Robust speech recognition using fusion techniques and adaptive filtering

Syed Abdul Rahman Al-Haddad, Salina Abdul Samad, Aini Hussain, Khairul Anuar Ishak, A. O A Noor

Research output: Contribution to journalArticle

22 Citations (Scopus)

Abstract

The study proposes an algorithm for noise cancellation by using recursive least square (RLS) and pattern recognition by using fusion method of Dynamic Time Warping (DTW) and Hidden Markov Model (HMM). Speech signals are often corrupted with background noise and the changes in signal characteristics could be fast. These issues are especially important for robust speech recognition. Robustness is a key issue in speech recognition. The algorithm is tested on speech samples that are a part of a Malay corpus. It is shown that the fusion technique can be used to fuse the pattern recognition outputs of DTW and HMM. Furthermore refinement normalization was introduced by using weight mean vector to obtain better performance. Accuracy of 94% on pattern recognition was obtainable using fusion HMM and DTW compared to 80.5% using DTW and 90.7% using HMM separately. The accuracy of the proposed algorithm is increased further to 98% by utilization the RLS adaptive noise cancellation.

Original languageEnglish
Pages (from-to)290-295
Number of pages6
JournalAmerican Journal of Applied Sciences
Volume6
Issue number2
Publication statusPublished - 2009

Fingerprint

Adaptive filtering
Hidden Markov models
Speech recognition
Fusion reactions
Pattern recognition
Electric fuses

Keywords

  • DTW
  • HMM
  • RLS
  • Word bounder
  • Zero crossing technique

ASJC Scopus subject areas

  • General

Cite this

Robust speech recognition using fusion techniques and adaptive filtering. / Al-Haddad, Syed Abdul Rahman; Abdul Samad, Salina; Hussain, Aini; Ishak, Khairul Anuar; Noor, A. O A.

In: American Journal of Applied Sciences, Vol. 6, No. 2, 2009, p. 290-295.

Research output: Contribution to journalArticle

@article{2a7d06b4413842808f0f3fc75cebe9de,
title = "Robust speech recognition using fusion techniques and adaptive filtering",
abstract = "The study proposes an algorithm for noise cancellation by using recursive least square (RLS) and pattern recognition by using fusion method of Dynamic Time Warping (DTW) and Hidden Markov Model (HMM). Speech signals are often corrupted with background noise and the changes in signal characteristics could be fast. These issues are especially important for robust speech recognition. Robustness is a key issue in speech recognition. The algorithm is tested on speech samples that are a part of a Malay corpus. It is shown that the fusion technique can be used to fuse the pattern recognition outputs of DTW and HMM. Furthermore refinement normalization was introduced by using weight mean vector to obtain better performance. Accuracy of 94{\%} on pattern recognition was obtainable using fusion HMM and DTW compared to 80.5{\%} using DTW and 90.7{\%} using HMM separately. The accuracy of the proposed algorithm is increased further to 98{\%} by utilization the RLS adaptive noise cancellation.",
keywords = "DTW, HMM, RLS, Word bounder, Zero crossing technique",
author = "Al-Haddad, {Syed Abdul Rahman} and {Abdul Samad}, Salina and Aini Hussain and Ishak, {Khairul Anuar} and Noor, {A. O A}",
year = "2009",
language = "English",
volume = "6",
pages = "290--295",
journal = "American Journal of Applied Sciences",
issn = "1546-9239",
publisher = "Science Publications",
number = "2",

}

TY - JOUR

T1 - Robust speech recognition using fusion techniques and adaptive filtering

AU - Al-Haddad, Syed Abdul Rahman

AU - Abdul Samad, Salina

AU - Hussain, Aini

AU - Ishak, Khairul Anuar

AU - Noor, A. O A

PY - 2009

Y1 - 2009

N2 - The study proposes an algorithm for noise cancellation by using recursive least square (RLS) and pattern recognition by using fusion method of Dynamic Time Warping (DTW) and Hidden Markov Model (HMM). Speech signals are often corrupted with background noise and the changes in signal characteristics could be fast. These issues are especially important for robust speech recognition. Robustness is a key issue in speech recognition. The algorithm is tested on speech samples that are a part of a Malay corpus. It is shown that the fusion technique can be used to fuse the pattern recognition outputs of DTW and HMM. Furthermore refinement normalization was introduced by using weight mean vector to obtain better performance. Accuracy of 94% on pattern recognition was obtainable using fusion HMM and DTW compared to 80.5% using DTW and 90.7% using HMM separately. The accuracy of the proposed algorithm is increased further to 98% by utilization the RLS adaptive noise cancellation.

AB - The study proposes an algorithm for noise cancellation by using recursive least square (RLS) and pattern recognition by using fusion method of Dynamic Time Warping (DTW) and Hidden Markov Model (HMM). Speech signals are often corrupted with background noise and the changes in signal characteristics could be fast. These issues are especially important for robust speech recognition. Robustness is a key issue in speech recognition. The algorithm is tested on speech samples that are a part of a Malay corpus. It is shown that the fusion technique can be used to fuse the pattern recognition outputs of DTW and HMM. Furthermore refinement normalization was introduced by using weight mean vector to obtain better performance. Accuracy of 94% on pattern recognition was obtainable using fusion HMM and DTW compared to 80.5% using DTW and 90.7% using HMM separately. The accuracy of the proposed algorithm is increased further to 98% by utilization the RLS adaptive noise cancellation.

KW - DTW

KW - HMM

KW - RLS

KW - Word bounder

KW - Zero crossing technique

UR - http://www.scopus.com/inward/record.url?scp=52949122456&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=52949122456&partnerID=8YFLogxK

M3 - Article

VL - 6

SP - 290

EP - 295

JO - American Journal of Applied Sciences

JF - American Journal of Applied Sciences

SN - 1546-9239

IS - 2

ER -