Usable speech detection using a context dependent Gaussian mixture model classifier.
Robert E. YantornoBrett Y. SmolenskiAnanth N. IyerJashmin K. ShahPublished in: ISCAS (5) (2004)
Keyphrases
- gaussian mixture model
- context dependent
- speaker recognition
- speaker identification
- mixture model
- feature space
- probabilistic neural network
- feature vectors
- maximum likelihood
- em algorithm
- mouth region
- background subtraction
- low level
- expectation maximization
- detection method
- gaussian mixture
- speech recognition
- high level
- linear classifiers
- feature selection
- object detection
- feature set
- object detectors
- bayesian information criterion
- generative model
- support vector machine
- probabilistic model
- speech signal
- pairwise
- broadcast news
- training set
- object recognition
- training data
- gaussian mixture modeling
- learning algorithm