In this paper is presented analyses in automatic speech recognition (ASR) to find out what is the state of the arts in this direction and, eventually, it can be a starting point for the implementation of a real ASR system. In the second chapter of this work, it is revealed the structure of a typical speech recognition system and the used methods for each step of the recognition process, and in special, there are described two kinds of speech recognition algorithms, namely, Dynamic Time Warping (DTW) and Hidden Markov Model (HMM). The work continues with some results of ASR, in order to make conclusions about what is needed to be improved and what is more eligible to implement an ASR system.
|