R. Pieraccini and A. E. Rosenberg. Coarticulation models for continuous digit recognition. In Proc. Acoust. Soc. Am., page 106, May 1990. Home/Search Document Not in Database Summary Related Articles |
This paper is cited in the following contexts:
Discriminative Feature And Model Design For Automatic.. - Rahim, Bengio, LeCun (1997) (2 citations) (Correct)
....time derivatives, resulting in a vector of 39 features per frame. Following feature analysis, each feature vector in the baseline system is directly passed to the recognizer which models each word (i.e. digit) in the vocabulary by a set of left to right continuous density quasitriphonic HMMs [7]. Each word is divided into three units, namely, head, body and tail. To model interword coarticulation, each word is made to have a single body with multiple heads and tails, resulting in a total of 274 subword models. Each subword model consists of 3 to 4 states, with each state having a mixture ....
R. Pieraccini and A. E. Rosenberg. Coarticulation models for continuous digit recognition. In Proc. Acoust. Soc. Am., page 106, May 1990.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC