Speaker adaptation intensively weighted on mis-recognized speech segments.
Takahiro OkuYuya FujitaAkio KobayashiToru ImaiPublished in: APSIPA (2012)
Keyphrases
- speaker adaptation
- automatic speech recognition
- speech segments
- speech recognition
- speech retrieval
- maximum likelihood
- speech signal
- hidden markov models
- multi modal
- neural network
- feature vectors
- speech recognizer
- speaker independent
- feature space
- emotional speech
- question answering
- broadcast news
- speaker diarization
- audio stream