Skip to main content
Log in

Word spotting based ona posterior measure of keyword confidence

  • Correspondence
  • Published:
Journal of Computer Science and Technology Aims and scope Submit manuscript

Abstract

In this paper, an approach of keyword confidence estimation is developed that well combines acoustic layer scores and syllable-based statistical language model (LM) scores. Ana posteriori (AP) confidence measure and its forward-backward calculating algorithm are deduced. A zero false alarm (ZFA) assumption is proposed for evaluating relative confidence measures by word spotting task. In a word spotting experiment with a vocabulary of 240 keywords, the keyword accuracy under the AP measure is above 94%, which well approaches its theoretical upper limit. In addition, a syllable lattice Hidden Markov Model (SLHMM) is formulated and a unified view of confidence estimation, word spotting, optimal path search, and N-best syllable re-scoring is presented. The proposed AP measure can be easily applied to various speech recognition systems as well.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

References

  1. Nguyen P, Gelin P, Junqua J C, Chien J T. N-best based supervised and unsupervised adaptation for native and non-native speakers in cars. InICASSP99, Phoenix.

  2. Koo M W, Lee C H, Juang B H. A new decoder based on a generalized confidence score. InICASSP98, Seattle, pp. 213–216.

  3. Sukkar R A. Subword-based minimum verification error (SB-MVE) training for task independent utterance verification. InICASSP98, Seattle, pp. 229–232.

  4. Ohtsuki K, Matsutoka T, Matsunaga S, Furui S. Topic extraction with multiple topic-words in broadcast-news speech. InICASSP98, Seattle, pp. 329–332.

  5. Jitsuhiro T, Takahashi S, Alkawa K. Rejection of out-of-vocabulary words using phoneme confidence likelihood. InICASSP98, Seattle, pp. 217–220.

  6. Junqua J C, Haton J P. Robustness in Automatic Speech Recognition—Fundamentals and Applications. Kluwer Academic Publishers, 1996, p. 338.

  7. Rose R, Paul D. A hidden Markov model based keyword recognition system. InICASSP90, pp. 129–132.

  8. Young S R, Ward W. Recognition confidence measures for spontaneous spoken dialog. InEUROSPEECH, 1993, pp. 1177–1179.

  9. Young S R. Estimating recognition confidence: Methods for conjoining acoustics, semantics, pragmatics and discourse. InICSLP, pp. 2159–2162.

  10. Bellman R E. Dynamic Programming. Princeton University Press, Princeton, NJ, 1957.

    Google Scholar 

  11. Yang H R, Wang Z Y, Lu D Y. PinyinN-best search in Mandarin speech recognition.Acta Electronica Sinica, 1999, 27(4): 58–62.

    Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Hao Jie.

Additional information

This work is supported by the National Outstanding Youth Science Fund of China (Grant No.69625103).

HAO Jie received his B.S. degree in radio technology and information systems in 1996 and the M.S. and Ph.D. degrees in signal and information processing in 2001, all from the Department of Electronic Engineering of Tsinghua University, China. His current research interests include speech coding, speech recognition, image coding, multimedia and Internet applications. He has published 6 books and several papers on these areas.

LI Xing received his B.Sc. degree in radio electronics from Tsinghua University in 1982, and his MS. and Ph.D. degrees in electrical engineering from Drexel University, Philadelphia in 1985 and 1989, respectively. He is now a professor of the Department of Electronic Engineering of Tsinghua University and the deputy director of China Education and Research Network (CERNET) Center. He holds memberships of the following: Communication Expert Committee of the China National “863” High-Tech Project, Technical Board of CERNET, China Communication Institute, IEEE Signal Processing Society, IEEE Computer Society, ISOC, Board of Executive Council of APNIC, Board of Steering Committee of APIA, Sigma Xi, and he is a vice chairman of APNG.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Hao, J., Li, X. Word spotting based ona posterior measure of keyword confidence. J. Comput. Sci. & Technol. 17, 491–497 (2002). https://doi.org/10.1007/BF02943289

Download citation

  • Received:

  • Revised:

  • Issue Date:

  • DOI: https://doi.org/10.1007/BF02943289

Keywords

Navigation