GTAINet: Graph neural network-based two-stage anomaly identification for locking wire point clouds using hierarchical attentive edge convolution.
Zhixue WangYu ZhangTongfei LvLin LuoPublished in: Int. J. Appl. Earth Obs. Geoinformation (2022)
Keyphrases
- high resolution
- point cloud
- laser scanner
- weighted graph
- image processing
- surface reconstruction
- structure from motion
- edge weights
- point sets
- undirected graph
- point cloud data
- neural network
- edge detection
- anomaly detection
- normal vectors
- triangular mesh
- architectural scenes
- multi view stereo
- icp algorithm
- stereo camera
- graph matching
- data objects
- shortest path
- image registration
- graph structure
- database systems
- multi view reconstruction
- directed graph