Efficient prior and incremental beam width control to suppress excessive speech recognition time based on score range estimation.
Satoshi KobashikawaTakaaki HoriYoshikazu YamaguchiTaichi AsamiHirokazu MasatakiSatoshi TakahashiPublished in: SLT (2012)
Keyphrases
- speech recognition
- language model
- speech synthesis
- speech signal
- pattern recognition
- speech processing
- speech recognizer
- handwriting recognition
- automatic speech recognition
- speech understanding
- noisy environments
- hidden markov models
- speech recognition technology
- speaker identification
- keyword spotting
- speech recognizers
- speaker dependent
- information retrieval
- speaker recognition
- bayesian networks
- computer vision
- data mining
- isolated word