Dimensions overall score 7.0

• Hg-I2P significantly outperforms existing methods in generalization and accuracy on six indoor and outdoor benchmarks under cross-domain setups.result 90%
• A heterogeneous graph representation that maps segmented 2D and 3D regions enhances cross-modal feature interaction and improves feature discriminability.method 85%
• Modeling consistency among vertices and edges within the heterogeneous graph enables pruning of unreliable 2D-3D correspondences.method 85%
• The method uses 2D and 3D SAM models to segment images and point clouds into non-overlapping regions, which form the vertices of the heterogeneous graph.technical 90%
• The heterogeneous edge set E_I2P is predicted by mining multi-path correlation features using an attention mechanism.technical 80%
• The HE-adapting module refines cross-modal features via heterogeneous edge-guided cross-attention and message interaction.method 85%
• The method addresses the challenge of the drastic modality gap between images and point clouds in I2P registration.technical 90%

Competitive landscape

Competitor map is still being generated for this paper. Enable generation or check back soon.

Startup potential card