FLOW-BASED POINT CLOUD COMPLETION NETWORK WITH ADVERSARIAL REFINEMENT

Point cloud completion is the task of estimating the complete point cloud from the partial observation. Most of the existing methods tend to recover global shapes of 3D objects and usually lack local details. These methods rely merely on distance metrics between point sets as loss functions, which have the insuf?cient capability of supervising ?ne structures. In this work, we propose a coarse-to-?ne approach to complete the partial point cloud with two stages: 1) Flow-based Completion Network, a principled probabilistic model that built on continuous normalizing ?ow to generate coarse completions conditioned on partial inputs. 2) Adversarial Re?nement Network, a hierarchical re?nement network constrained by the proposed patch discriminator to re?ne local details based on coarse completions. Experimental results show that our method can progressively complete 3D point clouds with ?ne details. Compared with other competitive methods, our method achieves better results on both quantitative and qualitative evaluations.