PDE-GCN: Novel Architectures for Graph Neural Networks Motivated by Partial Differential Equations.
Moshe EliasofEldad HaberEran TreisterPublished in: NeurIPS (2021)
Keyphrases
- partial differential equations
- neural network
- anisotropic diffusion
- level set
- image denoising
- image processing
- numerical solution
- image enhancement
- fourth order
- multiscale
- differential equations
- pattern recognition
- nonlinear diffusion
- machine learning
- energy functional
- weighted graph
- mathematical morphology
- graph structure
- numerical algorithms
- numerical scheme
- finite difference
- diffusion equation
- high order
- reaction diffusion
- boundary value problem
- denoising
- numerical methods
- curve evolution
- image analysis
- natural images
- active contours
- methods in computer vision
- lattice boltzmann
- input image
- level set method
- scale space
- heat equation
- mumford shah model