The Projected Power Method: An Efficient Algorithm for Joint Alignment from Pairwise Differences.
Yuxin ChenEmmanuel J. CandèsPublished in: CoRR (2016)
Keyphrases
- pairwise
- high accuracy
- cost function
- preprocessing
- detection algorithm
- computationally efficient
- clustering method
- dynamic programming
- experimental evaluation
- similarity measure
- segmentation algorithm
- computational cost
- estimation algorithm
- selection algorithm
- global alignment
- optimization method
- computational complexity
- synthetic and real images
- matching algorithm
- significant improvement
- theoretical analysis
- high efficiency
- k means
- recognition algorithm
- single pass
- improved algorithm
- similarity matrix
- highly efficient
- detection method
- support vector machine svm
- optimization algorithm
- input data
- classification method
- objective function
- sequence alignment
- tree structure
- probabilistic model
- segmentation method
- dissimilarity matrix
- classification algorithm
- support vector machine
- image alignment
- convergence rate
- simulated annealing
- semi supervised
- reconstruction method
- em algorithm
- energy function
- absolute difference
- search space
- expectation maximization
- particle swarm optimization
- np hard
- matrix representation
- error accumulation
- similarity function