Regularized Dynamic Optimal Transport for Spatial-Temporal Image Generation
SIAM Journal on Applied Mathematics · 30.06.2026 · 86 (3) · 985-1007
DOI: 10.1137/25m1750998 ↗Kurz gefasst
Fachartikel in der Zeitschrift SIAM Journal on Applied Mathematics, online erschienen am 28.05.2026. Verfasst von Chen Li und 4 weiteren Autor:innen. Die Arbeit nennt 43 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Ministry of Science and Technology of China, Natural Science Foundation of Jiangsu Province.
Abstract (Originalsprache)
Optimal transport (OT) theory and models have been widely applied in the fields of computer vision, machine learning, and image analysis. Among three typical OT model formulations—the Monge, Kantorovich, and dynamical ones—the Kantorovich formulation with entropic regularization is most commonly used in practice due to its simple computational scheme. The other two formulations are rarely regularized since adding regularization terms brings additional difficulties to their theoretical analysis and numerical algorithms. In this paper, we propose a regularized OT model based on the dynamic formulation. We incorporate a [Formula: see text] regularization term on the momentum field to promote smoothness, and a total variation regularization term on the density field to enhance robustness to noise. We establish the existence and uniqueness of solutions to the regularized dynamic OT model and provide a theoretical analysis based on [Formula: see text]-convergence. Moreover, we develop a primal-dual algorithm to efficiently solve the resulting optimization problem. Numerical experiments on spatial-temporal image generation tasks demonstrate the effectiveness of the proposed model and the stability and efficiency of the proposed numerical algorithm. The proposed regularized dynamical OT model provides a flexible and theoretically grounded framework for image generation tasks in imaging science.
Förderung
- National Natural Science Foundation of China
- Ministry of Science and Technology of China
- Natural Science Foundation of Jiangsu Province
- China Postdoctoral Science Foundation
- Fundamental Research Funds for the Central Universities
Zitieren
APA
Li, C., Zhao, Q., Gao, Y., Liu, H., & Nie, Z. (2026). Regularized Dynamic Optimal Transport for Spatial-Temporal Image Generation. SIAM Journal on Applied Mathematics, 86(3), 985-1007. https://doi.org/10.1137/25m1750998
Vancouver
Li C, Zhao Q, Gao Y, Liu H, Nie Z. Regularized Dynamic Optimal Transport for Spatial-Temporal Image Generation. SIAM Journal on Applied Mathematics. 2026;86(3):985-1007. doi:10.1137/25m1750998
BibTeX
@article{Li2026,
author = {Li, Chen and Zhao, Qing and Gao, Yiming and Liu, Hairong and Nie, Ziwei},
title = {Regularized Dynamic Optimal Transport for Spatial-Temporal Image Generation},
journal = {SIAM Journal on Applied Mathematics},
year = {2026},
volume = {86},
number = {3},
pages = {985--1007},
doi = {10.1137/25m1750998}
}
EUVIMED