Variational Bayesian Model Selection for GMM-Speaker Verification Using Universal Background Model.
Timur PekhovskyAlexandra LokhanovaPublished in: INTERSPEECH (2011)
Keyphrases
- model selection
- background subtraction
- gaussian mixture model
- mixture model
- exponential family
- cross validation
- moving objects
- hyperparameters
- foreground objects
- parameter estimation
- object detection
- expectation maximization
- density estimation
- image sequences
- motion segmentation
- unsupervised learning
- machine learning
- em algorithm
- gaussian process
- feature selection
- bayesian methods
- feature vectors
- matrix factorization
- appearance model
- image processing
- object tracking
- image segmentation
- input image
- probability distribution