HMM/NN hybrids for continuous speech recognition

The main goal of automatic speech recognition (ASR) is to produce a machine which will recognize accurately normal human speech from any speaker. The recognition system may be classified as speaker-dependent or speaker-independent and isolated-word or connected word. There are three approaches to research in automatic speech recognition (ASR); the acoustic-phonetic approach, the pattern recognition approach, and the database statistical approach. Two approaches of this kind: hidden Markov model (HMM) and artificial neural network (ANN) are presented in this paper.