Text location in camera-captured guidepost images

Qiaoyu Sun; Yue Lu

doi:10.1109/CCPR.2010.5659136

Text location in camera-captured guidepost images

Qiaoyu Sun, Yue Lu

School of Communication and Electronic Engineering

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

3 Scopus citations

Abstract

A method is proposed to locate text in camera- captured guidepost images. Firstly, due to its advantage of smoothing the low contrast information, mean shift method is applied to remove some complex background. In order to improve the time efficiency, we modify the traditional mean shift method. Secondly, two stage features are used for the edge map of image to classify it into candidate text blocks and background by K-means algorithm. Finally, the text blocks are identified according to some empirical rules and project profile analysis. Experimental results demonstrate that the proposed method can efficiently locate the text region contained in camera-captured guidepost images and be robust for font, size, color, language, alignment and complexity of background.

Original language	English
Title of host publication	2010 Chinese Conference on Pattern Recognition, CCPR 2010 - Proceedings
Pages	98-101
Number of pages	4
DOIs	https://doi.org/10.1109/CCPR.2010.5659136
State	Published - 2010
Event	2010 Chinese Conference on Pattern Recognition, CCPR 2010 - Chongqing, China Duration: 21 Oct 2010 → 23 Oct 2010

Publication series

Name	2010 Chinese Conference on Pattern Recognition, CCPR 2010 - Proceedings

Conference

Conference	2010 Chinese Conference on Pattern Recognition, CCPR 2010
Country/Territory	China
City	Chongqing
Period	21/10/10 → 23/10/10

Keywords

Connected component analysis
K-means
Mean shift
Text detection
Text location

Access to Document

10.1109/CCPR.2010.5659136

Cite this

@inproceedings{0853d51a7f1c4cd7ba0d1f7aa32f0f1b,

title = "Text location in camera-captured guidepost images",

abstract = "A method is proposed to locate text in camera- captured guidepost images. Firstly, due to its advantage of smoothing the low contrast information, mean shift method is applied to remove some complex background. In order to improve the time efficiency, we modify the traditional mean shift method. Secondly, two stage features are used for the edge map of image to classify it into candidate text blocks and background by K-means algorithm. Finally, the text blocks are identified according to some empirical rules and project profile analysis. Experimental results demonstrate that the proposed method can efficiently locate the text region contained in camera-captured guidepost images and be robust for font, size, color, language, alignment and complexity of background.",

keywords = "Connected component analysis, K-means, Mean shift, Text detection, Text location",

author = "Qiaoyu Sun and Yue Lu",

year = "2010",

doi = "10.1109/CCPR.2010.5659136",

language = "英语",

isbn = "9781424472109",

series = "2010 Chinese Conference on Pattern Recognition, CCPR 2010 - Proceedings",

pages = "98--101",

booktitle = "2010 Chinese Conference on Pattern Recognition, CCPR 2010 - Proceedings",

note = "2010 Chinese Conference on Pattern Recognition, CCPR 2010 ; Conference date: 21-10-2010 Through 23-10-2010",

}

Sun, Q & Lu, Y 2010, Text location in camera-captured guidepost images. in 2010 Chinese Conference on Pattern Recognition, CCPR 2010 - Proceedings., 5659136, 2010 Chinese Conference on Pattern Recognition, CCPR 2010 - Proceedings, pp. 98-101, 2010 Chinese Conference on Pattern Recognition, CCPR 2010, Chongqing, China, 21/10/10. https://doi.org/10.1109/CCPR.2010.5659136

TY - GEN

T1 - Text location in camera-captured guidepost images

AU - Sun, Qiaoyu

AU - Lu, Yue

PY - 2010

Y1 - 2010

N2 - A method is proposed to locate text in camera- captured guidepost images. Firstly, due to its advantage of smoothing the low contrast information, mean shift method is applied to remove some complex background. In order to improve the time efficiency, we modify the traditional mean shift method. Secondly, two stage features are used for the edge map of image to classify it into candidate text blocks and background by K-means algorithm. Finally, the text blocks are identified according to some empirical rules and project profile analysis. Experimental results demonstrate that the proposed method can efficiently locate the text region contained in camera-captured guidepost images and be robust for font, size, color, language, alignment and complexity of background.

AB - A method is proposed to locate text in camera- captured guidepost images. Firstly, due to its advantage of smoothing the low contrast information, mean shift method is applied to remove some complex background. In order to improve the time efficiency, we modify the traditional mean shift method. Secondly, two stage features are used for the edge map of image to classify it into candidate text blocks and background by K-means algorithm. Finally, the text blocks are identified according to some empirical rules and project profile analysis. Experimental results demonstrate that the proposed method can efficiently locate the text region contained in camera-captured guidepost images and be robust for font, size, color, language, alignment and complexity of background.

KW - Connected component analysis

KW - K-means

KW - Mean shift

KW - Text detection

KW - Text location

UR - https://www.scopus.com/pages/publications/78651423854

U2 - 10.1109/CCPR.2010.5659136

DO - 10.1109/CCPR.2010.5659136

M3 - 会议稿件

AN - SCOPUS:78651423854

SN - 9781424472109

T3 - 2010 Chinese Conference on Pattern Recognition, CCPR 2010 - Proceedings

SP - 98

EP - 101

BT - 2010 Chinese Conference on Pattern Recognition, CCPR 2010 - Proceedings

T2 - 2010 Chinese Conference on Pattern Recognition, CCPR 2010

Y2 - 21 October 2010 through 23 October 2010

ER -

Text location in camera-captured guidepost images

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this