Cumulative moving averaged bottleneck speaker vectors for online speaker adaptation of CNN-based acoustic models.
Tsubasa OchiaiMarc DelcroixKeisuke KinoshitaAtsunori OgawaTaichi AsamiShigeru KatagiriTomohiro NakataniPublished in: ICASSP (2017)
Keyphrases
- speech recognition
- speaker adaptation
- acoustic models
- speaker independent
- speech recognizer
- automatic speech recognition
- speaker dependent
- hidden markov models
- speaker identification
- speech signal
- speech synthesis
- pattern recognition
- digit recognition
- spoken language
- broadcast news
- maximum likelihood
- language model
- speaker recognition
- audio visual
- speech recognition systems
- feature vectors
- machine learning
- noisy environments
- speaker diarization
- feature extraction