Large-vocabulary speaker-independent continuous speech recognition using HMM | IEEE Conference Publication | IEEE Xplore