Convex Regularization Behind Neural Reconstruction.
Arda SahinerMorteza MardaniBatu OzturklerMert PilanciJohn M. PaulyPublished in: CoRR (2020)
Keyphrases
- horizontal and vertical projections
- neural network
- risk minimization
- convex optimization
- convex formulation
- image reconstruction
- network architecture
- minimization problems
- reconstruction process
- three dimensional
- penalty functions
- discrete tomography
- inverse problems
- reconstruction method
- neural model
- compressive sensing
- prior information
- piecewise linear
- image processing
- convex relaxation
- bio inspired
- tomographic reconstruction
- bregman divergences
- graph cuts
- loss function
- total variation
- augmented lagrangian
- biologically inspired
- regularization method
- convex sets
- compressed sensing
- regularization term
- globally optimal
- data dependent