Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Bidirectional deep architecture for Arabic speech recognition

Bidirectional deep architecture for Arabic speech recognition AbstractNowadays, the real life constraints necessitates controlling modern machines using human intervention by means of sensorial organs. The voice is one of the human senses that can control/monitor modern interfaces. In this context, Automatic Speech Recognition is principally used to convert natural voice into computer text as well as to perform an action based on the instructions given by the human. In this paper, we propose a general framework for Arabic speech recognition that uses Long Short-Term Memory (LSTM) and Neural Network (Multi-Layer Perceptron: MLP) classifier to cope with the nonuniform sequence length of the speech utterances issued fromboth feature extraction techniques, (1)Mel Frequency Cepstral Coefficients MFCC (static and dynamic features), (2) the Filter Banks (FB) coefficients. The neural architecture can recognize the isolated Arabic speech via classification technique. The proposed system involves, first, extracting pertinent features from the natural speech signal using MFCC (static and dynamic features) and FB. Next, the extracted features are padded in order to deal with the non-uniformity of the sequences length. Then, a deep architecture represented by a recurrent LSTM or GRU (Gated Recurrent Unit) architectures are used to encode the sequences of MFCC/FB features as a fixed size vector that will be introduced to a Multi-Layer Perceptron network (MLP) to perform the classification (recognition). The proposed system is assessed using two different databases, the first one concerns the spoken digit recognition where a comparison with other related works in the literature is performed, whereas the second one contains the spoken TV commands. The obtained results show the superiority of the proposed approach. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Open Computer Science de Gruyter

Bidirectional deep architecture for Arabic speech recognition

Download PDF
11 pages

Loading next page...
 
/lp/de-gruyter/bidirectional-deep-architecture-for-arabic-speech-recognition-NCALIqKv0A
Publisher
de Gruyter
Copyright
© 2019 Naima Zerari et al., published by De Gruyter Open
eISSN
2299-1093
DOI
10.1515/comp-2019-0004
Publisher site
See Article on Publisher Site

Abstract

AbstractNowadays, the real life constraints necessitates controlling modern machines using human intervention by means of sensorial organs. The voice is one of the human senses that can control/monitor modern interfaces. In this context, Automatic Speech Recognition is principally used to convert natural voice into computer text as well as to perform an action based on the instructions given by the human. In this paper, we propose a general framework for Arabic speech recognition that uses Long Short-Term Memory (LSTM) and Neural Network (Multi-Layer Perceptron: MLP) classifier to cope with the nonuniform sequence length of the speech utterances issued fromboth feature extraction techniques, (1)Mel Frequency Cepstral Coefficients MFCC (static and dynamic features), (2) the Filter Banks (FB) coefficients. The neural architecture can recognize the isolated Arabic speech via classification technique. The proposed system involves, first, extracting pertinent features from the natural speech signal using MFCC (static and dynamic features) and FB. Next, the extracted features are padded in order to deal with the non-uniformity of the sequences length. Then, a deep architecture represented by a recurrent LSTM or GRU (Gated Recurrent Unit) architectures are used to encode the sequences of MFCC/FB features as a fixed size vector that will be introduced to a Multi-Layer Perceptron network (MLP) to perform the classification (recognition). The proposed system is assessed using two different databases, the first one concerns the spoken digit recognition where a comparison with other related works in the literature is performed, whereas the second one contains the spoken TV commands. The obtained results show the superiority of the proposed approach.

Journal

Open Computer Science – de Gruyter

Published: Jan 1, 2019

References

  • UCIMachine Learning Repository University of California http archive ics uci edu ml
    Lichman
  • Bidirectional recurrent end - to - end neural network classifier for spoken Arab digit recognition International Conference on Natural Language and Speech Processing
    Zerari
  • - spectral cepstral coefficients for robust speech recognition international conference on acoustics speech and signal processing
    Kumar
  • Deep neural networks for acoustic modeling in speech recognition processing magazine
    Hinton
  • speech and - speaker identifcation system : feature extraction description and classification of speech signal image transactions on industrial
    Saeed
  • recognition using mfcc Conference on and Modeling
    Ittichaichareon
  • Young The HTK hiddenMarkov model toolkit University of Department of
  • Dropout a simpleway to prevent neural networks from over - fitting of
    Srivastava
  • Pattern classification
    Duda
  • Fundamentals of speech recognition Cliffs
    Rabiner
  • Automatic Speech Recognition A Brief History of the Technology Development Georgia Institute of Technology Atlanta Rutgers University and the University of California Santa
    Juang
  • Speaker - independent isolated word recognition based on emphasized spectral dynamics International Conference on Acoustics speech Processing
    Furui
  • The python deep learning Source Code
    Chollet
  • networks for pattern recognition University
    Bishop
  • Comparison of parametric representations for monosyllabic word recognition in continuously spoken sentences Transactions on acoustics speech and signal processing
    Davis
  • step beyond local observations with a dialog aware bidirectional GRU network for Spoken Language Understanding th Annual Conference of the International Speech Association
    Vukotic
  • Hybrid speech recognition with deep bidirectional LSTM workshop on automatic speech recognition and understanding
    Graves
  • networks learning machines Upper Saddle River
    Haykin
  • Theory and applications of digital speech processing Upper Saddle River
    Rabiner
  • Deep learning
    LeCun
  • Introduction to arabic speech recognition using cmu sphinx system arXiv preprint arXiv
    Satori
  • detection using cepstrum features
    Lalitha
  • Chip design of MFCC extraction for speech recognition the
    Wang
  • Classification of speech dysfluencies with features with
    Hariharan
  • Improved tree model for arabic speech recognition International Conference on Technology
    Hammami
  • complete kaldi recipe for building arabic speech recognition systems spoken language technology workshop
    Ali
  • models for sequence tagging arXiv preprint arXiv
    Huang
  • Automatic Analysis of High dimensional Signals Advanced Wind Speed Forecasting Techniques Academic Publishing
    Bouzgou
  • Speech recognition with deep recurrent neural networks IEEE International conference on acoustics speech and signal processing
    Graves
  • Wavelet lpc with neural network for spoken arabic digits recognition system of Technology
    Daqrouq
  • The challenge multi - dialect broadcast media recognition Spoken Language Technology Workshop
    Ali
  • Speech recognition challenge in the wild Automatic Speech Recognition and Understanding Workshop
    Ali
  • Context - dependent pretrained deep neural networks for large - vocabulary speech recognition on audio and language processing
    Dahl
  • Connectionist temporal classification : labelling unsegmented sequence data with recurrent neural networks The rd International Conference on Machine Learning
    Graves
  • Statistical methods for speech recognition MIT press
    Jelinek
  • Recent progress in Arabic broadcast news transcription at Ninth European Conference on Technology
    Afify
  • Gated feedback recurrent neural networks International Conference on Machine
    Chung
  • Confidence measures for speech recognition survey communication
    Jiang
  • Kaldi system for challenge using diarization audio - transcript alignment and transfer learning Automatic Speech Recognition and Understanding Workshop
    Manohar
  • Deep gate recurrent neural network Asian Conference on Machine
    Yuan Gao
  • San Feature extraction fromsmartphone inertial signals for human activity segmentation Processing
    Segundo
  • Feature extraction and classifcation techniques for speech recognition of Emerging Technology and Advanced
    Desai
  • of neural networks for speech recognition computation
    Lippmann
  • analysis and experimental evaluation of block based transformation in MFCC computation for speaker recognition
    Sahidullah