Name
Abstract | ||
|---|---|---|
When moving a speech recognition system whose models were trained in a clean laboratory condition to real environments, one of most important issues is how to modify the models according to the changing environments. Using an HMM composition technique we present an algorithm to compensate the dynamic cepstral coefficients for HMM based speech recognition systems in noise environments. Noise compensation for acceleration parameters and for dynamic parameters which are calculated using longer linear regression are discussed. The experimental results show a clear improvement when the algorithm was applied to a speech database recorded in a car. A noise compensation system based realtime speech recognizer using the TMS320C40 was implemented and achieves a good performance in noisy environments. |
Year | DOI | Venue |
|---|---|---|
1996 | 10.1109/ICASSP.1996.540287 | ICASSP |
Keywords | Field | DocType |
acceleration parameter,improved noise compensation algorithm,speech recognition system,noise compensation system,hmm composition technique,dynamic parameter,noise compensation,speech database,realtime speech recognizer,dynamic cepstral coefficient,noise environment,speech processing,parameter estimation,hidden markov models,speech recognition,covariance matrix,acoustic noise,acceleration,vectors,linear regression | Speech enhancement,Mel-frequency cepstrum,Speech processing,Computer science,Artificial intelligence,Estimation theory,Noise,Pattern recognition,Algorithm,Speech recognition,Acceleration,Covariance matrix,Hidden Markov model | Conference |
ISBN | Citations | PageRank |
0-7803-3192-3 | 8 | 1.51 |
References | Authors | |
7 | 2 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Ruikang Yang | 1 | 82 | 12.85 |
| P. Haavisto | 2 | 46 | 8.00 |