Regularization by architecture: A deep prior approach for inverse problems.
Sören DittmerTobias KluthPeter MaassDaniel Otero BaguerPublished in: CoRR (2018)
Keyphrases
- inverse problems
- image reconstruction
- global optimization
- regularization methods
- regularization method
- convex optimization
- optimization methods
- optimization problems
- early vision
- partial differential equations
- smoothness constraint
- prior information
- emission tomography
- computationally expensive
- edge preserving
- image restoration
- medical images
- simulated annealing
- evolutionary algorithm
- image processing
- neural network