Abstract | ||
---|---|---|
Model based compensation schemes are a powerful approach for noise robust speech recognition. Recently there have been a number of investigations into adaptive training, and estimating the noise models used for model adaptation. This paper examines the use of EM-based schemes for both canonical models and noise estimation, including discriminative adaptive training. One issue that arises when estimating the noise model is a mismatch between the noise estimation approximation and final model compensation scheme. This paper proposes FA-style compensation where this mismatch is eliminated, though at the expense of a sensitivity to the initial noise estimates. EM-based discriminative adaptive training is evaluated on in-car and Aurora4 tasks. FA-style compensation is then evaluated in an incremental mode on the in-car task. |
Year | DOI | Venue |
---|---|---|
2011 | 10.1109/ICASSP.2011.5947427 | ICASSP |
Keywords | Field | DocType |
fa-style compensation,speech recognition,jud noise estimation,noise robustness,noise robust speech recognition,canonical models,adaptive training,em-based discriminative adaptive training,factor analysis,vts,aurora4 tasks,noise,speech,second order,canonical model,estimation | Pattern recognition,Noise measurement,Computer science,Load modeling,Speech recognition,Canonical model,Artificial intelligence,Discriminative model | Conference |
ISSN | ISBN | Citations |
1520-6149 E-ISBN : 978-1-4577-0537-3 | 978-1-4577-0537-3 | 6 |
PageRank | References | Authors |
0.63 | 6 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Federico Flego | 1 | 55 | 6.19 |
Mark J. F. Gales | 2 | 3905 | 367.45 |