Intelligibility Enhancement Via Normal-to-Lombard Speech Conversion With Long Short-Term Memory Network and Bayesian Gaussian Mixture Model.
Gang LiXiaochen WangRuimin HuHuyin ZhangShanfa KePublished in: IEEE Trans. Multim. (2021)
Keyphrases
- gaussian mixture model
- speaker recognition
- speaker identification
- maximum likelihood
- mixture model
- speech recognition
- expectation maximization
- em algorithm
- feature space
- gaussian distribution
- probability density function
- background subtraction
- feature vectors
- gaussian mixture modeling
- image processing
- probability distribution
- pairwise
- pattern recognition
- multiscale
- speech signal
- gaussian model
- bayesian networks
- speaker verification
- mixture distribution
- learning algorithm