Multi-font recognition of printed Arabic using the BBN BYBLOS speech recognition system

Author

LaPre, Christopher ; Zhao, Ying ; Raphael, Christopher ; Schwartz, Richard ; Makhoul, John

Author_Institution

BBN Syst. & Technol. Corp., Cambridge, MA, USA

Volume

4

fYear

1996

fDate

7-10 May 1996

Firstpage

2136

Abstract

We use a hidden Markov model (HMM) based continuous speech recognition system to perform off-line character recognition (OCR) of Arabic printed text. The HMM trainer and recognizer are used without change, however we modify the feature extraction stage to compute features relevant to OCR. Although we begin by segmenting the page into a collection of lines, no further segmentation is necessary for either recognition or training. Experiments on the ARPA Arabic data corpus yield a range of character error rates from under one percent for a single computer font to 2.8% for multiple-font recognition of a wide range of material from books, magazines and newspapers

Keywords

feature extraction; hidden Markov models; image segmentation; optical character recognition; speech recognition; ARPA Arabic data corpus; BBN BYBLOS speech recognition system; HMM; HMM recognizer; HMM trainer; books; character error rates; continuous speech recognition system; experiments; feature extraction; hidden Markov model; magazines; multifont recognition; newspapers; off-line character recognition; page segmentation; printed Arabic; single computer font; training; Character recognition; Error analysis; Feature extraction; Handwriting recognition; Hidden Markov models; Histograms; Optical character recognition software; Optical materials; Speech recognition; Text recognition;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1996. ICASSP-96. Conference Proceedings., 1996 IEEE International Conference on

Conference_Location

Atlanta, GA

ISSN

1520-6149

Print_ISBN

0-7803-3192-3

Type

conf

DOI

10.1109/ICASSP.1996.545738

Filename

545738