A nested infinite Gaussian mixture model for identifying known and unknown audio events.
Yoko SasakiKazuyoshi YoshiiSatoshi KagamiPublished in: WIAMIS (2013)
Keyphrases
- gaussian mixture model
- speaker identification
- mixture model
- mouth region
- speaker recognition
- em algorithm
- feature vectors
- expectation maximization
- event detection
- soccer video
- gaussian mixture
- background subtraction
- maximum likelihood
- gaussian distribution
- visual information
- mixture of gaussians
- gaussian mixture modeling
- feature space
- image segmentation
- emotion recognition
- active learning