Generative model-based speaker clustering via mixture of von Mises-Fisher distributions.
Hao TangStephen M. ChuThomas S. HuangPublished in: ICASSP (2009)
Keyphrases
- von mises fisher
- mixture model
- unit sphere
- generative model
- multi variate
- noise model
- unsupervised classification
- gaussian mixture model
- unsupervised learning
- partially supervised
- probabilistic model
- expectation maximization
- model selection
- em algorithm
- gaussian distribution
- pattern recognition
- image processing
- image segmentation
- noise level
- high dimensional
- computer vision
- semi supervised
- hidden markov models