DNN-based feature transformation for speech recognition using throat microphone.
Shengke LinTakashi TsunakawaMasafumi NishidaMasafumi NishimuraPublished in: APSIPA (2017)
Keyphrases
- feature transformation
- speaker identification
- automatic speech recognition
- speech signal
- speech recognition
- broadcast news
- speaker diarization
- noisy environments
- gaussian mixture model
- feature extraction
- linear classifiers
- scale invariant
- sound source
- feature space
- hidden markov models
- distance metric
- audio visual
- feature vectors
- visual information
- language model
- training process
- maximum likelihood
- preprocessing