KineticNet: Deep learning a transferable kinetic energy functional for orbital-free density functional theory.
Roman RemmeTobias KaczunMaximilian ScheurerAndreas DreuwFred A. HamprechtPublished in: CoRR (2023)
Keyphrases
- deep learning
- energy functional
- euler lagrange
- mumford shah functional
- active contours
- level set
- image segmentation
- unsupervised learning
- partial differential equations
- machine learning
- energy function
- region growing
- shape prior
- active contour model
- curve evolution
- weakly supervised
- multiscale
- evolution equation
- natural images
- denoising
- semi supervised
- statistical models
- image analysis
- reinforcement learning
- training data
- computer vision