Abstract: In this paper, we propose a novel network, SVDFormer, to tackle two specific challenges in point cloud completion: understanding faithful global shapes from incomplete point clouds and ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback