ClusterGNN: Cluster-based Coarse-to-Fine Graph Neural Network for Efficient Feature Matching

Keyphrases

• Attention Matrix 25%
• Cluster-based 100%
• Coarse-to-fine 100%
• Coarse-to-fine Paradigm 25%
• Competitive Performance 25%
• Complete Graph 25%
• Computer Vision Tasks 25%
• Cross-attention 25%
• Dense Detection 25%
• Efficient Feature Matching 100%
• Feature Matching 50%
• Graph Neural Network 100%
• Matching Task 25%
• Memory Consumption 25%
• Neural Network 25%
• Neural Network Architecture 25%
• Number of Features 25%
• Progressive Clustering 25%
• Quadratic Complexity 25%
• Self-attention 25%
• Sparse Representation 25%
• Visual Features 25%

Computer Science

• Computer Vision Task 25%
• Graph Neural Network 100%
• Memory Consumption 25%
• Neural Network Architecture 25%
• Resulting Graph 25%
• Sparse Representation 25%
• Subgraphs 25%
• Visual Feature 25%