An EKF-based algorithm for learning statistical hidden dynamic model parameters for phonetic recognition

ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

Volumn 1, Issue , 2001, Pages 465-468

  Togneri, Roberto ^a   Deng, Li ^b,c  

^a UNIVERSITY OF WESTERN AUSTRALIA   (Australia)

^b UNIVERSITY OF WATERLOO   (Canada)

^c MICROSOFT RESEARCH   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EXTENDED KALMAN FILTER; HIDDEN DYNAMIC MODEL; PARAMETER VECTORS; PHONETIC RECOGNITION;

ALGORITHMS; KALMAN FILTERING; MATHEMATICAL MODELS; PARAMETER ESTIMATION; SPEECH ANALYSIS; STATISTICAL METHODS;

SPEECH RECOGNITION;

EID: 0034855057     PISSN: 15206149     EISSN: None     Source Type: Journal    
DOI: 10.1109/ICASSP.2001.940868     Document Type: Article

References (9)

1. A dynamic, feature-based approach to the interface between phonology and phonetics for speech modeling and recognition
2. A statistical coarticulatory model for the hidden vocal-tract-resonance dynamics
3. Spontaneous speech recognition using a statistical coarticulatory model for the hidden vocal-tract-resonance dynamics
4. Multistage coarticulation model combining articulatory, formant and cepstral features
5. Asymptotic behavior of the extended Kalman filter as a parameter estimator for linear systems
6. A path-stack algorithm for optimizing dynamic regimes in a statistical hidden dynamic model of speech
7. Lessons in estimation theory for signal processing, communications, and control
8. Initial evaluation of hidden dynamic models on conversational speech
9. The HDM: A segmental hidden dynamic model of coarticulation