Self-supervised point cloud completion via inpainting
When navigating in urban environments, many of the objects that need to be tracked and
avoided are heavily occluded. Planning and tracking using these partial scans can be
challenging. The aim of this work is to learn to complete these partial point clouds, giving us
a full understanding of the object's geometry using only partial observations. Previous
methods achieve this with the help of complete, ground-truth annotations of the target
objects, which are available only for simulated datasets. However, such ground truth is …
avoided are heavily occluded. Planning and tracking using these partial scans can be
challenging. The aim of this work is to learn to complete these partial point clouds, giving us
a full understanding of the object's geometry using only partial observations. Previous
methods achieve this with the help of complete, ground-truth annotations of the target
objects, which are available only for simulated datasets. However, such ground truth is …
[CITATION][C] Self-Supervised Point Cloud Completion via Inpainting. arXiv 2021
H Mittal, B Okorn, A Jangid, D Held - arXiv preprint arXiv:2111.10701
Showing the best results for this search. See all results