A Novel Multi-oriented Chinese Text Extraction Approach from Videos

Author

Yang Liu ; Yonghong Song ; Yuanlin Zhang ; Quan Meng

Author_Institution

Inst. of Artificial Intell. & Robot., Xi´an Jiaotong Univ., Xi´an, China

fYear

2013

fDate

25-28 Aug. 2013

Firstpage

1355

Lastpage

1359

Abstract

Video texts contain useful high level information which contributes to video indexing and retrieval. This paper proposed a novel video text extraction method for Chinese text of any orientation. Firstly, candidate text regions are detected by a wavelet based algorithm. Secondly, horizontal, vertical, slant, curve or arc text lines are merged by color and space relationship based on arrangement structures of multi-oriented text lines in these candidate regions. And then character segmentation will automatically choose best fit strategies based on structure analysis, due to the complex structures (single, up-down, left-right, encircling) of Chinese characters. Finally, a SVM classifier eliminates false positives. The experimental results show the proposed approach is robust for Chinese texts of any orientation in videos.

Keywords

feature extraction; image classification; image colour analysis; image segmentation; natural language processing; optical character recognition; support vector machines; text analysis; text detection; wavelet transforms; Chinese characters; SVM classifier; arc text line; character segmentation; color relationship; curve text line; encircling structure analysis; false positive elimination; high-level information; horizontal text line; left-right structure analysis; multioriented Chinese text extraction approach; multioriented text line arrangement structures; single-structure analysis; slant text line; space relationship; text line merging; text region detection; up-down structure analysis; vertical text line; video indexing; video retrieval; video text extraction method; wavelet-based algorithm; Feature extraction; Gabor filters; Image color analysis; Image edge detection; Image segmentation; Merging; Videos; Chinese character; character segmentation; multi-oriented; text extraction; text line merging;

fLanguage

English

Publisher

ieee

Conference_Titel

Document Analysis and Recognition (ICDAR), 2013 12th International Conference on

Conference_Location

Washington, DC

ISSN

1520-5363

Type

conf

DOI

10.1109/ICDAR.2013.274

Filename

6628835