DGC-GNN: Descriptor-free Geometric-Color Graph Neural Network for 2D-3D Matching.
Shuzhe WangJuho KannalaDaniel BarathPublished in: CoRR (2023)
Keyphrases
- neural network
- geometric consistency
- graph matching
- attributed relational graph
- keypoints
- feature descriptors
- category recognition
- maximum cardinality
- bipartite graph matching
- color images
- invariant descriptors
- artificial neural networks
- sift descriptors
- geometric constraints
- partial matching
- scale and rotation invariant
- color descriptors
- bipartite graph
- color information
- dominant color
- graph theoretical
- maximum weight
- matching algorithm
- consistency check
- shape context
- color space
- object recognition
- weighted graph
- feature representation
- neural network is trained
- association graph
- image registration
- graph theory
- appearance descriptors
- scale invariant feature transform
- geometric transformations
- graph structure
- maximum matching
- image matching
- subgraph isomorphism
- back propagation
- pattern matching
- directed graph
- shape descriptors
- image retrieval
- shape matching
- matching process
- affine invariant
- relational structures
- color features
- color histogram
- image regions
- pose estimation
- random walk
- image features
- face recognition