Batch and Online Underdetermined Source Separation Using Laplacian Mixture Models.
Nikolaos MitianoudisTania StathakiPublished in: IEEE Trans. Speech Audio Process. (2007)
Keyphrases
- mixture model
- source separation
- mixing matrix
- blind source separation
- gaussian mixture model
- em algorithm
- generative model
- probabilistic model
- independent component analysis
- expectation maximization
- language model
- model selection
- maximum likelihood
- denoising
- audio features
- temporal structure
- information retrieval
- non stationary
- speech recognition
- multi modal
- audio visual
- low level
- prior knowledge
- source signals
- pattern recognition