Data-driven Arabic phoneme recognition using varying number of HMM states

Khalid M.O. Nahar; Wasfi G. Al-Khatib; Moustafa Elshafei; Husni Al-Muhtaseb; Mansour M. Alghamdi

doi:10.1109/ICCSPA.2013.6487258

Data-driven Arabic phoneme recognition using varying number of HMM states

Khalid M.O. Nahar^*, Wasfi G. Al-Khatib, Moustafa Elshafei, Husni Al-Muhtaseb, Mansour M. Alghamdi

^*Corresponding author for this work

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

7 Scopus citations

Abstract

Continuous Arabic Speech Recognition, appears in many real life applications. Its speed, accuracy and improvement are highly dependent on the accuracy of the language phonemes set. The main goal of this research is to recognize and transcribe the Arabic phonemes based on a data-driven approach. We built a phoneme recognizer based on a data driven approach using HTK tool. Different numbers of Gaussian mixtures with different numbers of HMM states were used in modeling the Arabic phonemes in order to reach the best configuration. The corpus used consists of about 4000 files, representing 5 recorded hours of modern standard Arabic of TV-News. The maximum phoneme recognition accuracy reached was 56.79%. This result is very encouraging and shows the viability of our approach as compared to using a fixed number of HMM states.

Original language	English
Title of host publication	2013 1st International Conference on Communications, Signal Processing and Their Applications, ICCSPA 2013
DOIs	https://doi.org/10.1109/ICCSPA.2013.6487258
State	Published - 2013

Publication series

Name	2013 1st International Conference on Communications, Signal Processing and Their Applications, ICCSPA 2013

Keywords

Arabic Speech Recognition
KFUPM Arabic speech
Phoneme recognition
corpus HMM

ASJC Scopus subject areas

Computer Networks and Communications
Signal Processing

Access to Document

10.1109/ICCSPA.2013.6487258

Cite this

Nahar, K. M. O., Al-Khatib, W. G., Elshafei, M., Al-Muhtaseb, H., & Alghamdi, M. M. (2013). Data-driven Arabic phoneme recognition using varying number of HMM states. In 2013 1st International Conference on Communications, Signal Processing and Their Applications, ICCSPA 2013 Article 6487258 (2013 1st International Conference on Communications, Signal Processing and Their Applications, ICCSPA 2013). https://doi.org/10.1109/ICCSPA.2013.6487258

Nahar, Khalid M.O. ; Al-Khatib, Wasfi G. ; Elshafei, Moustafa et al. / Data-driven Arabic phoneme recognition using varying number of HMM states. 2013 1st International Conference on Communications, Signal Processing and Their Applications, ICCSPA 2013. 2013. (2013 1st International Conference on Communications, Signal Processing and Their Applications, ICCSPA 2013).

@inproceedings{eeb234234a1e430fb5c5a06976249446,

title = "Data-driven Arabic phoneme recognition using varying number of HMM states",

abstract = "Continuous Arabic Speech Recognition, appears in many real life applications. Its speed, accuracy and improvement are highly dependent on the accuracy of the language phonemes set. The main goal of this research is to recognize and transcribe the Arabic phonemes based on a data-driven approach. We built a phoneme recognizer based on a data driven approach using HTK tool. Different numbers of Gaussian mixtures with different numbers of HMM states were used in modeling the Arabic phonemes in order to reach the best configuration. The corpus used consists of about 4000 files, representing 5 recorded hours of modern standard Arabic of TV-News. The maximum phoneme recognition accuracy reached was 56.79%. This result is very encouraging and shows the viability of our approach as compared to using a fixed number of HMM states.",

keywords = "Arabic Speech Recognition, KFUPM Arabic speech, Phoneme recognition, corpus HMM",

author = "Nahar, {Khalid M.O.} and Al-Khatib, {Wasfi G.} and Moustafa Elshafei and Husni Al-Muhtaseb and Alghamdi, {Mansour M.}",

year = "2013",

doi = "10.1109/ICCSPA.2013.6487258",

language = "English",

isbn = "9781467328210",

series = "2013 1st International Conference on Communications, Signal Processing and Their Applications, ICCSPA 2013",

booktitle = "2013 1st International Conference on Communications, Signal Processing and Their Applications, ICCSPA 2013",

}

Nahar, KMO, Al-Khatib, WG, Elshafei, M, Al-Muhtaseb, H & Alghamdi, MM 2013, Data-driven Arabic phoneme recognition using varying number of HMM states. in 2013 1st International Conference on Communications, Signal Processing and Their Applications, ICCSPA 2013., 6487258, 2013 1st International Conference on Communications, Signal Processing and Their Applications, ICCSPA 2013. https://doi.org/10.1109/ICCSPA.2013.6487258

Data-driven Arabic phoneme recognition using varying number of HMM states. / Nahar, Khalid M.O.; Al-Khatib, Wasfi G.; Elshafei, Moustafa et al.
2013 1st International Conference on Communications, Signal Processing and Their Applications, ICCSPA 2013. 2013. 6487258 (2013 1st International Conference on Communications, Signal Processing and Their Applications, ICCSPA 2013).

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

TY - GEN

T1 - Data-driven Arabic phoneme recognition using varying number of HMM states

AU - Nahar, Khalid M.O.

AU - Al-Khatib, Wasfi G.

AU - Elshafei, Moustafa

AU - Al-Muhtaseb, Husni

AU - Alghamdi, Mansour M.

PY - 2013

Y1 - 2013

N2 - Continuous Arabic Speech Recognition, appears in many real life applications. Its speed, accuracy and improvement are highly dependent on the accuracy of the language phonemes set. The main goal of this research is to recognize and transcribe the Arabic phonemes based on a data-driven approach. We built a phoneme recognizer based on a data driven approach using HTK tool. Different numbers of Gaussian mixtures with different numbers of HMM states were used in modeling the Arabic phonemes in order to reach the best configuration. The corpus used consists of about 4000 files, representing 5 recorded hours of modern standard Arabic of TV-News. The maximum phoneme recognition accuracy reached was 56.79%. This result is very encouraging and shows the viability of our approach as compared to using a fixed number of HMM states.

AB - Continuous Arabic Speech Recognition, appears in many real life applications. Its speed, accuracy and improvement are highly dependent on the accuracy of the language phonemes set. The main goal of this research is to recognize and transcribe the Arabic phonemes based on a data-driven approach. We built a phoneme recognizer based on a data driven approach using HTK tool. Different numbers of Gaussian mixtures with different numbers of HMM states were used in modeling the Arabic phonemes in order to reach the best configuration. The corpus used consists of about 4000 files, representing 5 recorded hours of modern standard Arabic of TV-News. The maximum phoneme recognition accuracy reached was 56.79%. This result is very encouraging and shows the viability of our approach as compared to using a fixed number of HMM states.

KW - Arabic Speech Recognition

KW - KFUPM Arabic speech

KW - Phoneme recognition

KW - corpus HMM

UR - http://www.scopus.com/inward/record.url?scp=84876036383&partnerID=8YFLogxK

U2 - 10.1109/ICCSPA.2013.6487258

DO - 10.1109/ICCSPA.2013.6487258

M3 - Conference contribution

AN - SCOPUS:84876036383

SN - 9781467328210

T3 - 2013 1st International Conference on Communications, Signal Processing and Their Applications, ICCSPA 2013

BT - 2013 1st International Conference on Communications, Signal Processing and Their Applications, ICCSPA 2013

ER -

Nahar KMO, Al-Khatib WG, Elshafei M, Al-Muhtaseb H, Alghamdi MM. Data-driven Arabic phoneme recognition using varying number of HMM states. In 2013 1st International Conference on Communications, Signal Processing and Their Applications, ICCSPA 2013. 2013. 6487258. (2013 1st International Conference on Communications, Signal Processing and Their Applications, ICCSPA 2013). doi: 10.1109/ICCSPA.2013.6487258

Data-driven Arabic phoneme recognition using varying number of HMM states

Abstract

Publication series

Keywords

ASJC Scopus subject areas

Access to Document

Fingerprint

Cite this