Data-Driven Regularization with Weak Convexity for Robust Image Reconstruction
SIAM Review · 30.09.2026 · 68 (3) · 611-634
DOI: 10.1137/26m1886638 ↗Kurz gefasst
Fachartikel in der Zeitschrift SIAM Review, online erschienen am 06.08.2026. Verfasst von Alexis Goujon und 3 weiteren Autor:innen. Die Arbeit nennt 78 Literaturangaben. Gefördert von: Swiss National Science Foundation, Horizon 2020 Framework Programme.
Abstract (Originalsprache)
We describe a practical framework for data-driven regularization in image reconstruction. It combines model expressivity with the guarantees of variational methods. The approach is based on a weakly convex ridge regularizer, defined as the composition of a convolutional filter bank and pointwise potentials constrained to be weakly convex. These potentials are implemented as learnable splines with a strict control of their weak-convexity modulus. The resulting denoisers outperform classic convex regularization techniques as well as competitive benchmarks such as BM3D, while they still correspond to the minimization of a convex energy. The learned regularizers further extend to general inverse problems with provable convergence to critical points. Overall, this framework shows that a controlled relaxation of convexity enables the design of learnable priors that achieve strong empirical performance while preserving mathematical guarantees.
Förderung
- Swiss National Science Foundation
- Horizon 2020 Framework Programme
Zitieren
APA
Goujon, A., Neumayer, S., Ducotterd, S., & Unser, M. (2026). Data-Driven Regularization with Weak Convexity for Robust Image Reconstruction. SIAM Review, 68(3), 611-634. https://doi.org/10.1137/26m1886638
Vancouver
Goujon A, Neumayer S, Ducotterd S, Unser M. Data-Driven Regularization with Weak Convexity for Robust Image Reconstruction. SIAM Review. 2026;68(3):611-634. doi:10.1137/26m1886638
BibTeX
@article{Goujon2026,
author = {Goujon, Alexis and Neumayer, Sebastian and Ducotterd, Stanislas and Unser, Michael},
title = {Data-Driven Regularization with Weak Convexity for Robust Image Reconstruction},
journal = {SIAM Review},
year = {2026},
volume = {68},
number = {3},
pages = {611--634},
doi = {10.1137/26m1886638}
}
EUVIMED