Recursive noise estimation using iterative stochastic approximation for stereo-based robust speech recognition
Tóm tắt
We present an algorithm for recursive estimation of parameters in a mildly nonlinear model involving incomplete data. In particular, we focus on the time-varying deterministic parameters of additive noise in the nonlinear model. For the nonstationary noise that we encounter in robust speech recognition, different observation data segments correspond to different noise parameter values. Hence, recursive estimation algorithms are more desirable than batch algorithms, since they can be designed to adaptively track the changing noise parameters. One such design based on the iterative stochastic approximation algorithm in the recursive-EM framework is described. This new algorithm jointly adapts time-varying noise parameters and the auxiliary parameters introduced to give a linear approximation of the nonlinear model. We present stereo-based robust speech recognition results for the AURORA task, which demonstrate the effectiveness of the new algorithm compared with a more traditional, MMSE noise estimation technique under otherwise identical experimental conditions.
Từ khóa
#Recursive estimation #Stochastic resonance #Noise robustness #Speech enhancement #Working environment noise #Iterative algorithms #Testing #Cepstral analysis #Acoustic noise #Piecewise linear approximationTài liệu tham khảo
10.1109/ICASSP.2001.940827
deng, 2000, Large-vocabulary speech recognition under adverse acoustic environments, Proc lCSLP, 3, 806
titterington, 1984, Recursive parameter estimation using incomplete data, J Royal Stat Soc, 46(b), 257
frey, 2001, ALGO-NQUIN: Iterating Laplace's method to remove multiple types of acoustic distortion for robust speech recognition, Proc EUROSPEECH, 10.21437/Eurospeech.2001-273
droppo, 0, Evaluation of the SPLICE algorithm on the Amora2 database, Proc Eurospeech 2001
10.1109/78.229888
kim, 1998, Nonstationary environment compensation based on sequential estimation, IEEE Sig Proc Letters, 5, 57, 10.1109/97.661559
10.1109/ICASSP.2001.940809
10.1109/ICASSP.1996.543225
acero, 2000, HMM adaptation using vector Taylor series for noisy speech recognition, Proc lCSLP, 3, 869