Tony Robinson
ABSTRACT
This paper presents a hybrid system using a connectionist model and a Markov model for the DARPA Resource Management task of large-vocabulary multiple-speaker continuous speech recognition. The connectionist model employs internal feedback for context modelling and provides phone state occupancy probabilities for a simple context independent Markov model. The system has been implemented in real-time on a workstation supported by a DSP board. The use of context independent phone models leads to the possibility of time-domain pruning and computationally efficient durational modelling, both of which are reported in the paper.
References
- J. S. Bridle and L. Dodd. An Alphanet approach to optimising input transformations for continuous speech recognition. In Proc. ICASSP, pages 277-280, 1991. Google Scholar
Cross Ref
- H. Gu, C. Tseng, and L. Lee. Isolated-utterance speech recognition using hidden Markov models with bounded state durations. IEEE Transactions on Signed Processing, 39(8):1743-1752, Aug. 1991.Google ScholarCross Ref
- V. N. Gupta, M. Lennig, P. Mermelstein, P. Kenny, F. Seitz, and D. O'Shaughnessy. Using phoneme duration and energy countour information to improve large vocabulary isolated-word recognition. In Proc. ICASSP, pages 341-344, 1991. Google Scholar
- F. Kubala and R. Schwartz. A new paradigm for speaker-independent training. In Proc. ICASSP, pages 833-836, 1991. Google ScholarCross Ref
- K.-F. Lee. Automatic Speech Recognition: The Development of the SPHINX System. Kluwer Academic Publishers, Boston, 1989. Google ScholarDigital Library
- S. E. Levinson. Continuously variable duration hidden Markov models for automatic speech recognition. Computer Speech and Language, 1(1):29-45, Mar. 1986. Google ScholarDigital Library
- N. Morgan and H. Bourlard. Continuous speech recognition uing multilayer perceptrons With hidden Markov models. In Proc. ICASSP, pages 413-416 1990. 1990.Google Scholar
- K. M. Ponting and S. M. Peeling. The use of variable frame rat.e analysis in speech recognition. Computer Speech and Longuage, 5:169-179, 1991.Google ScholarCross Ref
- P. Price, W. M. Fisher, J. Bernstein, and D. S. Pallett. The DARPA 1000-word Resource Management database for continuous speech recognition. In Proc. ICASSP, pages 651-654, 1988.Google ScholarCross Ref
- T. Robinson. Several improvements to a recurrent error propagation network phone recognition system. Technical Report CUED/F-INFENG/TR.82, Cambridge University Engineering Department, Sept. 1991.Google Scholar
- T. Robinson and F. Fallside. A recurrent error propagation network Speech recognition system. Computer Speech and Language, 5(3):259-274, July 1991.Google ScholarCross Ref
Index Terms
(auto-classified)A real-time recurrent error propagation network word recognition system
Comments