Speech enhancement for linear-predictive-analysis-by-synthesis coders
Marcin Kuropatwinski, Dieter Leckschat, Kristian Kroschel, Andrzej Czyzewski, Chaz Hales · 1999
In the conventional hidden Markov model, the model parameters are reestimated by an iterative procedure known as the Baum-Welch method. This paper proposes an alternative procedure using fuzzy estimation, which is generalised from the fuzzy c-means and the BaumWelch methods. An extension of this approach, which uses a garbage state to deal with outlier data is also proposed. Experiments using the TI46 speech data corpus show this approach can be applicable to speech and speaker recognition.