Simpler is better: spectral regularization and up-sampling techniques for variational autoencoders.
Sara BjörkJonas Nordhaug MyhreThomas Haugland JohansenPublished in: CoRR (2022)
Keyphrases
- denoising
- total variation regularization
- variational formulation
- random sampling
- variational framework
- basis pursuit
- sample size
- robust statistics
- image segmentation
- monte carlo
- hyperspectral images
- smoothness term
- spectral analysis
- sampling methods
- sampling algorithm
- regularization term
- image denoising
- multiscale
- image processing
- variational methods
- neural network
- energy functional
- regularization methods