Deakin University
Browse

File(s) under embargo

Weighted Point Cloud Normal Estimation

conference contribution
posted on 2023-10-03, 04:42 authored by W Wang, Xuequan Lu, Di ShaoDi Shao, Xiao LiuXiao Liu, Richard DazeleyRichard Dazeley, Antonio Robles-KellyAntonio Robles-Kelly, W Pan
Existing normal estimation methods for point clouds are often less robust to severe noise and complex geometric structures. Also, they usually ignore the contributions of different neighbouring points during normal estimation, which leads to less accurate results. In this paper, we introduce a weighted normal estimation method for 3D point cloud data. We innovate in two key points: 1) we develop a novel weighted normal regression technique that predicts point-wise weights from local point patches and use them for robust, feature-preserving normal regression; 2) we propose to conduct contrastive learning between point patches and the corresponding ground-truth normals of the patches' central points as a pre-training process to facilitate normal regression. Comprehensive experiments demonstrate that our method can robustly handle noisy and complex point clouds, achieving state-of-the-art performance on both synthetic and real-world datasets.

History

Volume

2023-July

Pagination

2015-2020

Location

Brisbane, Queensland

Start date

2023-07-10

End date

2023-07-14

ISSN

1945-7871

eISSN

1945-788X

ISBN-13

9781665468916

Language

eng

Publication classification

E1 Full written paper - refereed

Title of proceedings

ICME 2023 : Proceedings of the IEEE International Conference on Multimedia and Expo

Event

Multimedia and Expo. Conference (2023 : Brisbane, Queensland)

Publisher

IEEE

Place of publication

Piscataway, N.J.

Usage metrics

    Research Publications

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC