One Microphone Blind Dereverberation Based on Quasi-periodicity of Speech Signals.
Tomohiro NakataniMasato MiyoshiKeisuke KinoshitaPublished in: NIPS (2003)
Keyphrases
- speech signal
- automatic speech recognition
- sound source
- speech recognition
- hidden markov models
- speaker identification
- noisy environments
- non stationary
- linear prediction
- background noise
- spectral analysis
- adaptive filtering
- broadcast news
- vocal tract
- additive noise
- automatic speech recognition systems
- sound signals
- speech sounds
- word recognition
- speech quality
- speaker diarization
- multiscale
- noisy speech
- pattern recognition
- blind source separation
- visual information
- dynamic environments