N2 - Text appearing in video provides semantic knowledge and significant information for video indexing and retrieval system. This paper proposes an effective method for text detection in video based on the similarity in stroke width of text (which is defined as the distance between two edges of a stroke). From the observation that text regions can be characterized by a dominant fixed stroke width, edge detection with local adaptive thresholds is first devised to keep text- while reducing background-regions. Second, morphological dilation operator with adaptive structuring element size determined by stroke width value is exploited to roughly localize text regions. Finally, to reduce false alarm and refine text location, a new multi-frame refinement method is applied. Experimental results show that the proposed method is not only robust to different levels of background complexity, but also effective to different fonts (size, color) and languages of text.

AB - Text appearing in video provides semantic knowledge and significant information for video indexing and retrieval system. This paper proposes an effective method for text detection in video based on the similarity in stroke width of text (which is defined as the distance between two edges of a stroke). From the observation that text regions can be characterized by a dominant fixed stroke width, edge detection with local adaptive thresholds is first devised to keep text- while reducing background-regions. Second, morphological dilation operator with adaptive structuring element size determined by stroke width value is exploited to roughly localize text regions. Finally, to reduce false alarm and refine text location, a new multi-frame refinement method is applied. Experimental results show that the proposed method is not only robust to different levels of background complexity, but also effective to different fonts (size, color) and languages of text.

UR - http://www.scopus.com/inward/record.url?scp=38349078899&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=38349078899&partnerID=8YFLogxK

M3 - Conference contribution

SN - 9783540763857

VL - 4843 LNCS

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

SP - 200

EP - 209

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

ER -