On Dykstra's algorithm: finite convergence, stalling, and the method of alternating projections.
Heinz H. BauschkeRegina Sandra BurachikDaniel B. HermanC. Yalçin KayaPublished in: Optim. Lett. (2020)
Keyphrases
- high accuracy
- improved algorithm
- convergence rate
- preprocessing
- iterative algorithms
- dynamic programming
- experimental evaluation
- optimization algorithm
- computational complexity
- detection algorithm
- computational efficiency
- high efficiency
- objective function
- detection method
- computational cost
- rapid convergence
- clustering method
- theoretical analysis
- synthetic and real images
- segmentation algorithm
- cost function
- computationally efficient
- recognition algorithm
- optimization method
- probabilistic model
- segmentation method
- matching algorithm
- selection algorithm
- convergence property
- similarity measure
- classification algorithm
- em algorithm
- classification method
- support vector machine svm
- region of interest
- convergence theorem
- expectation maximization
- k means
- significant improvement
- estimation algorithm
- input data
- tree structure
- learning algorithm
- single pass
- np hard
- motion estimation
- simulated annealing
- projection matrix
- convergence speed
- particle filter
- convex hull