Lucent Automatic speech Recognition: A Speech Recognition Engine for Internet and telephony Service Applications

摘要

Based on Bell Labs speech recognition and understanding technology, we developed LASR3 (Lucent automatic Speech Recognition, Version 3), a speaker independent, software-based continuous speech recognition engine. It is compatible with Microsoft Speech Application Programming Interface (MS SAPI) [1]. LASR3 provides support for desktop, telephony, and internet applications requiring speech recognition. It is a multi-lingual, scalable multi-channel speech recognition engine. LASR3 runs on 32-bit Microsoft Windows platoforms. This paper discusses the LASR3 algorithms, features, and architecture design for scalable multi-channel speech recognition.