Revisiting convolutional neural network on graphs with polynomial approximations of Laplace-Beltrami spectral filtering.
Shih-Gu HuangMoo K. ChungAnqi QiuPublished in: Neural Comput. Appl. (2021)
Keyphrases
- convolutional neural network
- laplace beltrami
- heat kernel
- spectral analysis
- graph laplacian
- shape analysis
- face detection
- graph structure
- weighted graph
- point cloud
- euclidean distance
- minimum spanning tree
- d mesh
- neural network
- riemannian manifolds
- wiener filter
- adjacency matrix
- euclidean space
- directed graph
- basis functions
- image restoration
- complex shapes
- laplacian eigenmaps
- embedding space
- spanning tree
- neighborhood graph
- graph construction
- spectral clustering
- linear combination
- random walk
- multiscale