INTERSPEECH 2004 - ICSLP
Noise compensation methods for speech recognition provide a cleaned version of the speech representation. Usually this version is the expected value of the speech parameters given the observed noisy speech and the noise statistic. A more realistic representation should include a probability distribution of the cleaned speech instead of its expected value in order to represent the uncertainty associated to the compensation process due to the variability of the noise process. Recently, the inclusion of the uncertainty in the recognition process has been studied. In this paper we have developed a noise compensation technique that incorporates the variance of the cleaned speech into the speech representation. The variance is estimated using a Wiener filter during the speech feature enhancement process. This way of including the uncertainty implies the modification of the decoding rule. Experimental results demonstrate a improvement of the performance in the recognition system (about 21% WER reduction) when uncertainty is included.
Bibliographic reference. Segura, Josť Carlos / Torre, Angel De la / Ramirez, Javier / Rubio, Antonio J. / Benitez, Carmen (2004): "Including uncertainty of speech observations in robust speech recognition", In INTERSPEECH-2004, 137-140.