Deep Unrolling for Nonconvex Robust Principal Component Analysis - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

Deep Unrolling for Nonconvex Robust Principal Component Analysis

Caroline Chaux
Emmanuel Soubies
Vincent y F Tan

Résumé

We design algorithms for Robust Principal Component Analysis (RPCA) which consists in decomposing a matrix into the sum of a low rank matrix and a sparse matrix. We propose a deep unrolled algorithm based on an accelerated alternating projection algorithm which aims to solve RPCA in its nonconvex form. The proposed procedure combines benefits of deep neural networks and the interpretability of the original algorithm and it automatically learns hyperparameters. We demonstrate the unrolled algorithm's effectiveness on synthetic datasets and also on a face modeling problem, where it leads to both better numerical and visual performances.
Fichier principal
Vignette du fichier
2307.05893.pdf (246.96 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04160961 , version 1 (13-07-2023)

Licence

Paternité

Identifiants

Citer

Elizabeth Z. C. Tan, Caroline Chaux, Emmanuel Soubies, Vincent y F Tan. Deep Unrolling for Nonconvex Robust Principal Component Analysis. IEEE MLSP 2023, Sep 2023, Rome, France. ⟨10.1109/MLSP55844.2023.10285962⟩. ⟨hal-04160961⟩
93 Consultations
63 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More