Cantonese speaker recognition system based on EM algorithm in noisy environments.
Yu FanChin-Ta ChenChih-Chung YangPublished in: DSIT (2021)
Keyphrases
- em algorithm
- noisy environments
- speaker verification
- speech recognition
- speaker identification
- expectation maximization
- automatic speech recognition
- mixture model
- gaussian mixture model
- maximum likelihood
- noise reduction
- noisy speech
- speech enhancement
- maximum likelihood estimation
- generative model
- parameter estimation
- likelihood function
- speech signal
- digit recognition
- gaussian mixture
- feature extraction
- expectation maximisation
- incomplete data
- hyperparameters
- maximum a posteriori
- language model
- audio visual
- speaker diarization
- hidden markov models
- model based clustering
- probabilistic model
- maximum likelihood estimates
- prior knowledge
- finite mixture model
- feature transformation
- feature vectors
- mixture modeling
- unsupervised learning
- penalized likelihood
- image segmentation