Self-Supervised Global-Local Structure Modeling for Point Cloud Domain Adaptation with Reliable Voted Pseudo Labels

Page view(s)
35
Checked on Oct 25, 2024
Self-Supervised Global-Local Structure Modeling for Point Cloud Domain Adaptation with Reliable Voted Pseudo Labels
Title:
Self-Supervised Global-Local Structure Modeling for Point Cloud Domain Adaptation with Reliable Voted Pseudo Labels
Journal Title:
EEE / CVF Computer Vision and Pattern Recognition Conference (CVPR)
DOI:
Keywords:
Publication Date:
20 June 2022
Citation:
@inproceedings{fan2022self, title={Self-Supervised Global-Local Structure Modeling for Point Cloud Domain Adaptation With Reliable Voted Pseudo Labels}, author={Fan, Hehe and Chang, Xiaojun and Zhang, Wanyue and Cheng, Yi and Sun, Ying and Kankanhalli, Mohan}, booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition}, pages={6377--6386}, year={2022} }
Abstract:
In this paper, we propose an unsupervised domain adaptation method for deep point cloud representation learning. To model the internal structures in target point clouds, we first propose to learn the global representations of unlabeled data by scaling up or down point clouds and then predicting the scales. Second, to capture the local structure in a self-supervised manner, we propose to project a 3D local area onto a 2D plane and then learn to reconstruct the squeezed region. Moreover, to effectively transfer the knowledge from source domain, we propose to vote pseudo labels for target samples based on the labels of their nearest source neighbors in the shared feature space. To avoid the noise caused by incorrect pseudo labels, we only select reliable target samples, whose voting consistencies are high enough, for enhancing adaptation. The voting method is able to adaptively select more and more target samples during training, which in return facilitates adaptation because the amount of labeled target data increases. Experiments on PointDA (ModelNet-10, ShapeNet-10 and ScanNet-10) and Sim-to-Real (ModelNet-11, ScanObjectNN-11, ShapeNet-9 and ScanObjectNN-9) demonstrate the effectiveness of our method.
License type:
Publisher Copyright
Funding Info:
This research / project is supported by the Agency for Science, Technology and Research (A*STAR) - AME Programmatic Funding Scheme
Grant Reference no. : A18A2b0046
Description:
© 2022 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
ISBN:
17602-17611
Files uploaded:

File Size Format Action
cvpr2022-paper.pdf 1,001.21 KB PDF Open