Advanced search
1 file | 376.87 KB

Noise robust continuous digit recognition with reservoir-based acoustic models

Azarakhsh Jalalvand (UGent) , Kris Demuynck (UGent) and Jean-Pierre Martens (UGent)
Author
Organization
Abstract
Notwithstanding the many years of research, more work is needed to create automatic speech recognition (ASR) systems with a close-to-human robustness against confounding factors such as ambient noise, channel distortion, etc. Whilst most work thus far focused on the improvement of ASR systems embedding Gaussian Mixture Models (GMM)s to compute the acoustic likelihoods in the states of a Hidden Markov Model (HMM), the present work focuses on the noise robustness of systems employing Reservoir Computing (RC) as an alternative acoustic modeling technique. Previous work already demonstrated good noise robustness for continuous digit recognition (CDR). The present paper investigates whether further progress can be achieved by driving reservoirs with noise-robust inputs that have been shown to raise the robustness of GMM-based systems, by introducing bi-directional reservoirs and by combining reservoirs with GMMs in a single system. Experiments on Aurora-2 demonstrate that it is indeed possible to raise the noise robustness without significantly increasing the system complexity.

Downloads

  • IEEE Passed File.pdf
    • full text
    • |
    • open access
    • |
    • PDF
    • |
    • 376.87 KB

Citation

Please use this url to cite or link to this publication:

Chicago
Jalalvand, Azarakhsh, Kris Demuynck, and Jean-Pierre Martens. 2013. “Noise Robust Continuous Digit Recognition with Reservoir-based Acoustic Models.” In International Symposium on Intelligent Signal Processing and Communication Systems ISPACS, 204–209. New York, NY, USA: IEEE.
APA
Jalalvand, A., Demuynck, K., & Martens, J.-P. (2013). Noise robust continuous digit recognition with reservoir-based acoustic models. International Symposium on Intelligent Signal Processing and Communication Systems ISPACS (pp. 204–209). Presented at the 2013 International symposium on Intelligent Signal Processing and Communication Systems (ISPACS 2013), New York, NY, USA: IEEE.
Vancouver
1.
Jalalvand A, Demuynck K, Martens J-P. Noise robust continuous digit recognition with reservoir-based acoustic models. International Symposium on Intelligent Signal Processing and Communication Systems ISPACS. New York, NY, USA: IEEE; 2013. p. 204–9.
MLA
Jalalvand, Azarakhsh, Kris Demuynck, and Jean-Pierre Martens. “Noise Robust Continuous Digit Recognition with Reservoir-based Acoustic Models.” International Symposium on Intelligent Signal Processing and Communication Systems ISPACS. New York, NY, USA: IEEE, 2013. 204–209. Print.
@inproceedings{4192249,
  abstract     = {Notwithstanding the many years of research, more work is needed to create automatic speech recognition (ASR) systems with a close-to-human robustness against confounding factors such as ambient noise, channel distortion, etc. Whilst most work thus far focused on the improvement of ASR systems embedding Gaussian Mixture Models (GMM)s to compute the acoustic likelihoods in the states of a Hidden Markov Model (HMM), the present work focuses on the noise robustness of systems employing Reservoir Computing (RC) as an alternative acoustic modeling technique. Previous work already demonstrated good noise robustness for continuous digit recognition (CDR). The present paper investigates whether further progress can be achieved by driving reservoirs with noise-robust inputs that have been shown to raise the robustness of GMM-based systems, by introducing bi-directional reservoirs and by combining reservoirs with GMMs in a single system. Experiments on Aurora-2 demonstrate that it is indeed possible to raise the noise robustness without significantly increasing the system complexity.},
  author       = {Jalalvand, Azarakhsh and Demuynck, Kris and Martens, Jean-Pierre},
  booktitle    = {International Symposium on Intelligent Signal Processing and Communication Systems ISPACS},
  isbn         = {9781467363617},
  language     = {eng},
  location     = {Naha, Okinawa, Japan},
  pages        = {204--209},
  publisher    = {IEEE},
  title        = {Noise robust continuous digit recognition with reservoir-based acoustic models},
  url          = {http://dx.doi.org/10.1109/ISPACS.2013.6704547},
  year         = {2013},
}

Altmetric
View in Altmetric
Web of Science
Times cited: