Combining the Likelihood and the Kullback-Leibler Distance in Estimating the Universal Background Model for Speaker Verification Using SVM.
Zhenchun LeiPublished in: ICPR (2010)
Keyphrases
- kullback leibler distance
- speaker verification
- background model
- feature vectors
- background subtraction
- gaussian mixture model
- support vector
- moving objects
- motion detection
- foreground objects
- support vector machine
- knn
- image sequences
- feature extraction
- moving target
- pixel level
- feature selection
- segmentation algorithm
- object detection
- multi class
- video sequences
- maximum likelihood
- computer vision
- mixture model
- covariance matrix
- dimensionality reduction
- moving camera
- training set
- feature space