Discrete Gradient Flow Approximations of High Dimensional Evolution Partial Differential Equations via Deep Neural Networks.
Emmanuil H. GeorgoulisMichail LoulakisAsterios TsiourvasPublished in: CoRR (2022)
Keyphrases
- partial differential equations
- gradient flow
- difference equations
- high dimensional
- neural network
- level set
- numerical methods
- energy functional
- evolution equation
- anisotropic diffusion
- image denoising
- numerical solution
- image processing
- differential equations
- curve evolution
- multiscale
- pattern recognition
- discrete data
- image enhancement
- feature space
- parameter space
- high order
- variational methods
- active contours
- level set method
- regularization term
- natural images
- feature vectors