Training a Better Loss Function for Image Restoration.
Aamir MustafaAliaksei MikhailiukDan-Andrei IliescuVarun BabbarRafal K. MantiukPublished in: CoRR (2021)
Keyphrases
- image restoration
- loss function
- total variation
- pairwise
- support vector
- edge preserving
- regularization parameter
- image processing
- markov random field
- iterative image restoration
- super resolution
- regularization term
- regularization method
- risk minimization
- modified hopfield neural network
- reproducing kernel hilbert space
- higher order
- blurred images
- kernel function
- denoising
- supervised learning
- degraded images
- restored image
- blind image restoration