Post-Composition Primal-Dual Algorithm for Texture, Geometry, and Noise Separation
In this paper, we propose a new algorithm for separating geometry and texture in noisy images. The proposed algorithm uses a novel primal-dual approach exploiting postcomposition. We show that the latter exhibits significantly better computational performance than standard forward-backward iterations and provides a more flexible formulation, opening the door to future integration with proximal neural networks.
Authors
- Luis M. Briceño-Arias (ORCID: https://orcid.org/0000-0001-9609-2015)
- Adrien Meynard (ORCID: https://orcid.org/0000-0002-6523-4735)
- Sylvain Meignen (ORCID: https://orcid.org/0000-0002-6713-0593)
- Nelly Pustelnik (ORCID: https://orcid.org/0000-0001-7310-1927)
Institutions
- École Normale Supérieure de Lyon (FR)
- Centre National de la Recherche Scientifique (FR)
- Laboratoire de Physique de l'ENS de Lyon (FR)
- Laboratoire Jean Kuntzmann (FR)
- Federico Santa María Technical University (CL)
- Université Grenoble Alpes (FR)
Publication Details
- Published
- 2025-09-08
- DOI
- https://doi.org/10.23919/eusipco63237.2025.11226623
- Primary Topic
- Image and Signal Denoising Methods
- Type
- article
- Field-Weighted Citation Impact
- 0.00