HyperGCT: A Dynamic Hyper-GNN-Learned Geometric Constraint for 3D Registration
It addresses a domain-specific challenge in 3D registration by improving accuracy and robustness to noise, offering an incremental advance over existing graph-based methods.
The paper tackles the problem of noise in geometric constraints for 3D point cloud registration by proposing HyperGCT, a dynamic hyper-GNN-learned method that mines high-order consistency among correspondences, achieving state-of-the-art performance on datasets like 3DMatch and KITTI-LC.
Geometric constraints between feature matches are critical in 3D point cloud registration problems. Existing approaches typically model unordered matches as a consistency graph and sample consistent matches to generate hypotheses. However, explicit graph construction introduces noise, posing great challenges for handcrafted geometric constraints to render consistency. To overcome this, we propose HyperGCT, a flexible dynamic Hyper-GNN-learned geometric ConstrainT that leverages high-order consistency among 3D correspondences. To our knowledge, HyperGCT is the first method that mines robust geometric constraints from dynamic hypergraphs for 3D registration. By dynamically optimizing the hypergraph through vertex and edge feature aggregation, HyperGCT effectively captures the correlations among correspondences, leading to accurate hypothesis generation. Extensive experiments on 3DMatch, 3DLoMatch, KITTI-LC, and ETH show that HyperGCT achieves state-of-the-art performance. Furthermore, HyperGCT is robust to graph noise, demonstrating a significant advantage in terms of generalization.