Incorporating multiple-HMM acoustic modeling in a modular large vocabulary speech recognition system in telephone environment | Publicación