arXiv 2202.06688

Geometric Transformer for Fast and Robust Point Cloud Registration

By Zheng Qin, Hao Yu, et al.

Published 2022-02-14

Discussion

Read the public discussion and references gathered around this paper.

We study the problem of extracting accurate correspondences for point cloud registration. Recent keypoint-free methods bypass the detection of repeatable keypoints which is difficult in low-overlap scenarios, showing great potential in registration. They seek correspondences over downsampled superpoints, which are then propagated to dense points. Superpoints are matched based on whether their neighboring patches ove…

View the original paper on arXiv