ARMA lattice model for phoneme feature extraction

In this paper, the result of a study on phoneme feature extraction, under a noisy environment, using an auto-regressive moving average (ARMA) lattice model, is presented. The phoneme characteristics are modeled and expressed in the form of ARMA lattice reflection coefficients for classification. Experimental results, based on the TIMIT speech database and NoiseX-92 noise database, indicate that the ARMA lattice model achieves an improved noise-resistant capability on vowel phonemes and fricative phonemes as compared to those of the conventional mel-frequency cepstral coefficient (MFCC) method.

[1]  Shu Hung Leung,et al.  Feature extraction based on perceptually non-uniform spectral compression for speech recognition , 2003, Proceedings of the 2003 International Symposium on Circuits and Systems, 2003. ISCAS '03..

[2]  Monson H. Hayes,et al.  Statistical Digital Signal Processing and Modeling , 1996 .

[3]  Nobuhiro Miki,et al.  ARMA digital lattice filter based on new criterion , 1987 .

[4]  Hon Keung Kwan,et al.  Minimal normalized lattice structure for ARMA digital filter realization , 1991, [Proceedings] ICASSP 91: 1991 International Conference on Acoustics, Speech, and Signal Processing.

[5]  Hon Keung Kwan,et al.  Normalized ARMA Schur algorithm , 1989, Proceedings of the 32nd Midwest Symposium on Circuits and Systems,.

[6]  Y. Lim,et al.  On the synthesis of lattice parameter digital filters , 1983, ICASSP.

[7]  Hon Keung Kwan,et al.  ARMA lattice modeling for isolated word speech recognition , 2000, Proceedings of the 43rd IEEE Midwest Symposium on Circuits and Systems (Cat.No.CH37144).

[8]  Steven Kay,et al.  Fundamentals Of Statistical Signal Processing , 2001 .

[9]  M. Morf,et al.  Recursive ladder algorithms for ARMA modeling , 1980, 1980 19th IEEE Conference on Decision and Control including the Symposium on Adaptive Processes.