A Novel Bilingual OCR for Printed Malayalam-English Text Based on Gabor Features and Dominant Singular Values

Author

Philip, Bindu ; Samuel, R. D Sudhaker

Author_Institution

Dept. of Electron. & Commun., S.J. Coll. of Eng., Mysore, India

fYear

2009

fDate

7-9 March 2009

Firstpage

361

Lastpage

365

Abstract

In this paper a bilingual character recognition system is proposed for the characterization and classification of printed Malayalam-English characters. Indian scripts in general are rich in patterns and variations. Gabor features are extracted after the word level segmentation to identify the script and recognition is based on characterization using dominant singular values. A recognition rate of 96.5% was achieved for the two-stage classification approach.

Keywords

document image processing; feature extraction; image classification; image segmentation; natural language processing; optical character recognition; text analysis; Gabor feature extraction; Indian script; bilingual OCR; dominant singular value; image classification; optical character recognition system; printed Malayalam-English text; word level segmentation; Character recognition; Digital images; Educational institutions; Feature extraction; Gabor filters; Handicapped aids; Natural languages; Nearest neighbor searches; Optical character recognition software; Speech synthesis; Bilingual OCR; Dominant Singular Values; Gabor features; Segmentation; Two-stage classification;

fLanguage

English

Publisher

ieee

Conference_Titel

Digital Image Processing, 2009 International Conference on

Conference_Location

Bangkok

Print_ISBN

978-0-7695-3565-4

Type

conf

DOI

10.1109/ICDIP.2009.50

Filename

5190519