Adaptive Loss Function for Super Resolution Neural Networks Using Convex Optimization Techniques.
Seyed Mehdi AyyoubzadehXiaolin WuPublished in: CoRR (2020)
Keyphrases
- loss function
- super resolution
- empirical risk
- risk minimization
- hinge loss
- low resolution
- high resolution
- image reconstruction
- image restoration
- pairwise
- convex loss functions
- low resolution images
- resolution enhancement
- motion estimation
- super resolution reconstruction
- high quality
- depth map
- single image super resolution
- high resolution images
- support vector
- regularization term
- multi frame
- super resolved
- image processing
- light field
- image super resolution
- reproducing kernel hilbert space
- higher resolution
- image patches
- higher order
- linear combination
- feature vectors
- feature extraction
- super resolution algorithm
- image sequences
- feature selection