Unpaired Point Cloud Completion via Unbalanced Optimal Transport

Taekyung Lee, Jaemoo Choi, Jaewoong Choi, Myungjoo Kang
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:32965-32986, 2025.

Abstract

Unpaired point cloud completion is crucial for real-world applications, where ground-truth data for complete point clouds are often unavailable. By learning a completion map from unpaired incomplete and complete point cloud data, this task avoids the reliance on paired datasets. In this paper, we propose the Unbalanced Optimal Transport Map for Unpaired Point Cloud Completion (UOT-UPC) model, which formulates the unpaired completion task as the (Unbalanced) Optimal Transport (OT) problem. Our method employs a Neural OT model learning the UOT map using neural networks. Our model is the first attempt to leverage UOT for unpaired point cloud completion, achieving competitive or superior performance on both single-category and multi-category benchmarks. In particular, our approach is especially robust under the class imbalance problem, which is frequently encountered in real-world unpaired point cloud completion scenarios.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-lee25e, title = {Unpaired Point Cloud Completion via Unbalanced Optimal Transport}, author = {Lee, Taekyung and Choi, Jaemoo and Choi, Jaewoong and Kang, Myungjoo}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {32965--32986}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/lee25e/lee25e.pdf}, url = {https://proceedings.mlr.press/v267/lee25e.html}, abstract = {Unpaired point cloud completion is crucial for real-world applications, where ground-truth data for complete point clouds are often unavailable. By learning a completion map from unpaired incomplete and complete point cloud data, this task avoids the reliance on paired datasets. In this paper, we propose the Unbalanced Optimal Transport Map for Unpaired Point Cloud Completion (UOT-UPC) model, which formulates the unpaired completion task as the (Unbalanced) Optimal Transport (OT) problem. Our method employs a Neural OT model learning the UOT map using neural networks. Our model is the first attempt to leverage UOT for unpaired point cloud completion, achieving competitive or superior performance on both single-category and multi-category benchmarks. In particular, our approach is especially robust under the class imbalance problem, which is frequently encountered in real-world unpaired point cloud completion scenarios.} }
Endnote
%0 Conference Paper %T Unpaired Point Cloud Completion via Unbalanced Optimal Transport %A Taekyung Lee %A Jaemoo Choi %A Jaewoong Choi %A Myungjoo Kang %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-lee25e %I PMLR %P 32965--32986 %U https://proceedings.mlr.press/v267/lee25e.html %V 267 %X Unpaired point cloud completion is crucial for real-world applications, where ground-truth data for complete point clouds are often unavailable. By learning a completion map from unpaired incomplete and complete point cloud data, this task avoids the reliance on paired datasets. In this paper, we propose the Unbalanced Optimal Transport Map for Unpaired Point Cloud Completion (UOT-UPC) model, which formulates the unpaired completion task as the (Unbalanced) Optimal Transport (OT) problem. Our method employs a Neural OT model learning the UOT map using neural networks. Our model is the first attempt to leverage UOT for unpaired point cloud completion, achieving competitive or superior performance on both single-category and multi-category benchmarks. In particular, our approach is especially robust under the class imbalance problem, which is frequently encountered in real-world unpaired point cloud completion scenarios.
APA
Lee, T., Choi, J., Choi, J. & Kang, M.. (2025). Unpaired Point Cloud Completion via Unbalanced Optimal Transport. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:32965-32986 Available from https://proceedings.mlr.press/v267/lee25e.html.

Related Material