Adaptive blind audio source extraction supervised by dominant speaker identification using x-vectors.
Jakub JanskýJirí MálekJaroslav CmejlaTomás KounovskýZbynek KoldovskýJindrich ZdánskýPublished in: CoRR (2019)
Keyphrases
- speaker identification
- speech recognition
- gaussian mixture model
- speech processing
- speaker recognition
- speech signal
- feature extraction
- broadcast news
- noisy environments
- audio signals
- audio signal
- audio features
- unsupervised learning
- neural network
- feature selection
- supervised learning
- feature vectors
- audio visual
- semi supervised
- mixture model
- em algorithm
- multi modal
- expectation maximization
- maximum likelihood
- natural language processing
- information extraction