论文信息 - A second-order HMM for high performance word and phoneme-based continuous speech recognition

A second-order HMM for high performance word and phoneme-based continuous speech recognition

In the field of speech recognition by stochastic methods, it is conventional to pursue approaches using first-order-hidden Markov models (HMM1s). Despite the success of this approach, it is still worth investigating if some of the drawbacks of HMM1s can be overcome, e.g. by using higher-order Markov processes. In this paper, we show that second-order hidden Markov models (HMM2s) can yield high performances in the context of continuous speech recognition. We first present the underlying equations and complexity of HMM2s in the maximum likelihood estimation (MLE) paradigm. Then, we show that in a connected word recognition task, such as spelled name recognition over the telephone, HMM2s outperform HMM1s. In the field of phoneme-based continuous speech recognition, we show that context-independent HMM2s can achieve more than 69% phone accuracy.

[1] Hsiao-Wuen Hon,et al. Speaker-independent phone recognition using hidden Markov models , 1989, IEEE Trans. Acoust. Speech Signal Process..

[2] Abdelaziz Kriouile,et al. Some improvements in speech recognition algorithms based on HMM , 1990, International Conference on Acoustics, Speech, and Signal Processing.

[3] Ronald A. Cole,et al. A telephone speech database of spelled and spoken names , 1992, ICSLP.

[4] Jean-Luc Gauvain,et al. High performance speaker-independent phone recognition using CDHMM , 1993, EUROSPEECH.

[5] Jean-François Mari,et al. Hidden Markov models and selectively trained neural networks for connected confusable word recognition , 1994, ICSLP.

[6] Abdelaziz Kriouile,et al. Automatic word recognition based on second-order hidden Markov models , 1994, IEEE Trans. Speech Audio Process..

[7] Li Deng,et al. Use of generalized dynamic feature parameters for speech recognition: maximum likelihood and minimum classification error approaches , 1995, 1995 International Conference on Acoustics, Speech, and Signal Processing.

[8] Jean-François Mari,et al. An N-best strategy, dynamic grammars and selectively trained neural networks for real-time recognition of continuously spelled names over the telephone , 1995, 1995 International Conference on Acoustics, Speech, and Signal Processing.