A robust Lagrangian-DNN method for a class of quadratic optimization problems.
Naohiko ArimaSunyoung KimMasakazu KojimaKim-Chuan TohPublished in: Comput. Optim. Appl. (2017)
Keyphrases
- high accuracy
- image noise
- significant improvement
- high precision
- detection method
- clustering method
- experimental evaluation
- computational cost
- preprocessing
- optimization method
- energy function
- computationally efficient
- prior knowledge
- objective function
- cost function
- k means
- segmentation method
- pairwise
- computational complexity
- training process
- multi class classification
- neural network