Enhanced power-normalized features for mandarin robust speech recognition based on a voiced-unvoiced-silence decision.
Ying-Wei TanWenju LiuZhanlei YangMingming ChenPublished in: ChinaSIP (2014)
Keyphrases
- speech recognition
- speech signal
- noisy environments
- speech recognition systems
- speech processing
- automatic speech recognition
- cepstral coefficients
- speaker identification
- hidden markov models
- language model
- mel frequency cepstral coefficients
- background noise
- speech recognizer
- vocal tract
- speaker independent
- pattern recognition
- speech enhancement
- speaker recognition
- speech synthesis
- acoustic features
- feature set
- speech recognition technology
- feature extraction
- extracted features
- feature space
- speaker dependent
- speaker diarization
- feature vectors
- speech recognizers