An expectation maximization approach for formant tracking using a parameter-free non-linear predictor.
Issam BazziAlex AceroLi DengPublished in: ICASSP (1) (2003)
Keyphrases
- parameter free
- expectation maximization
- em algorithm
- categorical data
- fully automatic
- particle filter
- outlier detection
- neural network
- real time
- kalman filter
- probabilistic model
- object tracking
- appearance model
- visual tracking
- moving target
- linear predictors
- mean shift
- bayesian framework
- generative model
- parameter estimation
- gaussian mixture model
- motion model
- mixture model
- active learning
- k means
- pairwise
- particle filtering
- gaussian mixture
- image segmentation
- machine learning
- data mining
- databases