2Yuk D S, Che C W, Jin L. Environment-independent continuous speech recognition using neural networks and hidden markov models. IEEE International Conference on Acoustics, Speech, and Signal Processing, 1996, 6: 3358～3361
3Markov K, Nakamura S. Hybrid HMM/BN LVCSR system integrating multiple acoustic features. IEEE International Conference on Acoustics, Speech, and Signal Processing, 2003, 1(3): 840-843
