Wasserstein distributional harvesting for highly dense 3D point clouds

作者：

Highlights：

• Our method outputs the surface distributions and samples an arbitrary number of 3D points.

• Our stochastic instance normalization transfers the implicit distribution into other distributions.

• Our method trains the generative model using the progressive sampling strategy.

摘要

•Our method outputs the surface distributions and samples an arbitrary number of 3D points.•Our stochastic instance normalization transfers the implicit distribution into other distributions.•Our method trains the generative model using the progressive sampling strategy.

论文关键词：3D point cloud harvesting,Progressive sampling,Stochastic instance normalization

论文评审过程：Received 12 July 2021, Revised 27 June 2022, Accepted 13 August 2022, Available online 16 August 2022, Version of Record 19 August 2022.

论文官网地址：https://doi.org/10.1016/j.patcog.2022.108978