A novel multi-resolution based PSD estimation method based on expectation maximization algorithm.
Yasin MiarClaude D'AmoursTyseer AboulnasrPublished in: WCNC (2014)
Keyphrases
- expectation maximization
- em algorithm
- parameter estimation
- dynamic programming
- high accuracy
- estimation algorithm
- detection method
- clustering method
- significant improvement
- experimental evaluation
- maximum likelihood estimation
- computationally efficient
- cost function
- objective function
- optimization algorithm
- computational cost
- theoretical analysis
- k means
- support vector machine svm
- detection algorithm
- improved algorithm
- computational complexity
- gaussian mixture
- optimization method
- synthetic and real images
- segmentation method
- classification method
- convergence rate
- recognition algorithm
- probabilistic model
- multiresolution
- matching algorithm
- classification algorithm
- tree structure
- coarse to fine strategy
- segmentation algorithm
- preprocessing
- bayesian framework
- selection algorithm
- learning algorithm
- estimation accuracy
- similarity measure
- hierarchical representation
- particle swarm optimization
- constrained optimization
- closed form
- map estimation
- kalman filter
- newton raphson
- support vector machine
- decision directed