Latent Dirichlet Decomposition for Single Channel Speaker Separation.
Bhiksha RajMadhusudana V. S. ShashankaParis SmaragdisPublished in: ICASSP (5) (2006)
Keyphrases
- single channel
- latent dirichlet
- multi channel
- sound source
- audio visual
- independent component analysis
- em algorithm
- prior information
- speech recognition
- wiener filter
- frequency domain
- latent dirichlet allocation
- signal processing
- feature selection
- mixture model
- automatic speech recognition
- hidden markov models
- multi modal
- expectation maximization
- pattern recognition
- multiscale
- image segmentation