Analysis of HMM Temporal Evolution for Automatic Speech Recognition and Verification

  • Marta Casar
  • José A. R. Fonollosa
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4188)


This paper proposes a double layer speech recognition and utterance verification system based on the analysis of the temporal evolution of HMM’s state scores. For the lower layer, it uses standard HMM-based acoustic modeling, followed by a Viterbi grammar-free decoding step which provides us with the state scores of the acoustic models. In the second layer, these state scores are added to the regular set of acoustic parameters, building a new set of expanded HMMs. This new paremeter models the acoustic HMM’s temporal evolution. Using the expanded set of HMMs for speech recognition a significant improvement in performance is achieved. Next, we will use this new architecture for utterance verification in a ”second opinion” framework. We will consign to the second layer evaluating the reliability of decoding using the acoustic models from the first layer. An outstanding improvement in performance versus a baseline verification system has been achieved with this new approach.


Recognition Rate Speech Recognition State Score Automatic Speech Recognition Acoustic Model 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Huang, et al.: Spoken Language Processing, 1st edn. Prentice Hall PTR, Englewood Cliffs (2001)Google Scholar
  2. 2.
    Demuynck, et al.: Flavor: a flexible architecture for LVCSR. In: EUROSPEECH (2003)Google Scholar
  3. 3.
    Cox, Dasmahapatra: High-level approaches to confidence estimation in speech recognition. IEEE Trans. on Speech and Audio Processing 10, 460–471 (2002)CrossRefGoogle Scholar
  4. 4.
    Hernández-Ábrego, Mariño: A second opinion approach for speech recognition verification. In: Proceedings of the VIII SNRFAI, vol. I (1999)Google Scholar
  5. 5.
    Stemmer, et al.: Context-dependent output densities for Hidden Markov Models in speech recognition. In: EUROSPEECH (2003)Google Scholar
  6. 6.
    Moreno, Winksky: Spanish fixed network speech corpus. SpeechDat Project. LRE-63314 (1999)Google Scholar
  7. 7.
    Bonafonte, et al.: Ramses: el sistema de reconocimiento del habla continua y gran vocabulario desarrollado por la UPC. VIII Jornadas de Telecom I+D (1998)Google Scholar
  8. 8.
    Young: Detecting misrecognitions and out-of-vocabulary words. In: ICASSP, vol. I (1994)Google Scholar
  9. 9.
    Jiang, Huang: Vocabulary-independent word confidence measure using subword features. In: ICSLP (1998)Google Scholar
  10. 10.
    Rahim, et al.: Discriminative utterance verification using minimum string verification error (MSVE) training. In: ICASSP (1996)Google Scholar
  11. 11.
    Schaaf, Kemps: Confidence measures for spontaneous speech recognition. In: ICASSP, vol. II (1997)Google Scholar
  12. 12.
    Sanchis: PhD Thesis. Estimation and application of confidence measures for speech recognition (in Spanish). PhD thesis, Universidad Politécnica de Valencia. Departamento de Sistemas Informáticos y Computación (2004)Google Scholar
  13. 13.
    Egan: Signal detection theory and ROC analysis. Academic Press, London (1975)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2006

Authors and Affiliations

  • Marta Casar
    • 1
  • José A. R. Fonollosa
    • 1
  1. 1.TALP Research Center, Dept. of Signal Theory and CommunicationsUniversitat Politècnica de CatalunyaBarcelonaSpain

Personalised recommendations