作者: Se-Hoon Chin , Heui-Suck Jung , Tae-Young Roh
DOI:
关键词:
摘要: The present invention relates to a continuous speech recognition system that is very robust in noisy environment. In order recognize smoothly environment, the selects call commands, configures minimum network token, which consists of commands and mute intervals including noises, recognizes inputted continuously real time, analyzes reliability from speaker. When speaker delivers command, for detecting interval recognizing environment through real-time measures after by transferring following command speech-recognition engine at moment when command.