81 pages matching word error rate in this book
Results 1-3 of 81
What people are saying - Write a review
We haven't found any reviews in the usual places.
Speaker Speciﬁc Processing
Speech Recognition LVCSR
39 other sections not shown
accented acoustic model adaptation algorithm applied approach ASR system automatic speech recognition barge-in baseline bigram cepstral cepstrum classes classiﬁcation clean speech clustering coefﬁcients components computed conﬁdence measure conﬁdence scoring context corpus database decision tree decoding deﬁned detection dialogue domain estimated evaluation experiments extraction feature vector Figure ﬁltering ﬁnal ﬁnd ﬁrst ﬁxed frame frequency function Gaussian hidden Markov models HLDA IEEE improvement initial input iteration labels language model lattice likelihood linear mapping maximum likelihood method MFCC MLLR n-gram node noise obtained optimal output paper parameters parsing performance phoneme pitch posterior probability probability Proc pronunciation proposed prosody recognition results recognition system recognizer reduced robust segmentation selection semantic sentence sequence shows signiﬁcant signiﬁcantly speaker speciﬁc spectral speech signal spoken syllable Table task techniques test set tion training data training set transcription transform utterances values vocabulary word error rate