Noise-robust HMMs based on minimum error classification

The authors compare and contrast the noise-robustness of hidden Markov models (HMMs) trained using a discriminant minimum error classification (MEC) optimization criterion with that of HMMs trained using the conventional maximum likelihood (ML) approach. Isolated word recognition experiments were performed on the ATR 5240 Japanese word database. MEC continuous Gaussian mixture density HMMs trained in a specific noisy environment were found to be more robust to changes in the signal-to-noise ratio than conventional ML HMMs. MEC HMMs trained in various noisy environments were more robust in all environments than conventional ML HMMs.<<ETX>>

[1]  Chin-Hui Lee,et al.  Segmental GPD training of HMM based speech recognizer , 1992, [Proceedings] ICASSP-92: 1992 IEEE International Conference on Acoustics, Speech, and Signal Processing.

[2]  Kunio Nakajima,et al.  Optimal discriminative training for HMMs to recognize noisy speech , 1992, ICSLP.

[3]  L. E. Scales,et al.  Introduction to Non-Linear Optimization , 1985 .

[4]  David Rainton,et al.  Minimum error classification training of HMMs-Implementation details and experimental results.:Implementation details and experimental results , 1992 .

[5]  Biing-Hwang Juang,et al.  New discriminative training algorithms based on the generalized probabilistic descent method , 1991, Neural Networks for Signal Processing Proceedings of the 1991 IEEE Workshop.

[6]  Shigeki Sagayama,et al.  Minimum error classification training of HMMs , 1992 .

[7]  Biing-Hwang Juang,et al.  Vector equalization in hidden Markov models for noisy speech recognition , 1992, [Proceedings] ICASSP-92: 1992 IEEE International Conference on Acoustics, Speech, and Signal Processing.

[8]  Roger K. Moore,et al.  Hidden Markov model decomposition of speech and noise , 1990, International Conference on Acoustics, Speech, and Signal Processing.