ICASSP
2011
IEEE
12 years 9 months ago
2011
IEEE
In this paper, we present a novel technique for modeling the posterior probability estimates obtained from a neural network directly in the HMM framework using the Dirichlet Mixtu...