Blind speech signal separation based on non-stationary and colored characteristics.
Fengyu CongYanxue LiangShaoling JiYifan HuXizhi ShiPublished in: ISSPA (2005)
Keyphrases
- non stationary
- speech signal
- sound source
- automatic speech recognition
- noisy environments
- background noise
- speaker identification
- speaker recognition
- adaptive algorithms
- linear prediction
- speech recognition
- autoregressive
- automatic speech recognition systems
- blind source separation
- spectral analysis
- acoustic features
- empirical mode decomposition
- speech enhancement
- change point detection
- feature extraction
- random fields
- language model
- hidden markov models