A Provably Optimal, Distribution-Independent Parallel Fast Multipole Method.
Fatih Erdogan SevilgenNatsuhiko FutamuraSrinivas AluruPublished in: IPDPS (2000)
Keyphrases
- synthetic data
- segmentation method
- parallel implementation
- preprocessing
- cost function
- data sets
- image processing
- gaussian distribution
- significant improvement
- experimental evaluation
- computational cost
- detection method
- clustering method
- high accuracy
- high precision
- computationally efficient
- input data
- mutual information
- probabilistic model
- training set
- neural network