A typical speech recognition (SR) engine can be classed according to three criteria; speaker dependence, mode of operation and vocabulary size.
