Large Vocabulary Continuous Speech Recognition a Review

Considerable progress has been made in speech recognition technology over the last few years and nowhere has this progress been more evident than in the area of Large Vocabulary Recognition LVR Current laboratory systems are capable of transcribing continuous speech from any speaker with average word error rates of between and If speaker adaptation is allowed then after or minutes of speech the error rate will drop well below for most speakers Hitherto LVR systems have been limited to dictation applications since they were speaker dependent and they required words to be spoken with a short pause between them The capability to recognise natural continuous speech input from any speaker however opens up many more applications and as a result LVR technology appears to be on the brink of widespread deployment across a range of Information Technology IT systems This article will discuss the principles and architecture of current LVR systems and identify the key issues a ecting their future deployment To illustrate the various points raised the Cam bridge University HTK system will be described This is a modern design giving state of the art performance and it is typical of the current generation of recognition systems