Corpus ID: 6669963

Speech recognition using HMMs with quantized parameters

@inproceedings{Vasilache2000SpeechRU,
  title={Speech recognition using HMMs with quantized parameters},
  author={Marcel Vasilache},
  booktitle={INTERSPEECH},
  year={2000}
}
  • Marcel Vasilache
  • Published in INTERSPEECH 2000
  • Computer Science
  • In this paper we describe the structure and examine the performance of a recognition engine based on hidden Markov models (HMMs) with quantized parameters (qHMM). The main goal of qHMMs is to enable a low complexity implementation without sacrificing the classification performance. In the tests with a whole word digit dialler engine and a phoneme based isolated word recognizer we managed to preserve the performance of unquantized HMMs with qHMMs having as little as 5 bit for a mean component… CONTINUE READING

    Create an AI-powered research feed to stay up to date with new papers like this posted to ArXiv

    Citations

    Publications citing this paper.
    SHOWING 1-10 OF 18 CITATIONS

    PocketSUMMIT: small-footprint continuous speech recognition

    VIEW 2 EXCERPTS
    CITES BACKGROUND & METHODS
    HIGHLY INFLUENCED

    Multi-rate hmm quantization for speech recognition

    • Marcel Vasilache
    • Computer Science
    • 2008 IEEE International Conference on Acoustics, Speech and Signal Processing
    • 2008
    VIEW 4 EXCERPTS
    CITES BACKGROUND

    Quantized HMMs for low footprint text-to-speech synthesis

    VIEW 12 EXCERPTS
    CITES BACKGROUND, METHODS & RESULTS
    HIGHLY INFLUENCED

    Speaker adaptation of quantized parameter HMMs

    VIEW 3 EXCERPTS
    CITES METHODS & BACKGROUND

    Accelerometer-based activity recognition on a mobile phone using cepstral features and quantized gmms

    VIEW 1 EXCERPT
    CITES BACKGROUND

    Information Distance-Based Subvector Clustering for ASR Parameter Quantization

    VIEW 2 EXCERPTS
    CITES METHODS & BACKGROUND

    Data-Driven Subvector Clustering using the Cross-Entropy Method

    VIEW 1 EXCERPT
    CITES BACKGROUND

    Dynamic vocabulary prediction for isolated-word dictation on embedded devices

    VIEW 2 EXCERPTS
    CITES BACKGROUND

    A high-speed, low-resource ASR back-end based on custom arithmetic

    References

    Publications referenced by this paper.
    SHOWING 1-6 OF 6 REFERENCES

    On the use of scalar quantization for fast HMM computation

    VIEW 1 EXCERPT

    A recursive feature vector normalization approach for robust speech recognition in noise

    • Olli Viikki, David Bye, Kari Laurila
    • Computer Science
    • Proceedings of the 1998 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP '98 (Cat. No.98CH36181)
    • 1998
    VIEW 1 EXCERPT

    Fast decoding techniques for practical real-time speech recognition systems

    • J. Suontausta, J. Hakkinen, O. Viikki
    • In IEEE Automatic Speech Recognition and Understanding Workshop,
    • 1998
    VIEW 1 EXCERPT

    Noise robust speech recognition with state duration constraints

    • Kari Laurila
    • Computer Science
    • 1997 IEEE International Conference on Acoustics, Speech, and Signal Processing
    • 1997
    VIEW 1 EXCERPT

    Efficient speech recognition using subvector quantization and discrete-mixture HMMs

    VIEW 1 EXCERPT

    Vector quantization for the efficient computation of continuous density likelihoods

    • Enrico Bocchieri
    • Computer Science
    • 1993 IEEE International Conference on Acoustics, Speech, and Signal Processing
    • 1993
    VIEW 1 EXCERPT