The QCRI recognition system for handwritten Arabic

Felix Stahlberg, Stephan Vogel

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

10 Citations (Scopus)

Abstract

This paper describes our recognition system for handwritten Arabic. We propose novel text line image normalization procedures and a new feature extraction method. Our recognition system is based on the Kaldi recognition toolkit which is widely used in automatic speech recognition (ASR) research. We show that the combination of sophisticated text image normalization and state-of-the art techniques originating from ASR results in a very robust and accurate recognizer. Our system outperforms the best systems in the literature by over 20% relative on the abcde-s configuration of the IFN/ENIT database and achieves comparable performance on other configurations. On the KHATT corpus, we report 11% relative improvement compared to the best system in the literature.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer Verlag
Pages276-286
Number of pages11
Volume9280
ISBN (Print)9783319232331
DOIs
Publication statusPublished - 2015
Event18th International Conference on Image Analysis and Processing, ICIAP 2015 - Genoa, Italy
Duration: 7 Sep 201511 Sep 2015

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9280
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other18th International Conference on Image Analysis and Processing, ICIAP 2015
CountryItaly
CityGenoa
Period7/9/1511/9/15

Fingerprint

Speech recognition
Automatic Speech Recognition
Feature extraction
Normalization
Configuration
Feature Extraction
Line
Text

Keywords

  • Arabic
  • Handwriting recognition
  • Text image normalization

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Stahlberg, F., & Vogel, S. (2015). The QCRI recognition system for handwritten Arabic. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9280, pp. 276-286). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9280). Springer Verlag. https://doi.org/10.1007/978-3-319-23234-8_26

The QCRI recognition system for handwritten Arabic. / Stahlberg, Felix; Vogel, Stephan.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9280 Springer Verlag, 2015. p. 276-286 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9280).

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

Stahlberg, F & Vogel, S 2015, The QCRI recognition system for handwritten Arabic. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 9280, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9280, Springer Verlag, pp. 276-286, 18th International Conference on Image Analysis and Processing, ICIAP 2015, Genoa, Italy, 7/9/15. https://doi.org/10.1007/978-3-319-23234-8_26
Stahlberg F, Vogel S. The QCRI recognition system for handwritten Arabic. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9280. Springer Verlag. 2015. p. 276-286. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-23234-8_26
Stahlberg, Felix ; Vogel, Stephan. / The QCRI recognition system for handwritten Arabic. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9280 Springer Verlag, 2015. pp. 276-286 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{560889a748f843f59f6474d660ba047e,
title = "The QCRI recognition system for handwritten Arabic",
abstract = "This paper describes our recognition system for handwritten Arabic. We propose novel text line image normalization procedures and a new feature extraction method. Our recognition system is based on the Kaldi recognition toolkit which is widely used in automatic speech recognition (ASR) research. We show that the combination of sophisticated text image normalization and state-of-the art techniques originating from ASR results in a very robust and accurate recognizer. Our system outperforms the best systems in the literature by over 20{\%} relative on the abcde-s configuration of the IFN/ENIT database and achieves comparable performance on other configurations. On the KHATT corpus, we report 11{\%} relative improvement compared to the best system in the literature.",
keywords = "Arabic, Handwriting recognition, Text image normalization",
author = "Felix Stahlberg and Stephan Vogel",
year = "2015",
doi = "10.1007/978-3-319-23234-8_26",
language = "English",
isbn = "9783319232331",
volume = "9280",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "276--286",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - The QCRI recognition system for handwritten Arabic

AU - Stahlberg, Felix

AU - Vogel, Stephan

PY - 2015

Y1 - 2015

N2 - This paper describes our recognition system for handwritten Arabic. We propose novel text line image normalization procedures and a new feature extraction method. Our recognition system is based on the Kaldi recognition toolkit which is widely used in automatic speech recognition (ASR) research. We show that the combination of sophisticated text image normalization and state-of-the art techniques originating from ASR results in a very robust and accurate recognizer. Our system outperforms the best systems in the literature by over 20% relative on the abcde-s configuration of the IFN/ENIT database and achieves comparable performance on other configurations. On the KHATT corpus, we report 11% relative improvement compared to the best system in the literature.

AB - This paper describes our recognition system for handwritten Arabic. We propose novel text line image normalization procedures and a new feature extraction method. Our recognition system is based on the Kaldi recognition toolkit which is widely used in automatic speech recognition (ASR) research. We show that the combination of sophisticated text image normalization and state-of-the art techniques originating from ASR results in a very robust and accurate recognizer. Our system outperforms the best systems in the literature by over 20% relative on the abcde-s configuration of the IFN/ENIT database and achieves comparable performance on other configurations. On the KHATT corpus, we report 11% relative improvement compared to the best system in the literature.

KW - Arabic

KW - Handwriting recognition

KW - Text image normalization

UR - http://www.scopus.com/inward/record.url?scp=84944721756&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84944721756&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-23234-8_26

DO - 10.1007/978-3-319-23234-8_26

M3 - Conference contribution

SN - 9783319232331

VL - 9280

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

SP - 276

EP - 286

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

PB - Springer Verlag

ER -