A probabilistic union model with automatic order selection for noisy speech recognition

A critical issue in exploiting the potential of the sub-band-based approach to robust speech recognition is the method of combining the sub-band obser vations, for selecting the bands unaffected by noise. A new method for this purpose, i.e., the probabilistic union model, was recently introduced. Thi s model has been shown to be capable of dealing with band-limited corruptio n., requiring no knowledge about the band position and statistical distribu tion of the noise. A parameter within the model, which we call its order, g ives the best results when it equals the number of noisy bands. Since this information may not be available in practice, in this paper we introduce an automatic algorithm for selecting the order, based on the state duration p attern generated by the hidden Markov model (HMM). The algorithm has been t ested on the TIDIGITS database corrupted by various types of additive band- limited noise with unknown noisy bands. The results have shown that the uni on model equipped with the new algorithm can achieve a recognition performa nce similar to that achieved when the number of noisy bands is known. The r esults show a very significant improvement over the traditional full-band m odel, without requiring prior information on either the position or the num ber of noisy bands. The principle of the algorithm for selecting the order based on state duration may also be applied to other sub-band combination m ethods. (C) 2001 Acoustical Society of America.