A type of speech recognition which can recognise speech as it is naturally spoken rather. than. spoken. one. word. at. a. time. fullstop. This term is used to distinguish it from isolated word recognition. It's what is used in commercial products such as IBM's ViaVoice and Dragon's Naturally Speaking; although typically systems capable of continuous speech recognition can also be used for isolated word recognition.

They are usually implemented using hidden Markov models, together with some tricks such as bigrams and use of production rules to enhance accuracy.

Log in or register to write something here or to contact authors.