Unsupervised learning of finite full covariance multivariate generalized Gaussian mixture models for human activity recognition.
Fatma NajarSami BourouisNizar BouguilaSafya BelghithPublished in: Multim. Tools Appl. (2019)
Keyphrases
- human activity recognition
- unsupervised learning
- gaussian mixture model
- finite mixture models
- mixture model
- expectation maximization
- finite mixtures
- activity recognition
- covariance matrices
- human activities
- minimum message length
- supervised learning
- em algorithm
- semi supervised
- gaussian mixture
- machine learning
- gaussian distribution
- semi supervised learning
- mixture modeling
- speaker recognition
- dimensionality reduction
- density estimation
- feature space
- visual surveillance
- maximum likelihood
- model selection
- background subtraction
- feature vectors
- active learning
- text classification
- action recognition
- covariance matrix
- surveillance system
- image segmentation
- sensor networks
- decision trees
- pairwise
- probabilistic model
- computer vision